The AI

Decision Brief
Insights from Microsoft and AI leaders on
navigating the generative AI platform shift
 Table of contents 2

03 04 21 31
Introduction Chapter 01 Chapter 02 Chapter 03
Driving AI impact with an The current state The future potential Advancing generative AI
evolving tech paradigm of generative AI of generative AI in your organisation

05 Section 1.1: The generative AI 23 Section 2.1: Who’s driving innovation 32 Section 3.1: Best practices for
platform shift 25 H
 ow software development companies are paving implementing AI at scale
the way for AI transformation – Jason Graefe
07 
Empowering the future: AI access for all –
Brad Smith
33 Section 3.2: Advancing generative AI
26 Section 2.2: Partnering together to build
with the five drivers of value
08 Section 1.2: Navigating the AI trustworthy AI
platform shift 27 
How to stay ahead of emerging challenges and
cyberthreats – Vasu Jakkal 36 Section 3.3: See how organisations across
10 
How AI is revolutionising IT at Microsoft –
Nathalie D’Hers industries are using generative AI
28 Section 2.3: AI and sustainability
12 Section 1.3: Addressing challenges to
generative AI adoption 30 Section 2.4: Innovators on the
14 
Learnings on the business value of AI from IDC – forefront of AI
Alysa Taylor

15 Section 1.4: The winning formula

17 The future of work is AI-powered – Jared Spataro

18 Section 1.5: Building trust with

generative AI
20 M
 icrosoft’s commitment to supporting customers
on their AI journey – Judson Althoff
INTRODUCTION 3

Driving AI impact with

an evolving tech paradigm
Over the last few years, generative AI has or we invest in the research that drives cutting-
experienced phenomenal growth – creating edge advances, our mission to empower every
exciting new ways for organisations to increase person and every organisation on the planet
productivity, streamline processes and to achieve more remains constant.
accelerate innovation. As adoption continues
and generative AI applications become more As we support leading innovators and help
powerful, the question is no longer whether businesses move from talking about AI to
organisations will embrace generative AI – but translating it into real and lasting results, we’ve
how they can use it most effectively. gained a unique perspective on the generative AI
strategies that drive business impact.
Here at Microsoft, we’ve built our foundation on
the belief that we should create technologies In The AI Decision Brief, our goal is to share
that enable others to create their own. The insights to help navigate the AI platform shift and
technology itself is always evolving – this is show how generative AI is impacting businesses
known. This next generation of AI will reshape today. We also highlight what’s on the horizon
every software category and every business, and share key strategies you can implement to
including our own. As a company, we remain maximise your opportunities in embracing these
committed to helping every individual, powerful new tools – while avoiding the pitfalls
organisation and industry to adapt to each new that could limit your success.
technological paradigm shift. This commitment
has positioned us as leaders at the forefront of We hope these perspectives inspire,
the latest shift – the AI platform shift. energise and guide you as you continue
on your AI journey.
In all we do, whether we build the cloud
infrastructure that powers best-in-class
foundation models and services like ChatGPT,
 4

Chapter 01
The current state
of generative AI
 Chapter 01: 5
The current state of generative AI

SECTION 1.1

The generative AI platform shift

It’s easy to get excited over AI Scaling laws – the fundamental forces driving Moore’s Law, on the other hand, is a long-
the generative AI platform shift standing observation in the field of computer
innovation – we’ve entered a new age
of generative AI that’s transforming
Tech paradigm shifts science, named after Gordon Moore, cofounder

of the last 50 years With every platform shift, it’s helpful to of Intel. Under Moore’s Law, computing power
productivity for every individual, understand underlying forces driving the change roughly doubled every two years with the
organisation and industry on earth. and one crucial driver has been the evolution number of transistors in an integrated circuit.
But it’s also challenging to navigate of AI scaling laws. These refer to the principles Under new scaling laws, AI capabilities can
the changes as the speed and scale Client Mobile that describe how the performance of AI models double every six months.1 As organisations use AI
Internet and AI
of AI breakthroughs are driving server
Cloud
improves as the number of computational to tackle complex challenges, the impact of this
resources, data and model size (parameters) accelerated pace will be felt everywhere and it’s
a new technological paradigm shift
increases. These laws suggest that larger models this rate of change that makes the AI platform
that’s fundamentally reshaping our
trained on more data tend to perform better shift uniquely transformative.
businesses and society. Leaders on a variety of tasks. Scaling laws not only
who envision new possibilities and guide researchers and practitioners in building
embrace the full potential of AI better models but also drive the design of the AI
have an opportunity to achieve infrastructure needed to sustain AI’s growth. This
unprecedented impact and growth – concept is crucial for understanding the rapid
and to help us address some of our advancements in AI capabilities.

most pressing challenges.

 Chapter 01: 6
The current state of generative AI

The emergence of generative AI Today, generative AI is transforming business

operations, employee and customer experiences
Generative AI is a subset of AI that uses and even day-to-day life. Whether it’s automating
advanced models, such as large language mundane tasks and making business intelligence
models (LLMs), to generate novel outputs. more accessible or personalising customer
This allows users to work with data in powerful service and enabling a new wave of creativity,
new ways using natural language prompts, generative AI represents a paradigm shift from
enabling a wide range of applications. laborious process to extraordinary possibility.

75 %
Generative AI capabilities have already The historic pace of generative AI adoption
improved substantially since ChatGPT, the
first breakthrough example of this technology, Generative AI has been adopted with
was introduced in November 2022. unprecedented speed. While it took seven
years for the internet to reach 100 million users,
used AI in 2024 • While early generative AI tools used ChatGPT achieved that benchmark in two
natural language interfaces that could only months.2 And even though generative AI is
understand text inputs, today’s universal relatively new to the market, adoption is rapidly
interfaces can also understand images expanding as usage among enterprises jumped
and video. to 75% in 2024 from 55% in 2023, according to an
IDC study.3
• More sophisticated generative AI models can
handle increasingly complex tasks, allowing
them to move from only performing reasoning
Generative AI usage jumped from
processes to being able to support both
55% in 2023 to 75% in 2024 reasoning and planning.

• Generative AI applications now have

short-term and long-term memory, enabling
them to ‘remember’ important things
across conversations, devices and contexts.
 Chapter 01: 7
The current state of generative AI

LEADERSHIP PERSPECTIVES

Empowering the future:

HIGHLIGHTS

AI leaders drive business value and

AI access for all

innovation

Leaders who have embraced generative AI and made

significant progress in their AI adoption journey see
By Brad Smith, Vice Chair and President another benefit in embracing an AI-driven future:
disruption. Beyond simply making existing systems
more efficient, they see generative AI as a way to
reinvent processes and business models to create
Artificial intelligence is quickly becoming what platform or choose to move to another opportunities for workers, respects enduring a future that’s more personalised, with less friction
economists call a general-purpose technology provider. These principles are not just about values of individuals and addresses the impact and achieving higher quality outcomes.
(GPT) that will transform every sector of the common sense governance; they ensure that of AI on local resources such as land, energy
economy, like the profound and lasting impacts AI development and deployment promote and water. • Watch Karin Conde-Knape, Novo Nordisk SVP of
of the printing press and electricity. Since the innovation and fair competition for both existing Global Drug Discovery, share how AI is helping
the company deliver the next generation of
first industrial revolution in the 1700s, the broad enterprises and emerging new companies. Ultimately, the world needs AI that is not only
treatments faster.
adoption of GPTs has been the key to a country’s more powerful, but also broadly accessible.
economic growth and prosperity. Building AI skills across engineers, systems It will take new partnerships to build and deploy • Watch Kelle Fontenot, KPMG Chief Digital Officer,
integrators and users is foundational for expensive hyperscale data infrastructure, and explain how generative AI could completely reshape
As generative AI transforms industries, spreading the adoption of AI. Historical make sure the world’s most trustworthy AI how the company delivers services.
Microsoft’s commitment to responsible and examples show that developing worker skills platforms are available to all countries seeking to
• Watch Aditya Thandani, H&R Block Vice President
ethical AI implementation is crucial. That’s why has been critical to the spread of all major GPTs. promote security and prosperity both for their of Artificial Intelligence Platforms, share how
we introduced AI Access Principles that address Countries will need to develop national AI skilling citizens and people around the world. generative AI will enable H&R Block to be more
our growing role and responsibility as an AI strategies that build upon existing disciplines responsive to client needs and deliver a compelling
infrastructure provider and leader in innovation. like computer and data science to facilitate the Learn more about the Microsoft’s AI Access experience.
spread of AI across their economies. Principles including our Responsible AI principles
By publishing these principles, Microsoft and approach.
Learn more in the AI Challengers thought leadership
commits to providing broad technology access Trust in technology and providers is another
video series. Hear C-suite executives share how
to empower organisations and individuals key factor in the adoption of AI. It must solve they approach AI – not only as part of their business
worldwide to develop and use AI in ways real-world problems while ensuring safety strategy, but by integrating it into their culture
that serve the public good. We also commit and security. Broad social acceptance for AI and operations to drive business value and deliver
to fairness for developers who build on our will depend on ensuring that AI creates new innovation.
 Chapter 01: 8
The current state of generative AI

SECTION 1.2

Navigating the AI platform shift

Commercial AI Readiness shifts – exploring becoming
planning, scaling becoming realising

AI Business Readiness *Adjusted

Where organisations are today The five stages of generative AI readiness Implementing
Businesses are moving from proofs of
Our research shows that businesses adopting concept and pilots into production.
As more organisations feel confident
AI progress through five stages of readiness,
and comfortable with generative AI each with its own strategic priorities: exploring, Scaling
technology, AI usage and sentiment planning, implementing, scaling and Businesses have deployed generative
is trending upwards. People expect AI realising. For business leaders, correctly AI applications and systems for a few
to have a positive impact on society, identifying the stage of their organisation’s business functions and are scaling across
generating real business value. This readiness and implementing the appropriate the entire organisation.
sentiment is impacting AI readiness as strategies is critical to managing generative
AI transformation. Realising
more organisations are moving from
Businesses are realising repeatable,
exploring AI to planning AI strategy – Exploring measurable value. Not only have they
and progressing from scaling to Businesses are educating themselves and piloted, implemented and scaled
realising consistent AI value. experimenting with generative AI in select applications, but they’ve also built
areas of the organisation. strong processes and guardrails for
further development and have a robust
Planning development pipeline.
To get a sense of your organisation’s Businesses are actively assessing and
level of AI readiness, we’ve created defining a generative AI strategy. Building
a simple 10-question assessment. on their strategic foundation, they run
April 2024 June 2024 September 2024
Start your AI readiness assessment. proofs of concept and plan generative
AI deployments.
Exploring Implementing Realising
To explore the priorities and Planning Scaling
opportunities at each stage, see
Chapter 3. Adjusted scores based on AI Readiness overstatement adjustment. Base: Commercial (n=400).
Weighted. Bis_AIMATUR: On a scale of 1-5, how would you describe your organisation’s overall
adoption with regards to AI? Please answer where one is beginning stage of exploring the role of
AI in the organisation, and five is a relatively advanced stage where the organisation can realise
and measure the value driven by AI. Select only one. Data visualisation on this page is not to
mathematical scale.
 Chapter 01: 9
The current state of generative AI

The five drivers of AI value 

Technology and data strategy • Assessing whether the workforce has AI governance
An effective strategy for addressing the diverse skills and experiences to Businesses overcome the challenges
Microsoft has identified five key drivers of AI technology and data challenges focuses achieve generative AI priorities. associated with security and data
value: business strategy, technology and on ensuring access to the data and governance by developing enterprise-
data strategy, AI strategy and experience, infrastructure needed to fuel and run • Creating cross-functional teams wide AI governance policies that embrace
organisation and culture and AI governance.4 generative AI applications at scale. that work across business units to the framework of Trustworthy AI.
Businesses who establish cross-functional teams Businesses that implement generative AI build experience. AI governance involves:
to manage each of these drivers effectively see initiatives successfully focus on:
higher ROI from their generative AI investments • Applying the right models for each • Implementing effective and transparent
and can move more quickly through the • Access to quality data that’s accurate and generative AI initiative. processes, controls and accountability
stages of AI readiness to the point of realising properly formatted. structures.
lasting value.  rganisation and culture
O
• Where applications and data are stored. A focus on organisation and culture is • Governing data privacy, security and
Business strategy essential to manage change effectively, responsible use.
An effective business strategy focuses • Ensuring cloud infrastructure encourage use of generative AI tools and
generative AI adoption on achieving the supports scaling. establish a mindset that invites change and • Measuring and assessing the value and
organisation’s business goals, aligning seeks continuous growth. Businesses that impact of generative AI investments on
investment with desired, measurable • Determining how to make decisions manage organisation and culture most business processes and outcomes.
outcomes. Businesses that accelerate their about whether to buy or build generative effectively focus on:
generative AI journey use this strategy AI applications strategically. Learn how to build and optimise your strategic
to address: • Top-down leadership support. plan for AI. Read the AI Strategy Roadmap:
 strategy and experience
AI Navigating the stages of value creation.5
• The business outcomes they’re driving. Organisations employ generative AI • Clearly defining an operating model
with greater impact as they accumulate to industrialise generative AI adoption
• How generative AI can help them achieve experience and expertise. Businesses and use.
those outcomes. develop this resource through
a combination of hiring and skilling • Effective change management and
• The objectives and use cases that are programs and organisational change. continuous learning.
most meaningful. These efforts emphasise:
• Building strong business partnerships
• A clear investment plan for achieving to support and accelerate generative
strategic goals. AI initiatives.
 Chapter 01: 10
The current state of generative AI

LEADERSHIP PERSPECTIVES

How AI is revolutionising IT at Microsoft

By Nathalie D’Hers, CVP Microsoft Digital (IT)

Microsoft Digital powers, protects and our employees build new habits with Copilot. Recognising that agents are the next step in to get the answers and take action. In our testing,
transforms the digital employee experience We’ve empowered over 4,000 Copilot champions our AI journey also required a deployment employees were 36% to 42% more successful
across devices, applications and infrastructure. who share their knowledge and enthusiasm to and adoption strategy. First, we needed to using these AI-powered agents than using
At Microsoft, AI has been an incredible catalyst, inspire their peers. We’ve made change fun with consider how our employees could leverage conventional self-help. In fact, our testing was
accelerating our own digital transformation in programs like Camp Copilot that gamify the simple knowledge retrieval with services like so successful we’ve already announced that
ways that would have been impossible just a few learning experience. And we’ve documented our Agent Builder in SharePoint, then consider more we’re bringing these agents to market so our
years ago. deployment and adoption journey at Microsoft advanced scenarios, like agents that can also customers can take advantage too. We’re already
so you can learn from our experience. take action on an employee’s behalf or more working on additional self-service capabilities
Part of our mission is to be ‘Customer Zero’ – autonomous agents. Our guidance is to identify as part of this new suite.
Microsoft’s first and best customer – and then Enabling an agentic Microsoft. As revolutionary a few pain points in your environment and
share our learnings with customers. In that spirit, as Copilot is, the promise of agents is just as address those first, using the lessons to inform Learn how AI is already changing work –
we’ve learned a lot about enterprise AI, especially exciting. To enable an agentic Microsoft required your future investments. Microsoft included.
ways to accelerate time to employee value. us to ensure our Microsoft 365 estate was in
order. That meant careful governance of our Innovating with employee self-service.
Change is hard! A capability as revolutionary enterprise data to ensure it isn’t overshared At Microsoft, we identified tech support and
as Microsoft 365 Copilot requires careful change or overexposed. HR inquiries as two pain points ideally suited
management to maximise impact and minimise for agents. We built a new Copilot-powered
time to value. Microsoft Digital has dedicated capability called Employee Self-Service Agent
change management teams globally to help in Microsoft 365 Copilot that makes it easy
 Chapter 01: 11
The current state of generative AI

CUSTOMER STORIES

Organisations driving efficiency with AI tools

Dentsu reduces time to media insights Crediclub saves 96% per month Eaton saves more than 650 hours and
by 90% using Azure AI and Copilot in audit processes improves operational efficiency
Dentsu, a multinational advertising and media Crediclub, a Mexican financial services company, Eaton, a global leader in intelligent power management,
company, needed to speed up its complex data recognised the need to enhance the operational needed to manage thousands of standard operating
analytics process. Dentsu used Microsoft 365 Copilot, management of over 150 branches to maintain the procedures (SOPs) and improve data access across
Microsoft Azure AI Foundry and Azure OpenAI Service consistency and quality of its processes. To optimise teams, prompting the company to find a solution
to build a predictive analytics agent. According to the audit process and reduce costs, Crediclub made to streamline operations, automate documentation
dentsu, this solution has already led to employees a significant leap in its digital transformation by and support knowledge transfer. Eaton adopted
saving 15 to 30 minutes a day using Copilot for tasks implementing an AI-based solution. With Microsoft’s Microsoft 365 Copilot to automate the creation of
such as summarising chats, generating presentations support, the company developed an innovative 1,000 SOPs, streamline customer service operations
and building executive summaries. Takuya Kodama, system on Microsoft Azure infrastructure, integrating and improve data access. With Copilot, Eaton estimates
Business Strategy Manager at dentsu, has stated, services such as Microsoft Azure AI, Microsoft Azure that SOP creation time has dropped from one hour
“Copilot has transformed the way we deliver Cosmos DB and Microsoft Azure OpenAI Service. to just 10 minutes – an 83% reduction – saving more
creative concepts to our clients, enabling real-time “Thanks to this project, 800 sales advisers and than 650 hours and significantly improving operational
collaboration. Agility, security and uniqueness are 150 branch managers now have an additional efficiency across the 1,000 SOPs aided by Copilot.
crucial, but our goal is to lead this transformation 1,600 hours per day to interact directly with customers, “Copilot gave us the efficiency we needed to stay
company-wide, from top to bottom.” without the need to travel,” says Gabriel Ruelas, on track and document our processes,” says Tom
Director of Operations and Co-founder of Crediclub. Doyle, Senior Project Manager at Eaton.
Read the Azure story
Read the Copilot story Read the story Read the story
 Chapter 01: 12
The current state of generative AI

SECTION 1.3

Addressing challenges to generative AI adoption

Implementing generative AI and I dentifying where to start Properstar, a leading real estate platform, high demand for AI skillsets will naturally
The possibilities for generative AI seem used Microsoft Azure AI services to create drive the workforce to become more
achieving its benefits requires
virtually limitless – which means for many an app that analyses unstructured real AI-competent, business leaders currently
organisations to build a strong
organisations, it’s difficult to identify estate data, provides advanced filtering struggle to hire and train workers fast
foundation of infrastructure, where to start. To provide that answer, capabilities and delivers relevant search enough to support AI initiatives.7
knowledge and capabilities. As it’s important to consider the business results – cementing their position as
organisations move forward in value or strategic outcome you are aiming a leading player in the field of data-driven To effectively build AI skills within a diverse
their AI journeys, several common to achieve. Many organisations begin with real estate markets. workforce, it’s essential to remember
challenges have emerged that can a goal to boost employee productivity that every organisation is different and
or optimise existing workflows to be more  akeaway: Go beyond optimisation and
T with a technology category this dynamic,
hinder progress and delay their ability
efficient, as in the case of DLA Piper, one efficiency and discover transformative use there’s no one-size-fits-all approach to
to realise value. For business leaders,
of the world’s largest law firms. cases that deliver meaningful business learning and development. Experimenting
clear strategic planning and goal value. Read the AI Use Cases for Business with new initiatives and approaches will
setting are critical to maximise value  ut generative AI can also be a force
B Leaders eBook6 for more ideas on how to help organisations build comprehensive
and minimise risk. for true business transformation when realise value with AI. learning playbooks that evolve and serve
you look at more specific pain points or the shifting needs of employees.
opportunities for innovation. For example,  vercoming AI skill shortages and
O
Babson College used AI to transform building expertise Takeaway: Use real-world examples to
learning and the student journey. Starting Organisations need employees with build an approach to employee skilling that
with pilots to support struggling students the right skills to manage AI-optimised will meet your organisation’s unique needs.
and provide learning assistance, they infrastructure, build and train models Read 10 best practices from Microsoft’s
have now launched ‘The Generator’, and work effectively with AI tools. Today, skilling initiatives. Plus, see how Wipro
a grassroots community that empowers that talent is in short supply. 45% of approached its skilling challenge.
faculty to explore and integrate AI with organisations say a lack of personnel with
confidence, which is providing guidance specialised skills is the top challenge when
on which AI tools to use at scale. And implementing AI technology. Although
 Chapter 01: 13
The current state of generative AI

 rom security concerns to

F  akeaway: While generative AI is
T Takeaway: Organisations are balancing two Executive leadership needs to help drive
protecting with AI contributing to a more complicated goals – innovating with AI and remaining this initiative to establish the culture of the
AI adoption can present unique security cybersecurity environment, it can also be compliant with applicable regulation. At organisation.
and data governance challenges, ranging a key part of your strategy to protect your Microsoft, we’re ready to help customers
from concerns over data privacy and organisation. Learn more about safeguards do both things at once. Learn more about Takeaway: Successful generative AI change
confidentiality to potential vulnerabilities in that can help mitigate risk, and learn how our commitment to responsible AI and our management must encompass three
the models being used. Threat actors also organisations like Eastman and AXA solved approach to regulations like the EU AI Act. factors: leadership, human change and
understand the value of AI; they’re using it their security challenges. technical readiness. Explore best practices
to generate malicious code, increase the Change management across for change management and adoption in
effectiveness of social engineering attacks Regulation and compliance changes the organisation Chapter 3. Or, explore resources like the
and multiply the volume of attacks they will gather pace AI adoption drives monumental changes Copilot Success Kit or Azure AI Adoption
carry out. As a result, 27% of organisations As generative AI evolves, industry groups in the way organisations operate. With guidance and learn how Absa and PKSHA
say concerns over data or IP loss hold their and governments are setting standards so many ways to apply generative AI in embraced change across their organisation.
AI initiatives back, while 26% cite a lack of to guide its use. Organisations must adapt particular, this technology can become
governance and risk management.8 to comply with this complex and sometimes integrated into employees’ work at every
fragmented international regulatory level of the organisation. This means that
While challenges such as AI-assisted threats environment. Embracing this challenge change management isn’t a matter of just
and the vulnerability of some AI models and is crucial, and leveraging resources to adopting a new tool – workers need to
tools will never go away, the emergence of stay compliant and use AI responsibly completely rethink their existing workflows,
more sophisticated AI-powered security is essential. Microsoft is committed to requiring them to break old habits and
capabilities is improving threat detection, assisting customers in this effort. embrace new ones. Thus, both top-down
accelerating response times and providing and bottom-up approaches are essential
more in-depth incident analysis.9 to drive the adoption of generative AI.
 Chapter 01: 14
The current state of generative AI

LEADERSHIP PERSPECTIVES

Learnings on the business value of AI

from IDC
By Alysa Taylor, Chief Marketing Officer, Commercial Cloud and AI

Generative AI has become a key driver of  nhanced productivity has become

E  Generative AI adoption and value is Leaders are also implementing AI solutions
business success across industries, with table stakes. Employee productivity is growing across industries. Generative AI faster, with 29% deploying AI in less than
companies just beginning to explore its vast the primary business outcome companies adoption has increased from 55% in 2023 three months.
potential. To help guide organisations on their aim to achieve with AI. The study shows to 75% in 2024, with the highest ROI
AI transformation, we commissioned a study that 92% of AI users are leveraging AI seen in the financial services industry –  ooking ahead: Skilling remains a top
L
with IDC to identify some of the top trends for productivity, with 43% reporting the followed by media and telecom, mobile challenge. A lack of specialised AI skills
among enterprise leaders when it comes to the greatest ROI from these use cases. retail and consumer packaged goods, remains a top challenge, with 30%
AI platform shift. What’s generating value? What energy, manufacturing, healthcare and of respondents indicating a shortage
does ROI look like across regions and industries?  C
 ompanies are gravitating to more education industries. of in-house AI expertise.
What’s holding organisations back? advanced AI solutions. Companies
are moving towards more advanced AI AI leaders are seeing greater returns
 Explore the key findings of the IDC study and
Here are the study results and the top five trends solutions tailored to industry needs and and accelerated innovation. According be inspired by over 300 examples of how
we uncovered. I invite you to read the full study10 business processes, including custom to the study participants, companies using organisations are embracing Microsoft AI.
for a complete view of the research insights. copilots and AI agents. generative AI are averaging a 370% ROI,
with top leaders achieving a 1000% ROI.
 Chapter 01: 15
The current state of generative AI

SECTION 1.4 HIGHLIGHTS

The winning formula Leaders’ AI investments are paying off

and the difference between leaders and
Perhaps the most striking aspect of ‘AI-rich’ company differentiators laggards is stark
the current state of generative AI is AI leaders realise a higher ROI, a greater number of organisational impact areas and
The increasing availability and accessibility of
the difference between customers in monetisation avenues from AI initiatives and are more prepared to unlock AI capabilities.14
AI solutions has given rise to a new breed of
the exploring stage and those who organisation that uses AI across the business
have fully embraced the AI platform to streamline processes, create new products
shift and are consistently realising and reimagine the customer experience. These
value. Today, the highest-performing ‘AI-rich’ companies are based on an Emerald
What would you estimate your organisation’s ROI is for every USD 1 spent of AI projects or initiatives?
organisations realise almost four times Research Group study12, 13 which surveyed 49
(percentage of respondents)
the value from their AI investments IT, developer and business decision makers
who have deployed one or more custom-built
as those just getting started. They’re Leaders Laggards
generative AI solutions.
also monetising more use cases and 4.9× average ROI 2.9× average ROI

implementing generative AI projects 6%

1× return on investment 1%
in a fraction of the time.11
2× return on investment 7% 38%
Although every organisation is
3× return on investment 21% 37%
different, looking at how AI leaders
operate offers a view into the way all 4× return on investment 28% 14%
of us will work in the not-so-distant
5× return on investment 24% 1%
future.
More than 5× return on investment 18% 0%

No return on investment 0% 1%

Source: IDC InfoBrief: sponsored by Microsoft, 2024 Business Opportunity of AI, IDC# US52699124, November 2024.
 Chapter 01: 16
The current state of generative AI

HIGHLIGHTS

Practical impacts of generative AI

A study15 of over 6,000 employees using • People collaborate more with AI and each An AI leader is one whose They understand the key considerations
Microsoft 365 Copilot in their day-to-day work other. As employees become comfortable with AI development.
revealed some surprising changes in the work using generative AI, they create more enterprise-wide AI strategy is
The business user point of view in the
dynamic. documents with AI assistance, co-edit aligned to business goals and development process is critical to success.
documents with colleagues and comment
• Employees are becoming liberated from email. on content more frequently. whose reimagined business Specifically, they know that understanding the
With the introduction of Copilot, employees models repeatedly create user’s role, responsibilities and needs will help
drive success.
use email summaries to reduce the time spent
keeping up with their inboxes.
The AI Data Drop provides ongoing insights
into the practical applications and benefits of
business value.”16
generative AI adoption. Keep up with the latest They know data is a critical success factor.
• Meetings focus on value creation. Using information from Microsoft WorkLab’s AI at Work. While most organisations know their data is not
Copilot tends to focus meetings on What do AI-rich companies do differently? in perfect shape, they start by taking steps to
collaborative work and reduce the overall time get the data in order to achieve their goals.
employees spend in meetings.
They know early successes drives investment.
These organisations jump in, build with what They think about security and risk.
they have and prioritise productivity gains for While thinking about security and compliance,
Where work is headed internal users. these organisations set up generative AI-specific
See real-world company data on how Copilot is starting to reshape the workday policies and guardrails reforming their current
They bring people on the AI journey. framework.
They get buy-in from leadership to secure
funding. They know user engagement and They’re prepared to show value quickly.
Time reading email Time in meetings Word file co-creation
adoption are cross-functional efforts. They The sooner they can demonstrate impact, if not

-31 % -16 % 58 % also know technical training and staffing

are important.
ROI, the more likely they can free up the budget.

Copilot helped employees At one consulting company, People using Copilot at a law
at a consumer goods company people with access to Copilot firm produced more content.
spend less time reading email. spent less time in meetings.
 Chapter 01: 17
The current state of generative AI

LEADERSHIP PERSPECTIVES

The future of work is AI-powered

By Jared Spataro, Chief Marketing Officer, AI at Work

AI is revolutionising how we spend our time, For example, when using AI like Copilot to create right context, clarify instructions and evaluate AI will level the playing field, enabling everyone
reshaping workflows and daily patterns. A key a product pitch deck, a single prompt won’t outputs, redirecting as necessary to achieve to be data analysts, designers or editors and to
emerging trend is the shift from using AI as a suffice. The best results come from co-creating the best results. delegate tasks to AI. This co-creation with AI and
tool to collaborating with it as a colleague. with AI: providing details, pointing to relevant human colleagues will amplify team creativity
files and refining the output through multiple As this co-creation model takes hold, even and productivity, making work more strategic,
From command to co-creation. Traditionally, iterations. This process transforms every step junior employees will manage multiple AI agents, meaningful and inspiring.
computing followed a command-and-control of deck creation, from formatting to illustration, delegating routine tasks and leveraging AI’s
model: humans input instructions, and into a collaborative effort. expertise to generate value more efficiently. Read more of my top-of-mind insights and
computers output results. This was true for actionable guidance in my AI at Work newsletter.
calculators and early internet searches. However, A new collaboration pattern. This new The next generation. This transformation
with AI’s introduction, this paradigm is changing. collaboration pattern requires learning new skills offers tremendous opportunities for all career
Initially, AI interactions were based on ‘prompt and unlearning old habits. Moving beyond the stages. ‘Generation AI’ graduates will enter the
and response’ – input the right words and get ‘calculator approach’ means treating AI like workforce with a native understanding of AI
a result. Now, effective AI collaboration is more a human colleague. Effective collaboration collaboration, freeing up time for higher-order
iterative and conversational. involves setting expectations, reviewing tasks. Experienced employees can use their
work and providing feedback – similar to management skills to scale their impact with AI.
managing an employee. You need to give the
 Chapter 01: 18
The current state of generative AI

SECTION 1.5

Building trust with generative AI

With widespread adoption and public Perceptions of generative AI scepticism about the technology and more Takeaway: Users are more willing to use
because they’re still experimenting with generative AI tools when they trust their value
acceptance of generative AI, trust is
Earning trust starts with understanding that how to integrate AI into their workflows. and outputs. By building trust, organisations
more important than ever. Through
any organisation has a variety of perceptions of can encourage more sceptical users to engage
years of providing AI-optimised generative AI. Customers who deploy generative  xplorers are more familiar with AI and
E with generative AI applications and experience
infrastructure, cloud services and AI tools effectively recognise the value of report engaging with AI tools once a week their benefits.
products to customers at scale, addressing each of these perceptions to drive or a few times a month. As individuals
Microsoft has gained unique insight usage and realise value. move into exploration, they see more
into how trust is created and how productivity gains, encouraging them
to develop generative AI tools that Microsoft has identified four types of users to experiment more actively.
that successful generative AI deployments
maintain it. must satisfy, each with different needs and Power users are the most proficient and
approaches.17 enthusiastic adopters, using AI extensively.
They report significant productivity
 ceptics are individuals who have some
S benefits and are more likely to feel
familiarity with AI, but rarely use it. They motivated, creative and empowered by
may engage with AI tools occasionally, but AI tools. They’re moving beyond just using
aren’t convinced of their value or relevance. AI to streamline tasks and beginning
In surveys, these individuals are likely to to reimagine business processes and
report the least time savings from AI tools. workflows altogether.

 ovices are somewhat familiar with AI and

N
use it infrequently. They also report minimal
productivity benefits, but this is less due to
HIGHLIGHTS Chapter 01: 19
The current state of generative AI

AI for good

Whether it’s discovering ways to make communities

more resilient or supporting the fight against infectious
disease, AI is creating new opportunities to confront
the world’s most persistent and complex challenges.
At the Microsoft AI for Good Lab, our efforts to extend
fundamental rights, earn trust, expand opportunity and
Why trustworthy AI matters Security requires protecting AI systems, Safety includes security and privacy and
advance sustainability are driving meaningful change
data and users from threats, misuse and ensures that AI systems operate reliably and
around the world.
The AI platform shift opens extraordinary vulnerabilities. It addresses the risks posed by predictably across a range of environments and
Learn more about the AI for Good Lab. possibilities, but it can also involve certain threat actors using AI to make their attacks conditions. It addresses the risk of generative
risks. According to an ISM study,18 these risks more persistent and sophisticated and by AI applications producing content that’s
range from data leakage, privacy, biases and configuration challenges in complex cloud inconsistent, unreliable, harmful or culturally
backscatters. And depending on how you environments. Under this pillar, applications biased. Rooted in the principles of the Microsoft
are using AI within your organisation can also should be secure by design, secure by default Responsible AI Standard, this pillar also involves
determine the level of risk. For instance, using and subject to continuous monitoring and upholding the values of fairness, reliability,
generative AI to create an internal email is very improvement to secure operations. inclusiveness, transparency and accountability.
different than using it to create sales forecasts
Kenya Red Cross expands access or generating customer loans. Learn more about our commitments to security Learn more about our commitments to safety
to mental health support with Azure in the eBook Grow Your Business with AI You and responsible AI.
At Microsoft, it’s become clear that success Can Trust.
AI-powered chatbot
depends on the trustworthiness of the tools
The Kenya Red Cross Society faced a significant customers rely on. That’s the foundation of Privacy focuses on protecting the data
challenge in providing mental health support Trustworthy AI, a framework for developing and organisations collect and use throughout
due to a shortage of mental health professionals deploying generative AI applications. the application life cycle. It addresses the
and services, making care inaccessible for many risk of AI applications exposing sensitive or In some cases, AI isn’t just
individuals. To address this issue, the Kenya Red Cross Trustworthy AI comprises three pillars: proprietary data, which could violate regulatory a solution. It’s the only solution
used Microsoft Azure AI to develop an AI-powered requirements and severely undermine trust.
chatbot named Chat Care. This innovative solution
security, privacy and safety.
we have.”
Under this pillar, organisations must maintain
aims to bridge the gap in mental health care Juan Lavista Ferres, CVP and Chief Data Scientist,
control of their data, know where it resides and
availability and accessibility, particularly in regions Microsoft AI for Good Lab
with limited resources. ensure it remains secure at rest and in transit.

Read the story Learn more about our commitments to privacy

in the Microsoft Trust Centre.
 Chapter 01: 20
The current state of generative AI

LEADERSHIP PERSPECTIVES

Microsoft’s commitment to supporting

customers on their AI Transformation journey
By Judson Althoff, Executive Vice President and Chief Commercial Officer

The AI opportunity. The pace of AI innovation  nriching employee experiences to

E Bending the curve on innovation products by design, by default and
happening today is incredible. At Microsoft, significantly boost employee productivity by empowering your R&D, product and within our own operations as part of our
we remain focused on helping organisations and well-being by automating mundane engineering teams with deeper knowledge Secure Future Initiative.
around the world embrace AI to drive meaningful tasks and providing access to personalised and insights to create products faster,
business outcomes. By integrating the most insights and critical data. design them more effectively and get them There has never been a more opportune moment
advanced AI technology into every product to market quicker. to fulfil our mission of empowering every person
we create – along with our extensive partner Reinventing customer engagement and every organisation on the planet to achieve
ecosystem and co-innovation approach – by transforming customer experiences Our commitment. We believe that we are at our more. To find out more about how our customers
we are helping our customers accelerate AI with enhanced self-service options and best when we serve others and are committed are accelerating AI Transformation, you can
Transformation to make progress in ways that more personalised engagements. to helping our customers and partners unlock AI read my latest earnings blog.
matter most to people, businesses opportunity. Our approach to put a Copilot on
and industries.  eshaping business processes by
R every desk for every role, develop differentiated
automating workflows and optimising AI solutions and build a strong cybersecurity
Accelerating AI Transformation. We supply chains with differentiated foundation is helping organisations harness
believe there are four areas of opportunity AI solutions. AI responsibly, securely and with purpose. We
for organisations to empower their AI prioritise security above all else – securing our
Transformation and bring pragmatic
innovation to life:
 21

Chapter 02
The future potential
of generative AI
 22

4×
During the platform shifts of recent The AI platform shift promises to have an even
decades, technological breakthroughs greater impact. With capabilities doubling
at a rate four times that of historical progress,19
have led to profound changes in how
the effects of AI-powered automation, scientific
organisations operate. Client-server discovery and innovation will also accelerate.20
architecture laid the groundwork As functional and industry tools become
for decentralised computing. The more common, generative AI will revolutionise
internet globalised information, operations, enable new and disruptive business
communication and commerce. models and reshape the competitive landscape.
Mobile/cloud gave birth to our
always-on, always-connected world
and democratised access to scalable,
cost-efficient infrastructure.
AI capabilities doubling at a rate four
times that of historical progress19
 Chapter 02: 23
The future potential of generative AI

SECTION 2.1

Who’s driving innovation

For business leaders guiding their they’re focused on using AI to support tasks a high impact. Massive access to data provides AI-powered tools for every stage of the product
that require human judgement and creativity, a rich environment for training and optimising life cycle, from design and testing to just-in-time
organisations through this shift, the
maximising the value and impact of human skill. their models. production and fulfilment.24
possibilities can seem both inspiring
and daunting. By looking at the They’re redefining management. As they Like start-ups, they’re creating applications that They’re democratising industry expertise. By
organisations leading innovation build AI-powered systems to support uniquely will fundamentally shift the way businesses integrating their deep industry expertise into
today, it’s possible to understand human effort, they’re creating teams that include operate. The difference is that, with greater generative AI, software development companies
what to expect tomorrow – and both humans and generative AI. AI tools are resources and deep relationships with existing help make advanced, industry-specific AI
how to plan for it. elevated on these teams and become essential enterprise customers, they’re positioned solutions more accessible and effective for
collaborators with their human colleagues. to impact industries on a massive scale. a broader range of businesses. This ensures that
Start-ups: A view of the future the AI solutions are not only technically robust,
They’re reimagining industries. Generative AI How are they shaping the future? but also highly relevant to the specific challenges
AI-native start-ups provide unique insights allows smaller teams to develop applications and needs of each industry.25
into what’s on the horizon.21 These businesses with greater agility, enabling AI-native start-ups They’re empowering teams to do more.
are cloud native – with AI at the core of their to move rapidly from a big idea to producing In industries like healthcare where chronic talent
DNA. Rather than evolving operations to take and scaling a solution. In virtually every industry, shortages impact organisations’ ability to deliver
advantage of emerging technology, they’re these innovation-based businesses are poised to optimal outcomes, software development
creating operational models that are ahead of have a significant and disruptive impact.22 companies are using generative AI to provide
their time – and imagining where AI will take us insights and administrative tools that automate
in the coming years. Software development companies: time-consuming manual work. This enables
Change at scale teams to focus more on tasks requiring their
How are they shaping the future? unique skill set.23
Software development companies possess
They’re focused on what humans do best. deep domain knowledge earned through years They’re shortening product cycles. In industries
These start-ups aren’t concerned about of understanding and meeting their customers’ like manufacturing, software development
generative AI replacing human effort. Instead, needs. This gives them foundational expertise to companies are helping organisations accelerate
develop generative AI applications that will have product development with generative
 Chapter 02: 24
HIGHLIGHTS The future potential of generative AI

See Microsoft Research in action Co-innovation labs: Accelerating impact They’re demonstrating ROI. For some They’re addressing global issues. Much of the
organisations, challenges in predicting work undertaken by AI research organisations
Multimodal models move beyond text and numbers, Co-innovation labs help companies develop or measuring ROI are a barrier to progress, even focuses on using technology to solve
incorporating a broad spectrum of inputs and and refine generative AI solutions. They provide when the outcomes of a generative AI project global challenges.30 Aurora, for example,
advancing generative AI. support throughout the development process, are valued. By relying on a co-innovation lab’s is a foundational model of the earth’s
offering expertise and resources to overcome expertise, organisations can manage generative atmosphere that can perform the complex
Watch the video
challenges and bring innovative products AI projects more effectively to create tools that computations necessary to generate a weather
to market. cost less than the business value they generate.28 forecast 5,000 times faster than traditional
techniques – achieving more accurate results
Healthcare organisations use AI to drive innovation How are they shaping the future? Research organisations: than the current state-of-the-art forecasting
and efficiency, paving the way for AI transformation Exploring what’s possible tool 94% of the time.
across many industries.
They’re identifying best practices. With
co-innovation labs, technical experts establish As generative AI evolves as a powerful force for They’re perfecting human-AI interaction.
Watch the video
highly specific, in-depth partnerships with change, research organisations operate far Generative AI can empower productivity, reduce
organisations tackling challenging problems ahead of what’s commercially available, stress, increase efficiency and unlock exciting
tailored to their specific situation. By exploring the frontiers of what the technology opportunities. But sometimes, it makes mistakes.
Data is critical to successful AI deployment and collaborating with different companies can accomplish. The innovations they As research organisations study how humans
an essential driver of business decisions. Maximising to develop unique solutions, they can identify develop expand the scope of what’s possible, interact with generative AI, they provide insights
the value of data helps business leaders prepare making them invaluable thought leaders for and guidance to help organisations establish
common patterns that lead to valuable insights
for future innovation.
and best practices.26 organisations seeking to maximise the impact policies and safeguards.31 This helps users stay
Watch the video of their generative AI investments. vigilant about catching mistakes while remaining
They’re improving customer experiences. motivated to use AI tools.
Generative AI can radically simplify access How are they shaping the future?
to an organisation’s products or services. But
assembling the infrastructure and services They’re innovating with models. Research
necessary to optimise outcomes and make organisations are working to develop models
the process easy for users can be challenging. that deliver better results, reduce latency
Co-innovation labs are helping organisations and are optimised to specific use cases.29 For
dramatically shorten the timeframe required to example, as of September 2024, Phi-3-Vision
deliver extraordinary customer experiences.27 model outperforms all other open-source vision
models, and its significantly smaller size allows
it to operate more efficiently.
 Chapter 02: 25
The future potential of generative AI

LEADERSHIP PERSPECTIVES

How software development companies are

paving the way for AI transformation
By Jason Graefe, Corporate Vice President, ISV and Digital Natives

Empowerment through AI them as the companies to lead the future of The power of curiosity in AI-powered helped to serve customers better, reduce costs
software development. The critical differentiator software development and improve revenues. Companies that prioritise
AI has become a transformative force in various lies in their speed of adoption and response AI integration are better positioned to lead their
industries, and software development is no to this technological shift. In the context of AI, curiosity is indispensable, markets, offering cutting-edge solutions that
exception. What this evident shift indicates paving the way for breakthrough solutions. meet dynamic client needs and inspiring new
is that every company, regardless of their make, To effectively embrace AI and evolve into Software development companies are asking the behaviour patterns.
is integrating software into their DNA, some a software development company, it is essential pivotal question that every aspiring organisation
faster than others. This evolution is not just to grasp the distinct challenges and benefits. can learn from: What is now possible that wasn’t Conclusion
a trend, but a transformative journey that holds Established companies overcome bureaucratic before? A curious mindset unlocks the true
immense potential for growth. hurdles and product legacy, using their extensive potential of AI by making us more open to create The state of AI in software development
resources and infrastructure, allowing them to more scalable and intelligent processes – and companies is marked by a new beginning for
A new dawn in software development drive innovation at large. Meanwhile, smaller to improve accuracy and enhance the reliability all players. The ability to respond swiftly to AI’s
companies, though confronted with limited of software. potential, coupled with a culture of curiosity and
AI democratises opportunities, enabling capital and scaling constraints, can benefit from continuous learning, will define success in this
companies of all sizes to harness its power. greater agility, facilitating a significant edge in Championing growth and innovation transformative era. As companies embrace AI,
Whether they’re emerging start-ups or experimentation and innovation. Both types must they must recognise and support this evolution,
established independent software vendors move quickly to experiment, test and embrace In today’s fast-paced software development seizing the opportunities to shape the future
(ISVs), companies that integrate AI into their this new technological capability. landscape, AI adoption is not merely of software development together
products encounter significant opportunities for advantageous – it’s imperative. The significant
transformation and advancement, positioning advances in AI adoption in the healthcare, Learn more about the trends shaping the future
finance, retail and manufacturing industries have of software development in my blog post,
Unveiling the future of AI innovation for ISVs.
 Chapter 02: 26
The future potential of generative AI

SECTION 2.2

Partnering together to build trustworthy AI

HIGHLIGHTS

The emergence of AI agents

Employee productivity is the top business goal for
companies using AI. For example, tools like Copilot can
At Microsoft, we believe the Three trusted platforms for the AI age class of PCs built to be the fastest on the be used to summarise documents, take meeting notes
market with all-day battery life. Another and prioritise inboxes – becoming more personalised
transformative power of generative over time.
Microsoft has built three platforms that help example includes a new secure, admin-
AI has a significant role in driving
companies maximise their opportunity and serve less and passwordless device designed
innovation and efficiency. It’s now as the building blocks for any AI transformation: for cloud productivity, ensuring no data
AI agents, on the other hand, go beyond assistance
by working alongside or on behalf of users, handling
the key driver for companies to boost is stored on the device. tasks from simple queries to complex assignments
their productivity. Embracing this Copilot. A platform designed to serve and can be tailored for specific expertise. This
capability boosts productivity across various industries
technology swiftly and integrating as the user interface for AI. This platform Security above all else. Microsoft is
by automating routine tasks, allowing employees
it deeply into daily operations are becomes an organising layer for work prioritising security through its Secure to focus on more strategic activities and saving time
crucial strategies for businesses and how it gets done. It includes tools Future Initiative, focusing on the principles and money.
for creating agents that automate and of secure by design, secure by default
to stay competitive and lead AI agents are gaining more attention due to advances
transform business processes. and secure operations. This includes
in today’s market. continuous improvement and collaboration
in large language models (LLMs), which make AI tools
more useful by allowing broader communication with
Copilot and AI stack. A platform that with the security community through Zero AI. These agents, which act as layers on top of LLMs,
Trust is at the core of any AI exposes every layer of the tech stack Day Quest, a new hacking event with observe, collect information and generate action
innovation. As organisations approach to build custom AI solutions. This platform USD 4 million in rewards aimed plans, then communicate with the user or even act
on their own, if permitted.
this platform shift, it’s important to empowers users to build ambitious at securing cloud and AI.
carefully consider responsible AI products by leveraging advanced In the future, organisations will have numerous
practices as well as prioritise security technology and providing a unified Read Microsoft Chairman and CEO Satya AI agents working behind the scenes to enhance
experience for designing, customising Nadella’s perspective: Delivering three trusted processes and create new ones. These agents will
throughout your organisation. Learn assist in sales by researching leads and guiding
and managing AI applications and agents. platforms for the AI age.
more about trustworthy AI in the customer outreach, minimise supply chain disruptions
eBook Grow your business with AI Copilot devices. A platform for a new
by tracking performance and recommending
adjustments and overall, reduce costs, improve
you can trust.32 class of devices designed to use decision-making and foster continuous improvement.
distributed AI across cloud and edge.
For example, Copilot devices include a new Get started today by creating your own agent in
Microsoft 365 Copilot, and learn how pre-built agents
can support common business workflows.
 Chapter 02: 27
The future potential of generative AI

LEADERSHIP PERSPECTIVES

How to stay ahead of emerging challenges

and cyberthreats
By Vasu Jakkal, Corporate Vice President, Microsoft Security Business

With the evolving cyberthreat landscape shaped According to a recent study by Microsoft To mitigate these risks, we encourage companies security solutions provide expertise on Zero
by global events and AI advancements, security conducted in November 2024, IT administrators, to enforce role-based data permissions, use Trust for AI along with governance and
is more crucial than ever. Security professionals on average, were 29.79% faster and 34.53% more governance solutions to track AI interactions compliance solutions.
face an overwhelming volume of alerts and accurate when tested on common IT admin and incidents and apply data life cycle
complex attacks, straining their resources. The scenarios using Copilot.33, 34 management. Ambiguous guidelines and global AI safety. When using generative AI, follow
global shortage of 4.7 million skilled security differences pose AI regulation challenges, responsible principles by validating outputs,
professionals further exacerbates this issue, One of our customers, QNET, also noticed an so we recommend establishing ethics ensuring transparency and incorporating ethical
stretching teams thin as they deploy AI safely. increase in security response efficiency by 60% committees and using frameworks like NIST’s AI practices. Evaluate systems for ungrounded
Moving forward with AI, it’s essential to adopt using Microsoft security solutions. Risk Management Framework. responses and monitor for false positives and
an end-to-end approach to security, focusing negatives. Human oversight and feedback
on three key areas to minimise risk and maximise Security for AI. A recent survey by ISM35 As organisations gear up for digital are crucial for correcting errors. AI red teaming
benefits. highlights executives’ concerns about integrating transformation, they should likewise prepare is essential for assessing accuracy, identifying
generative AI, focusing on data security, for AI transformation by assessing their vulnerabilities and minimising reliance
AI for security. AI for security empowers teams governance and regulatory compliance. Key security infrastructure, implementing Zero on inaccurate outputs.
to protect at the speed and scale of AI by turning issues include data leakage, over-permissioned Trust principles, identifying skill gaps and
global threat intelligence, best practices and an data and internal sharing. collaborating with AI providers. Microsoft Azure Learn more about AI in security, and how
organisations’ security data into tailored insights offers tools and guidance for the responsible to secure your AI, minimise risk and reap
to improve their security posture, strengthen management of AI systems, while Microsoft the benefits of AI.
team expertise and outpace adversaries.
 Chapter 02: 28
The future potential of generative AI
Simulating decarbonisation SECTION 2.3

opportunities

A global leader in manufacturing, Shaeffler employs

data and AI solutions to quickly identify which
decarbonisation measures will have the biggest
AI and sustainability
impact on their emissions reduction.

Read the story

AI stands as the latest – and Microsoft’s commitment to sustainability See left panel for more examples of how
leading companies are innovating with AI
potentially most powerful – general-
At Microsoft, our sustainability journey started to accelerate sustainability
purpose technology (GPT), offering an
with getting our own house in order. In 2020,
unprecedented opportunity to drive Advancing the sustainability of AI
Reducing food waste and helping we set ambitious targets to be a carbon-
toward societal transformations we negative, water-positive, zero-waste company
customers eat healthy urgently need to achieve the world’s that protects ecosystems by 2030. We’re In alignment with our commitment to
As the Netherlands’ leading supermarket chain,
sustainability goals. AI can significantly committed to sharing our progress, challenges trustworthy AI and our ambitious sustainability
Albert Heijn uses generative AI for customer enhance the development and and learnings through our annual Environmental goals, Microsoft is determined to tackle the
personalisation to food waste projects, helping deployment of sustainability solutions, Sustainability Report. challenge of sustainability so the world can
customers live a healthier lifestyle while minimising harness the full benefits of generative AI.37
making them faster, cheaper and
its environmental footprint. We also recognise that our actions alone will not There are three areas where we’re deeply
more effective. solve the climate crisis. As a global technology invested and increasing our focus. The first is
Read the story provider, we believe we have a role to play in optimising data centre energy, water, waste and
Generative AI is already helping to create a more
supporting the thousands of customers and ecosystem efficiency. The second is advancing
sustainable future with tools that help integrate
partners who look to Microsoft to provide the low-carbon materials, creating global markets
renewable energy into grids, develop energy
innovation and solutions to support them on to help advance sustainability across industries.
Planting a seed for climate storage solutions, reduce food waste and
their own sustainability journeys. The third is improving the energy efficiency of
neutrality support socially and environmentally beneficial
generative AI and cloud services, empowering
actions.36 As sustainability leaders increase their
Microsoft Cloud for Sustainability and our customers and partners with tools for
An association of 52,000 forest owners in Sweden, access to data and AI capabilities progress, we’re
sustainability data solutions in Microsoft Fabric collective progress.
Södra employs data and AI capabilities to collect and seeing an emergence of innovative solutions to
analyse sustainability data for every part of their value
offer data and AI capabilities to support the use
social, environmental and economic challenges.
chain, helping them reveal insights, meet reporting of environmental, social and governance (ESG) Learn more in our Sustainable by Design
requirements and advance their sustainability goals. data intelligence to make informed decisions, blog series, starting with Advancing the
create new efficiencies and look for new sustainability of AI.
Read the story growth opportunities.
 Chapter 02: 29
The future potential of generative AI

Companies accelerating sustainability projects with AI

Reducing water loss from leaks in our Optimising grids to take Accelerating the search for less toxic
municipal water systems on more renewables battery materials
Approximately 30% of the world’s piped water is Researchers at PNNL and Microsoft collaborated to
LineVision is using AI to enhance the capacity
lost from leaks and theft. discover a potential new battery material in weeks
of transmission lines, helping utilities get more
versus years with AI and high-performance computing.
renewables on the grid.
Microsoft partner FIDO Tech uses AI-enabled acoustic
analysis to seek out and identify leaks, so technicians The new material may reduce the amount of lithium
UK’s National Grid used this technology to increase
can identify big leaks and fix them faster. used in a battery by up to 70%.
capacity by 60% and add an additional 600 MW
of offshore wind capacity.
Read the story Read the story

Read the story

Learn more about how AI is helping advance sustainability today in the playbook Accelerating sustainability with AI.
 Chapter 02: 30
The future potential of generative AI

SECTION 2.4

Innovators on the forefront of AI Making cultural heritage more

accessible with AI

St. Peter’s Basilica, a historic and architectural marvel,

Today, Microsoft customers on the What generative AI means for the future Unlimited business models. Beyond the short-
term gains achieved by improving efficiency, used advanced photogrammetry and AI to create a 3D
leading edge of the AI platform shift replica of St. Peter’s Basilica using over 400,000 high-
Unprecedented efficiency. As decisions become increasing productivity and extracting value
are opening a window to an inspiring, resolution images. The data was processed and analysed
more data-driven – and as data-driven insights from data, generative AI is a platform for
hopeful world. Across industries, become available to non-technical employees innovation. By removing barriers, simplifying
using AI tech in the Azure AI Foundry. The project
resulted in an ultra-precise digital twin of the Basilica,
they’re demonstrating what’s possible at all levels – organisations will experience application development and shortening allowing visitors to explore every intricate detail from
now and creating what will be unprecedented gains in productivity and development cycles, generative AI is putting anywhere in the world, making it possible to preserve
possible in the future. Whether it’s efficiency. This will be more than an evolution; it the future squarely in the hands of innovators. and share its cultural heritage in an unprecedented way.

medical and scientific breakthroughs will be a system reset, enabling a fundamentally This will lead to the disruption and reinvention
of business models at a scale that exceeds the Read the story
or advances in transportation and different mode of operations. As organisations
embrace more potent generative AI tools, they platform shifts of the past.
communications, these innovators are
will accomplish initiatives that might once have
giving every business leader a reason seemed impossible, from dramatically reducing Unimaginable growth and opportunity. Using AI to personalise teaching and
to dream about what generative AI environmental impacts to creating frictionless As generative AI drives a tectonic shift in help make learning fun
could mean to their organisation. customer experiences. operations, it’s creating opportunities for
organisations that embrace the promise of Khan Academy saw the opportunity to enhance education
AI to achieve once-in-a-lifetime competitive with AI solutions and tools, introducing Khanmigo for
Teachers, an AI-powered teaching assistant, developed
advantages. As it changes the nature of work,
using Azure AI and Azure OpenAI Service. This initiative
it frees individuals to focus on high-level,
led to significant time savings for teachers, more effective
rewarding challenges, enabling them to achieve learning experiences for students and the development
more. And as it breaks down longstanding of scalable and adaptable AI tutoring capabilities.
barriers to entering the digital workforce, it
invites historically underrepresented groups to Read the story
participate more fully in the next phase of global
progress.
 31

Chapter 03
Advancing generative
AI in your organisation
 Chapter 03: 32
Advancing generative AI in your organisation

SECTION 3.1

In Chapters 1 and 2, we explored the current state of generative AI and potential

impacts the AI platform shift will have on the future. This chapter highlights key
elements of approaching generative AI adoption strategically, based on your
organisation’s resources, goals and culture, to maximise the value you realise.

Best practices for implementing AI at scale  stablish AI readiness and governance.

E
Implement AI policies, skill development,
There are a few overarching steps every governance and adoption plans. Appoint
organisation will need to accomplish on its leaders (for example, a Chief AI Officer)
journey to using AI at scale. and manage AI operations at scale (use
case approval, technical path, buy/build
Define your AI strategy and direction. considerations, responsible AI policies).
Establish the vision for AI within the
organisation and the strategy for its Deploy AI to scale. This includes
implementation, as well as ensuring understanding AI user interfaces,
alignment among the C-Suite executives. applications, supporting infrastructure,
Read Judson Althoff’s perspective on the data integration and security. It also entails
four areas of AI opportunity. building environments for AI to scale
effectively as use cases progress towards
Target a top business challenge self-service.
or opportunity. Once you identify
a significant business challenge or Expand AI transformation. Scale
opportunity, implement a solution to operations, deliver ROI and identify
achieve ROI and demonstrate value, key use cases for reuse. Promote broad
ensuring it’s measurable with KPIs. adoption through internal and external
communications, measure impact and
report back to the C-suite to ensure
alignment with the vision or make
adjustments as necessary.
 Chapter 03: 33
HIGHLIGHTS SECTION 3.2
Advancing generative AI in your organisation

The AI Centre of Excellence

Establishing an AI Centre of Excellence (AI CoE)

Advancing generative AI with the five
drivers of value
– a dedicated, cross-functional team tasked with
driving AI initiatives – is an important early step
in a well-structured AI transformation. An AI CoE
will significantly enhance the acceleration and
success rate of AI projects within organisations –
by clearly defining AI commitments, centralising In Chapter 1, we discussed the five stages of generative AI readiness (exploring, planning, implementing, scaling and realising)
resources and expertise, fostering a culture of and how the five drivers of AI value (business strategy, technology and data strategy, AI strategy and experience, organisation
innovation and continuous improvement and and culture and AI governance) can be used to accelerate your readiness journey and overcome the challenges of generative
managing the organisation’s overall adoption
strategy. An AI CoE helps ensure organisations are
AI adoption. These charts provide a quick reference for the strategic priorities that are most important to focus on as you
focused on a broad AI transformation strategy, progress through the stages of readiness. For a more detailed examination of the five drivers, read The AI Strategy Roadmap:
whether it’s prioritising use cases or driving Navigating the Stages of AI Value Creation.38
widespread adoption to measuring value –
building the foundation for ongoing AI success.

Business strategy

Keys to an effective AI Centre Exploring → Planning Planning → Implementing Implementing → Scaling Scaling → Realising

of Excellence • Discuss how generative AI aligns with • Review and prioritise the generative • Discuss the results of AI PoCs, • Identify areas where generative AI
your organisation’s efficiency, growth AI use cases identified in the including key metrics like ROI, has delivered significant value and
and innovation goals. planning stage. efficiency gains and customer impact. improvements are needed.
Clear mandates: It owns progress across the five
drivers of AI value. • Set clear objectives for generative • Align use cases with specific business • Ensure that AI efforts continue • Evaluate how generative AI contributes
AI projects and how they’ll support goals, including real-time decision- to be closely aligned with the to broader business outcomes,
Cross-functional teams: It’s composed of leaders strategic business goals. making and customer engagement. broad strategy. including revenue growth, innovation
with diverse skillsets. and customer satisfaction.
• Identify potential generative AI use • Identify the resources necessary to • Identify new growth-oriented use
Growth-oriented mindset: It continuously monitors cases that can provide quick wins advance the highest-priority projects. cases for generative AI. • Identify additional opportunities for
and optimises performance. in terms of cost savings or operational • Define success criteria for proofs generative AI to drive new products,
efficiency. of concept (PoCs), including metrics services or business models.
for efficiency, cost savings and
customer impact.
 Chapter 03: 34
Advancing generative AI in your organisation

Technology and data strategy

96 %
Exploring → Planning Planning → Implementing Implementing → Scaling Scaling → Realising

• Review the current state of your data • Assess readiness for generative • Develop a roadmap for scaling • Optimise cloud resources to support
quality and availability. AI deployment, focusing on cloud successful PoCs, including identifying generative AI at scale.
• Identify gaps in data that need to be infrastructure and data management. blockers and challenges encountered • Assess the feasibility and impact
addressed to support generative AI • Ensure access to the right data, during implementation. of new AI capabilities as they become
initiatives. formatted correctly for generative • Enhance data access and availability. available.
• Consider the need for cloud migration AI modelling. • Assess whether current cloud and • Continually improve data quality
if not already in place. • Design the PoC for one or more data infrastructure can support scaled standards and data formatting.
use cases. AI operations.
• Plan for infrastructure upgrades
Companies that use generative AI to innovate as needed to ensure performance
and expand their offerings report significantly and security.
higher returns, with 96% seeing measurable
value once they reach the ‘Realising’ stage.39

Generative AI strategy and experience

Exploring → Planning Planning → Implementing Implementing → Scaling Scaling → Realising

• Establish diverse, cross-functional • Establish clear, repeatable • Expand the AI team to include • Ensure that standardised generative
teams in your organisation with a mix processes and workflows for additional data scientists, engineers AI processes deliver repeatable and
of technical and business expertise generative AI projects. and subject matter experts. measurable results across departments.
centred around the CoE. • Identify the best AI models for • Continually refine processes • Look for opportunities to improve
• Define the roles and responsibilities your use cases. and workflows. efficiency by refining workflows and
of key personnel involved in generative AI models.
AI projects. • Set goals for further generative
• Identify experience and knowledge AI development to maintain
gaps and develop a strategy for filling a competitive edge.
those gaps through recruitment,
training or supplier expertise.
 Chapter 03: 35
Advancing generative AI in your organisation

Appointing a Chief AI Officer (CAIO) can Organisation and culture

signal generative AI’s strategic importance.
The presence of a CAIO helps coordinate
generative AI efforts across the business, Exploring → Planning Planning → Implementing Implementing → Scaling Scaling → Realising
ensuring alignment with broader
business objectives. • Secure executive leadership • Communicate the AI vision • Communicate non-confidential • Continue to foster a generative
commitment to generative throughout your organisation to results of AI proofs of concept, AI-driven culture, ensuring
AI as a critical part of your generate awareness and support including key metrics. all departments understand
organisation’s future. for generative AI projects. • Implement training programmes how generative AI supports
• Discuss long-term investment • Identify skills gaps within to help employees integrate business strategy.
and resource allocation for AI. your organisation. generative AI into their workflows. • Focus on leadership’s role in
• Define a vision for AI in your • Set up generative AI-specific training maintaining momentum and
organisation based on strategic goals, and upskilling programs to prepare prioritising generative AI.
desired business outcomes, technical employees for implementation.
capabilities and available resources.

At the most advanced stages of generative AI

readiness, AI governance becomes a critical
Generative AI governance
enabler of trust, both internally and externally.
Organisations that excel in AI governance not
Exploring → Planning Planning → Implementing Implementing → Scaling Scaling → Realising
only mitigate risks, but also build trust with
customers, regulators and stakeholders. This • Develop controls for ensuring • Develop generative AI governance • Review AI governance frameworks • Review and refine AI governance
trust is vital for scaling generative AI solutions transparency, explainability and policies, ensuring data privacy, and update as necessary to ensure policies to manage increasingly
and ensuring long-term success. interpretability of results. security and compliance with they meet the needs of scaled complex AI systems.
• Establish security and compliance industry regulations. generative AI operations. • Establish clear performance metrics
protocols. • Create transparency and explainability • Establish compliance mechanisms for generative AI projects, focusing
• Establish generative AI system guidelines for using generative AI in to prevent bias and ensure generative on transparency, fairness and security.
accountability for bias, impacts, decision-making processes. AI deployments remain secure
safety and security. and ethical.
 Chapter 03: 36
Advancing generative AI in your organisation

SECTION 3.3

See how organisations across industries are using generative AI

BCI empowers employees with Canadian Tire Corporation Dow reimagines productivity Teladoc Health: Transforming
Microsoft 365 Copilot, realising uses Azure AI services to boost and supply chain efficiency with telehealth operations with Copilot
10%-20% productivity boost employee productivity Microsoft 365 Copilot
Teladoc Health wanted to increase operational
British Columbia Investment Management Canadian Tire Corporation (CTC) needed to As a global leader in materials science, Dow efficiency and process automation to manage
Corporation (BCI) needed to enhance enhance employee productivity and create continually focuses their efforts to enhance rapid growth and acquisitions. By integrating
employee productivity and job satisfaction unique customer experiences in a competitive cost management and operational efficiency Microsoft 365 Copilot, Teladoc Health
while maintaining strict security and regulatory retail industry. They aimed to relieve their within their supply chain. This includes facing automated routine tasks, boosted efficiency and
requirements in their investment management workforce from mundane tasks and focus challenges with billing accuracy and costly increased productivity, resulting in considerable
services. BCI adopted Microsoft 365 Copilot on higher-value customer experiences. CTC overcharges in logistics spending. By adopting time savings, faster client service response times
and the Microsoft Azure ecosystem, including utilised Microsoft Azure AI services, including Microsoft 365 Copilot, Dow empowered and streamlined onboarding processes. The
Azure OpenAI Service, to streamline operations Azure OpenAI Service and Azure Data Lake employees with AI-driven insights and outcome has been a significant improvement in
and automate manual tasks. This implementation Storage, to build digital assistants. This AI-driven automated tasks across departments. The results operational capabilities, leadership engagement
of generative AI has allowed BCI to transform its transformation has allowed CTC to redefine were significant, with Dow anticipating millions of and employee satisfaction, setting a new
operations, empowering employees to focus on productivity, efficiency and business growth, dollars in cost reduction on shipping operations standard in the telehealth industry.
strategic and creative tasks, ultimately benefiting setting new benchmarks for engaging customers within the first year of implementation, along
their clients and the communities they serve. across Canada. with time savings, streamlined processes and Read the story
greater productivity for employees.
Read the story Read the story
Read the story
 Chapter 03: 37
Advancing generative AI in your organisation

We’re all on the AI Microsoft is committed to sharing our learning

in the age of generative AI to further our
mission of empowering every person and
We hope these perspectives from Microsoft
thought leaders and insights from leading
organisations provide you with resources,

journey together
every organisation on the planet to do more. strategies and inspiration as we travel together
on this road.
As the real-world impacts of the AI platform
shift take shape, organisations of all kinds will
need to learn, adapt and refine their approaches.
Collaboration, partnership and shared learning
will be vital for everyone.

Next steps
Keep up with the latest AI trends and news: The Shift newsletter.
Assess your company’s AI readiness: AI Readiness Wizard.
Explore Microsoft AI.
 38
Endnotes:

1 
Satya Nadella, ‘Delivering 3 trusted platforms for the AI age’, LinkedIn, Microsoft, 22
J ared Spataro, ‘What Can AI-Native Start-ups Teach the Rest of Us? A Lot: This new
November 19, 2024. wave of companies can show all companies a clear path to reinvention’, WorkLab,
2 
Benj Edwards, ‘ChatGPT sets record for fastest-growing user base in history, report Microsoft, September 2024.
says: Intense demand for AI chatbot breaks records and inspires new USD 20/mo 23
E
 ric Boyd, ‘Microsoft and Epic expand AI collaboration to accelerate generative
subscription plan‘, Ars Technica, February 1, 2023. AI’s impact in healthcare, addressing the industry’s most pressing needs‘, Official
3 
IDC InfoBrief: sponsored by Microsoft, 2024 Business Opportunity of AI, IDC# Microsoft Blog, Microsoft, August 22, 2023.
US52699124, November 2024. 24
M
 ichael Boland, ‘Siemens makes a splash at Microsoft Ignite‘, Siemens Team Centre
4 
‘The AI Strategy Roadmap: Navigating the Stages of AI Value Creation‘, Microsoft, Blog, Siemens, December 17, 2023.
2024. 25
S
 atish Thomas, ‘Microsoft introduces new adapted AI models for industry‘, Official
5 
‘The AI Strategy Roadmap: Navigating the Stages of AI Value Creation‘, Microsoft, Microsoft Blog, Microsoft, November 13, 2024.
2024. 26
B ill Briggs, ‘Innovation happens here: A novel program helps Microsoft customers
6
‘AI Use Cases for Business Leaders: Realise Value with AI‘, Microsoft, 2024. create solutions they once thought unreachable‘, Microsoft Source, Microsoft,
IDC InfoBrief: sponsored by Microsoft, 2024 Business Opportunity of AI, IDC#
7 
April 3, 2024.
US52699124, November 2024. ‘How the BBC played up in education technology with Microsoft AI Co-Innovation
27 

8 
IDC InfoBrief: sponsored by Microsoft, 2024 Business Opportunity of AI, IDC# Lab‘, Microsoft Co-Innovation Lab Case Study, Microsoft, 2024.
US52699124, November 2024. ‘Nearly 98% Automated Accurate Match Rate! Microsoft Azure Accelerates
28 

9 
‘Microsoft Digital Defence Report 2024‘, Microsoft Security Insider website, Global Pharmaceutical Data Processing‘, Microsoft Co-Innovation Lab Case Study,
Microsoft, March 2024. Microsoft, 2024.
10 
IDC InfoBrief: sponsored by Microsoft, 2024 Business Opportunity of AI, IDC# 29
J ianfeng Gao, Keynote: Phi-3-Vision: A highly capable and “small” language vision
US52699124, November 2024. model Microsoft, A highly capable and ‘small‘ language vision model Microsoft
11 
IDC InfoBrief: sponsored by Microsoft, 2024 Business Opportunity of AI, IDC# Research Forum Brief, Microsoft, January 6, 2025.
US52699124, November 2024. 30
M
 egan Stanley, ‘Project Aurora: The first large-scale foundation model of the
12 
Emerald Research Group: sponsored by Microsoft, GenAI Development Journey, atmosphere‘, Microsoft Research Forum Brief, Microsoft, September 3, 2024.
November 2024. This research study was conducted by Emerald Research Group 31
M
 ihaela Vorvoreanu, ‘Fostering appropriate reliance on AI‘, Microsoft Research
(a subsidiary of the Harris Poll) from October 24-November 22, 2024, among Forum Brief, Microsoft, September 3, 2024.
twenty-nine (29) IT decision makers, fourteen (14) developer decision makers and 32
‘Grow Your Business with AI You Can Trust‘, Microsoft, 2024.
six (6) business decisions makers in enterprise organisations. Qualified participants 33
B
 ono, James and Alec Xu. “Randomised Controlled Trials for Security Copilot for
have deployed one or more GenAI custom built solutions into a production IT Administrators‘, Microsoft, January 2024. Study consisted of 182 subjects split
environment and the participants must be actively involved throughout their entire into control and treatment groups. Subjects recruited via Upwork, a marketplace
GenAI development process. Participants also had a mix of GenAI development for freelancers. Subjects offered performance incentives for speed and accuracy.
tools and were not required to have used any specific development tools. Average statistics are statistically significant 0 < p ≤ 0.01. Throughout, statistical
13 
The opinions and experiences expressed in the online interviews are not significance is indicated using asterisks, where * indicates 0.05 < p ≤ 0.1, ** indicates
representative of the entire market, and only represent those of each individual who 0.01 < p ≤ 0.05, and *** indicates 0 < p ≤ 0.01.
participated. All qualitative findings should be interpreted as directional only. 34
B
 ono, James and Alec Xu. ‘Randomised Controlled Trials for Security Copilot for IT
14 
IDC InfoBrief: sponsored by Microsoft, 2024 Business Opportunity of AI, IDC# Administrators‘, Microsoft, January 2024.
US52699124, November 2024. 35
Information Security Media Group. "First Annual Generative AI Study: Business
15 
‘ AI Data Drop: Three Key Insights from Real-World Research on AI Usage‘, Rewards vs. Security Risks‘, December 2023.
Microsoft, 2024. 36
A
 my Luers et al., ‘Accelerating Sustainability with AI: A Playbook‘, Microsoft
16 
IDC InfoBrief: sponsored by Microsoft, 2024 Business Opportunity of AI, IDC# Advance Sustainability, Microsoft, January 2025.
US52699124, November 2024. 37
M
 elanie Nakagawa, ‘Sustainable by design: Advancing the sustainability of AI‘,
17 
“2024 Work Trend Index Annual Report: AI at Work Is Here. Now Comes the Hard Official Microsoft Blog, Microsoft, April 2, 2024.
Part: Employees want AI, leaders are looking for a path forward.‘ Microsoft, May 8, ‘The AI Strategy Roadmap: Navigating the Stages of AI Value Creation‘, Ipsos,
38 

2024. commissioned by Microsoft, 2024. IPSOS created the analyses and models
18
Information Security Media Group. ‘First Annual Generative AI Study: Business described in this paper using multinomial logit analyses to predict the AI readiness
Rewards vs. Security Risks,‘ December 2023. level of each driver using the items in the survey for each driver and then the overall
19
S
 atya Nadella, ‘Delivering three trusted platforms for the AI age,‘ LinkedIn, AI readiness from the predicted assessments of the five drivers. The multinomial
Microsoft, November 19, 2024. logit analysis produces probabilities for each level of AI readiness, which allows for
20
S
 atya Nadella, ‘My annual letter: Relevance and reinvention,‘ Microsoft, October 24, identifying the differential opportunities that exist for each. To determine the next
2024. best readiness opportunity, we took the predicted probabilities and calculated the
21
J ared Spataro and Nicholas Thompson, Learn from AI-Native Firms – or Get Left expected value of AI in each case, then estimated the increase in the expectation
Behind: Lessons from the new wave of companies with AI at their core, WorkLab, from increasing each item respectively.
Microsoft, September 2024. ‘The AI Strategy Roadmap: Navigating the Stages of AI Value Creation‘, Ipsos,
39 

commissioned by Microsoft, 2024.

© 2025 Microsoft Corporation. All rights reserved. This document is provided ’as is’. Information and views expressed in this document,
including URL and other Internet website references, may change without notice. You bear the risk of using it. This document does not
provide you with any legal rights to any intellectual property in any Microsoft product. You may copy and use this document for your internal,
reference purposes...