You are on page 1of 2

Deploying AI to Produce

Real-World Business Value Written by Aleksandar Velkoski, Ph. D.

xecutives are increasingly focused on en- CROSS-FUNCTIONAL COLLABORATION. cross-functional collaboration


hancing artificial intelligence (AI) adop- Start the journey by conven- to accelerate AI adoption. The
tion within their organizations. Accord- ing a cross-functional team. It may Data Science Group (DSG) and
ing to the 2020 Gartner CIO Survey, AI seem like a trivial starting point, the Marketing, Communications
ranked number one on CIOs’ list of top 10 game but for AI innovation to succeed, and Events (MCE) Group recently
changing technologies. Data analytics, which is teams must come together from worked together to identify a
typically discussed within the context of core AI and the beginning and work to break- high-impact business opportunity
supporting technologies, ranked second on the list. down silos if silos exist. In fact, related to increasing member en-
Although increased AI adoption is central to the most digitally advanced com- gagement. To ensure that the pro-
modern organization strategy, organizations con- panies — those successfully de- cess went smoothly, and that dis-
tinue to fail to deliver business value from AI. Of ploying digital technologies and parate needs were met, we openly
the biggest challenges to AI adoption, a lack of use capabilities to improve processes, and honestly discussed everything
cases, as well as an inability to measure or value engage talent across the organi- from technical capabilities and
success, continues to permeate. In addition, there zation and drive new value-gener- limitations to business unit goals
is often misalignment among technologists who ating business models — are far and objectives.
are responsible for AI innovation and business line more likely to perform cross-func- Through our collaboration, we
leaders who are responsible for decision-making. tional collaboration. identified a specific business op-
The misalignment is so predominant that through At the National Association portunity to enhance member
2021 80% of business line leaders are expected to of REALTORS®, we leverage engagement. We agreed that
override AI.
To increase AI adoption, executives should
take tangible steps toward crafting projects that
leverage AI, with a heightened focus on business
implementation. One path commonly traveled, as
described in Figure 1, is to leverage AI to improve
the efficiency and accuracy of human actions and
decision-making.
Given the challenges outlined above, though,
how can we identify prediction problems that trans-
late to real-world business value? The key is to pro-
mote cross-functional collaboration, identify clear
business opportunities that can be transformed into
AI solutions, and implement AI solutions through-
out the business.
increasing attendance at the asso- tastes and conversion preferenc- Increasing AI adoption is a for-
ciation’s REALTORS® Conference es, to the MCE Group to develop midable goal. Although it comes
& Expo, which is the largest annual a comprehensive strategy for im- with its own set of unique chal-
event for real estate professionals, plementation. lenges, the rewards can be nota-
was a great way to deliver real-world ble. Throughout your journey, you
value. When members attend the BUSINESS IMPLEMENTATION. Value will surely encounter breakdowns
conference and consume confer- comes from action. At minimum, in communication, differences
ence educational and networking business implementation is about of opinions and questions about
opportunities, they’re also more augmented decision-making; us- the true value of the work. Push
likely to be engaged and retained, ing insights derived from machine through those challenges and stay
and that serves as a win-win for intelligence to make better ad on course. Also, remember that
members themselves and the asso- hoc decisions. At best, it’s AI at perfection is the enemy of prog-
ciation. scale. ress.
In our case, the MCE Group Promote cross-functional col-
FROM BUSINESS OPPORTUNITIES TO AI took the DSG’s findings and laboration, identify clear business
SOLUTIONS. crafted a sophisticated strategy opportunities that can be trans-
It’s one thing to identify a busi- to encourage members within formed into AI solutions, and im-
ness opportunity, and it’s another the segment to attend the con- plement AI solutions throughout
to establish an AI solution. In order ference. The strategy included a the business. If you do so, you’ll
to go from business opportunity to campaign with personalized mes- be well on your way to generating
AI solution, you must have a good saging that spoke to members’ real-world business value with AI.
sense of your organization’s AI ca- unique needs and interests. The
pabilities. This is one of the most MCE Group divided the overall Dr. Aleksandar Velkoski
important reasons why you need segment into five distinct catego- serves as the director
to establish a cross-functional team, ries, mainly by decision making of Data Science at the
particularly one that integrates tech- styles, communication tastes and National Association
nology stakeholders as partners in conversion preferences. Then the of REALTORS®. In
the process. team created custom messag- addition, he is an
In our case, we transformed our ing for those groups. Groups in- adjunct professor in the College
business opportunity into a simple cluded personas like Proud to Be of Computing and Digital Media at
prediction question; namely, “Who an American Pam, Glory Seeking DePaul University, and co-founder
is likely to attend the REALTORS® Gloria, Discount Deadline Dave, and organizer emeritus of Chicago
Conference & Expo”. We gathered Rose-Colored Glasses Rosie and ML, a community of machine
a significant amount of data, includ- Generic George. learning scientists. He serves as
ing core member engagement data, Ultimately our goal was to en- a distinguished faculty advisor
demographic data and behavioral courage members to attend the and mentor at the Innovation
data, and then we leveraged tech- conference. After our campaign Development (iD) Lab at DePaul
niques in AI to build a robust pre- was complete, and the confer- University, and as a mentor at
dictive model. Once the model was ence itself had finished, we ob- Second Century Ventures.
built, it was deployed in production served a segment attendance rate
via an API. Finally, we used the API of nearly 1.5% and an increase in
to append to each member record revenue of roughly $100k. That
the output of the predictive model, may not seem significant, but
which was an individualized score given the complexity of the call to
that represented each member’s action — that is to say it required
likelihood to attend the conference. that members incur a collection of
Through leveraging our predic- direct costs (conference fee, flight
tive model, the DSG was able to and hotel) and indirect costs (time
identify an important segment of commitment and lost business)
members that not only were likely to — the rate and revenue gain was
attend the conference, but also had significant. This is especially true
not attended the conference previ- given that segment attendance
ously. The DSG provided the mem- outperformed almost all others.
ber segment, including information Based on this and a variety of
about each member’s unique deci- other factors, it was clear that the
sion-making styles,communication effort was successful.

You might also like