You are on page 1of 21

Operations Management Research

https://doi.org/10.1007/s12063-023-00390-z

Enabling artificial intelligence for sustainable food grain supply


chains: an agri 5.0 and circular economy perspective
Sumanta Das1 · Akhilesh Barve2 · Naresh Chandra Sahu3 · Kamalakanta Muduli4

Received: 8 November 2022 / Revised: 2 May 2023 / Accepted: 2 June 2023


© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023

Abstract
The majority of the food grain supply chain (FGSC) is run in a linear fashion, requiring substantial inputs that produce
mostly inedible by-products, environmental damage and wastage. Moreover, population increase, declining food resources,
shifting weather patterns, and dwindling supplies pose serious problems to the FGSC. Effective usage and consumption
of resources to harmonize ecological, economic, and social elements is the need of the hour from the Agri 5.0 and circular
economy (CE) perspective. Fortunately, modern technological developments like artificial intelligence (AI) might represent a
paradigm change in this context. However, enablers for AI adoption haven't been studied sufficiently despite AI's popularity.
Hence, the fundamental objective of this research is to identify and examine key enablers that facilitate rapid AI adoption in
FGSC, empowering Agri 5.0 and CE in India. The primary facilitators for AI adoption have been explored via a literature
review and expert interviews followed by a questionnaire survey. The fuzzy decision-making trial and evaluation laboratory
(F-DEMATEL) approach was then used to create a causal model of the identified enablers. The F-DEMATEL method helps
resolve the uncertainty of researching enabler interactions. Research findings suggest that “Legal and regulatory interven-
tions from the government (E7)” and “Green IoT-driven total automation (E5)” have a significant influence in integrating
AI in FGSC. The results have major ramifications for policymakers. The results may be used to justify future investments
and will also aid decision-makers in India in advancing AI initiatives.

Keywords AI-adoption · Agri 5.0 · Circular Economy (CE) · Food Grain Supply Chain (FGSC) · Fuzzy-DEMATEL

1 Introduction

The management, protection, and availability of agri-


food supplies are key concerns for countries all over the
* Kamalakanta Muduli globe (Das et al. 2021; Mangla et al. 2018). The advent
kamalakanta.muduli@pnguot.ac.pg of globalization and the resulting increase in competi-
Sumanta Das tiveness has pushed businesses toward the digital world,
sd38@iitbbs.ac.in where technologies enabled by Industry 5.0 are playing an
Akhilesh Barve increasingly important role (Akundi et al. 2022; Garske
akhileshbarve@yahoo.com et al. 2021; Liu et al. 2021). AI is one such technology; it
Naresh Chandra Sahu includes all programming codes, systems, computational
naresh@iitbbs.ac.in methods, and automated systems that exhibit intelligent
behavior (Barros et al. 2020; Vincent et al. 2019). In a
1
School of Mechanical Sciences, IIT Bhubaneswar, nutshell, AI is the study and development of computer sys-
Bhubaneswar, Odisha, India
tems that mimic human intellect to generate new forms of
2
Department of Mechanical Engineering, MANIT Bhopal, information and resolve issues (Fraga-Lamas et al. 2021;
Bhopal, Madhya Pradesh, India
Naz et al. 2021). By 2030, AI may add up to $15.7 trillion
3
School of Humanities, Social Sciences & Management, IIT to the world's economy, making it essential for the devel-
Bhubaneswar, Bhubaneswar, Odisha, India
opment of any given country’s economic growth (PWC
4
Department of Mechanical Engineering, Papua New Guinea 2018). AI-driven technologies have found widespread use
University of Technology, Lea, Papua New Guinea

13
Vol.:(0123456789)
S. Das et al.

in a variety of fields, including medicine (Haleem et al. As a natural component, agriculture is integral to the
2019a, b), engineering (Pan and Zhang 2021), business circular economy's emphasis on reusing and recycling
(Rana et al. 2022), etc. According to Belhadi et al., (2021), resources; its proximity to human consumption patterns
supply chain management can be one of the potential areas makes it ideally suited to facilitate the coordinated human-
which may benefit most from AI integration. Despite a natural relationships and ensuing sustainable human devel-
recent flurry of studies, the full capabilities of AI in SCs opment (Fraga-Lamas et al. 2021). Agri 5.0 deals with the
have yet to be investigated (Garske et al. 2021). smooth collaboration between humans and robots whereas
Given the complicated and ambiguous interactions of CE deals with the recycling and optimization of resources.
SC, AI can help develop a synchronized SC that shares The adoption of AI with CE and Agri 5.0 ideas and method-
information and resources cooperatively (Belhadi et al. ologies is necessary to address environmental degradation,
2021). According to Taylor and Fearne (2006), using big ecological harm, and resource depletion in today's traditional
data is necessary to provide a unified method for forecast- farming practices (Liu et al. 2021).
ing demand activity in FSC. Utilizing and capitalizing on Although modern farming has greatly increased yield, it
such massive datasets to make informed decisions is one has also come at a high cost in the form of excessive use of
area where AI is expected to be very useful. Better data energy and other resources; destruction of the natural agri-
management with the use of AI might also be beneficial cultural ecosystem has been witnessed (Cadden et al. 2021;
for SC members (Bhat and Huang 2021; Talari et al. 2022). Dora et al. 2021; Vincent et al. 2019). In this context, envi-
Also, AI may be used to save expenses, give businesses ronmentalism and sustainable development are two exam-
an edge, and boost SC efficiency (Haleem et al. 2019a, ples of social concerns that AI is expected to address. This
b; Holmes et al. 2022; PWC 2018). While the preceding could be done, for example, by increasing the efficiency with
studies have highlighted the relevance of AI in SCs, they which resources and energy are used, streamlining logistics
lack domain-specific insights that management may utilize throughout the entire SC, or by better integrating activities
when considering AI deployment in particular organiza- and ecological use from the Agri 5.0 and CE perspective
tions. This is because organizational variables affect AI (Akundi et al. 2022; Belhadi et al. 2021; Liu et al. 2021).
integration in SCs. AI has enormous potential in the food industry. Previous
There are four phases of FGSCs: inputs, harvesting, work has mostly concentrated on the use and advantages of
processing, and distribution. An AI-integrated FGSC AI in FSC and has not sufficiently investigated the crucial
would allow real-time surveillance of material flows, feature of AI-based data exchange for boosting SC efficiency
would boost responsiveness and efficiency, make it easier (Manning et al. 2022; Monteiro and Barata 2021; Vincent
to spot dangers, and help in strategy development. For et al. 2019). There are limited articles on AI uptake and
instance, the FGSC uses both digital and physical con- utilization in developing areas like India. According to exist-
nectivity, dealing with large concentrations, low profit- ing research on AI in FGSC, key success criteria that might
ability, limited asset allocation, and tight schedules; this is influence and promote AI adoption from the much-needed
where AI technologies can assist in optimizing the systems Agri 5.0 and CE perspectives have not been explored. Many
and enable coordination among stakeholders to meet the organizations still don't know how to account for the impact
requirement of the stocks (Vincent et al. 2019). FGSC is of AI adoption on their overall company strategy (Dora et al.
crucial for the Indian economy because of the many issues 2021). A comprehensive evaluation of such variables for AI
triggered by climate change and biodiversity loss, such deployment in FGSC may be crucial in using this technology
as pollution, broken nutrient cycles, lack of groundwa- to enhance an organization's reputation, profitability, and
ter, starvation, economic hardship, and resource warfare competitive edge. Therefore, this article is primarily moti-
(Alberti et al. 2021; Bhat and Huang 2021; Khandelwal vated by the very low rate at which businesses embrace AI.
and Chavhan 2019). FGSC entails constant quality modi- In this context, we intend to propose a meaningful theo-
fications until customers receive the food grains. Hence, retical and empirical framework employing many funda-
large post-harvest losses (PHL) have a detrimental effect mental variables from existing literature and a dedicated
on food security (Das et al. 2023a, b; Chandrasekaran and questionnaire survey for AI adoption in FGSC in a growing
Ranganathan 2017; Gunasekera et al. 2017). Ensuring the market like India from the Agri 5.0 and CE perspective; this
safety and quality of food grains along the SC becomes can improve economic growth by using the natural ecosys-
a problem (Ling and Wahab 2020; Rathore et al. 2020). tem's material recycling process. Hence the research aims to
Cutting-edge technology like AI may increase SC visibility provide answers to the following research questions:
and traceability, while addressing food quality and safety
RQ 1: What are the enablers that can help facilitate
(Das et al. 2021; Faisal and Talib 2016; Haleem et al. 2019a,
the adoption of AI in the context of Agri 5.0 and the
b). Figure 1 shows the current problems in FGSC and how
CE in FGSC?
AI can eliminate these problems to improve SC efficiency.

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

Fig. 1  Benefits of AI integrated FGSC

RQ 2: How are these enablers interdependent or con- chain management (Das et al. 2021), the manufacturing industry
nected to each other? (Parmar and Desai 2020) etc. Finally, the findings of the research
RQ 3: How can understanding the relationships between will be helpful to policymakers, solution providers, and supply
these enablers be useful for decision-makers seeking to chain experts in assisting them to concentrate on the most crucial
integrate AI from an Agri 5.0 and CE perspective? elements that impact AI adoption in the FGSC.
The remainder of the paper is as follows. Section 2
This paper contributes to the following areas of knowledge:
shows the literature review. Section 3 depicts the research
Firstly, it expands the AI adoption literature by reviewing AI ena-
gaps, followed by data collection in Section 4. The
blers. Secondly, the study adopted an F-DEMATEL approach
research methodology is highlighted in Section 5. Results
to interpret the interrelationships among the enablers. F-DEM-
and discussion are highlighted in Section 6. Section 7
ATEL has the advantage of classifying identified variables into
shows the sensitivity analysis for validation of the data
cause and effect groups; this provides insights into prioritizing
analysis. Finally, Section 8 discusses the research implica-
the enablers when decisions are being made. This methodology
tions and directions for further research.
has been extensively used in different fields such as food supply

13
S. Das et al.

2 Literature review the readily available data, farmers can now cultivate crops
that are both healthy and resource-efficient due to advances
2.1 Role of AI in FSC in AI technology (Belhadi et al. 2021; Bhat and Huang
2021). Hence, technological advancements play a pivotal
In recent years, technological advancements have enabled role in sharing data, especially in creating data-sharing and
a massive rise in the production of food grains. Still, con- computational infrastructures. AI may boost the productivity
cerns over food production and supply are rising dramati- of FSCs by facilitating improved integration among channel
cally across the globe (FAO 2021; Mithun Ali et al. 2019). partners (Bhat and Huang 2021; Maddikunta et al. 2022). It
Problems include a shortage of storage space and poor com- may aid in accelerating the intricate procedures of an FGSC,
munication among members of the supply chain; these are making it more trustworthy as a result. Table 1 gives a short
at the heart of inefficient FGSCs and are responsible for the description of previous literature on AI in the field of SCs.
majority of losses (Chandrasekaran and Ranganathan 2017; The identified enablers are shown in Table 2.
Gunasekera et al. 2017). An estimated 8–10% of worldwide
greenhouse gas emissions are related to wasted food (UNEP
2021). In emerging markets like India, post-harvest losses 3 Research gaps and highlights
account for 40% of food loss due to inadequate storage facili-
ties; this has serious ecological, social, and economic con- By completing an analysis of previously published research,
sequences (Chauhan et al. 2020; Mangla et al. 2018; Sharon we were able to identify a great number of topics that need
et al. 2014). At the same time, traceability issues have also more investigation. Several studies have looked into the
been reported by many researchers, causing diversion of adoption of the latest technological models in multiple areas
food grains and circumventing the planned beneficiaries. (Bestelmeyer et al. 2020; Cadden et al. 2021; Dora et al.
An estimation of 42% of food grains was diverted during 2021; Durrant et al. 2021). Although there has been a growth
2011–12 (Drèze and Khera 2015). It is very important to in research into cutting-edge technologies such as the Inter-
have a proper tracking and traceability system. It is vital net of Things (IoT), big data analytics (BDA), drones etc.,
to check whether the entitled quantity of grains is reaching the role of key enablers in the adoption of AI in FSC has
the beneficiaries; if not. the efficiency of the SC must be received little attention (Barros et al. 2020; Dora et al. 2021).
improved (Das et al. 2021). Due to demand fluctuations, per- According to the available research, AI has a great deal of
ishability, periodicity, and temperature sensitivities, FGSC untapped potential for driving transformation inside FSCs.
provides a particularly attractive environment for study (Das In contrast, an adequate integration system is needed since
et al. 2023a, b; Kamble et al. 2020). Indian FGSC is complex, unorganized, quasi, and includes a
AI has the capacity and is more effective in taking into wide range of intermediaries (Nyamah et al. 2017; Rathore
account the specific difficulties encountered by FGSC. AI et al. 2020). Technologies like AI might play a significant
gives superior technological methods for increasing agricul- role in this context.
tural output of produce, distribution, reducing food waste, However, the adoption of AI in FGSC is crucial as an
and maintaining a safe food supply (Dora et al. 2021; Jain advanced technological transformation has significant eco-
et al. 2021). For example, traditional approaches to crop logical footprints. Therefore, AI adoption needs to be energy
health monitoring are time-consuming and arduous. AI is efficient and reduce the greenhouse gas effect (Fraga-Lamas
a useful tool for checking soil for nutrients and detecting et al. 2021). However, there has been no study on the impact
any problems with the crop. Deep learning is used to assess of key enablers for AI adoption in FGSC from the Agri 5.0
agricultural crop health characteristics. AI-enabled apps help and CE perspective. There has only been investigation into
analyze soil quality, pest problems, and diseases (Bhat and AI integration in Indian FSC from the CE perspective. This
Huang 2021; Syeda et al. 2021). To aid farmers and land study has not considered some of the significant factors like
owners, researchers have highlighted the need for an AI- ‘Customer satisfaction in demand volatility’, ‘Economic
driven system that helps in decision-making in terms of site incentives from the government and private parties’, and
selection with the use of big data science (Vincent et al. ‘Green IoT-driven total automation’; these are all signifi-
2019). These resources help farmers anticipate pest and dis- cant aspects of Agri 5.0 and CE (Fraga-Lamas et al. 2021;
ease outbreaks and facilitate more informed decisions about Garske et al. 2021; Vincent et al. 2019). This study has only
how to best manage their crops. Every day, farms generate a prioritized the AI adoption enablers; it has not explored
mountain of data at ground level. AI has made it possible for any interrelationships between the enablers. Therefore, an
farmers to examine data in real time, allowing them to make industry-specific study is needed to understand the case bet-
more informed choices based on factors like weather, tem- ter and to provide more logical insights (Barros et al. 2020;
perature, water consumption, and soil properties. Utilizing Cadden et al. 2021).

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

Table 1  Previous literature on AI in SC


Authors Field of the study Research type Methodology Work done

(Manning et al. 2022) Generic FSC Qualitative Discussion The study highlighted the importance of
creating a common language for the
deployment of technology throughout
the SC while considering the ethical
implications of AI in the food industry
(Kollia et al. 2021) Belgium Quantitative Deep learning The study adopted a deep learning
algorithm that can recognize and
validate expiration dates on food
package photographs collected during
manufacture
(Durrant et al. 2021) Agri food sector Qualitative Discussion Highlighted that using semantic web,
distributed ledger, machine learning,
and information security will allow
future transformational agri-food
data-sharing infrastructures
(Jain et al. 2021) Indian food processing sector Quantitative Statistical analysis The study highlighted that operational
efficiency can be increased through
the adoption of AI in the SC for SMEs
(Liu et al. 2021) Agriculture sector Qualitative Discussion The study highlighted different chal-
lenges and enabling factors that are
needed for Industry 4.0 transformation
in agriculture
(Garske et al. 2021) Europe Qualitative Discussion The study highlighted that a trustwor-
thy legal framework for safety, data
privacy, access, and security is vital
for the adoption of AI
(Cadden et al. 2021) UK Quantitative Statistical analysis The study highlighted the cultural ena-
blers in the integration of AI into SC
of the manufacturing sector
(Dora et al. 2021) Indian FSC Quantitative Fuzzy SWARA​ The study analyzed the critical success
factors for AI adoption in Indian FSC
(Bhat and Huang 2021) Agriculture Qualitative Discussion The study focused on different chal-
lenges in big data and AI adoption
in agriculture, and also highlighted
the importance of this cutting-edge
technology in farming
(Barros et al. 2020) Agriculture and CE Literature review Discussion The study highlighted different trends
and practices in agriculture from the
CE perspective and concluded that
agricultural waste must be minimized
or recycled
(Bestelmeyer et al. 2020) Agri food sector Qualitative Discussion Highlighted that integrating AI into
agricultural research on a larger scale
can improve the efficiency of the SC;
AI will need long-term, wide-ranging
agricultural observational data
(Vincent et al. 2019) Agriculture sector Quantitative Neural network and This study integrates smart sensors with
Multi-Layer Perception AI systems like neural networks and
(MLP) MLP to evaluate agricultural land
compatibility for farming

Hence, identifying and analyzing the key enablers for enablers into cause and effect groups. Hence, the study pro-
AI adoption from both the Agri 5.0 and CE perspective is vides the causal relationships among identified factors; this
what makes this study unique, adding to the growing body can create a vision and provide insights for industry man-
of empirical literature on this topic. This study not only pri- agers and government to draft more informed policy and
oritizes the enablers for AI adoption but can classify these improve decision making.

13
S. Das et al.

Table 2  Identified AI enablers for sustainable FGSC


Enablers How it influences the adoption of AI References of the evidence in prior studies

Development of computing infrastructure • Transfer of data across borders and legal (Kollia et al. 2021; Manning et al. 2022;
authority Monteiro and Barata 2021)
• Reduced cost
• Increased data storage capacity
• IT connectivity
• Smooth flow of information
• Stability of internet and low latency
Coordination and trust among organiza- • Ability to reach consensus (Bestelmeyer et al. 2020; Durrant et al. 2021)
tions • Better accountability and trust
• Efficient networking
• Smooth flow of information
• Strategic collaboration and inter-firm
networks
Economic incentives from the government • More people-centric approach (UNEP 2021; Wong et al. 2020)
and private parties • Develop building trust and reliability
• Providing resources
• Helps formulate the policy India needs right
now
• Easily accessibility of Agri 5.0 products
Customer satisfaction in demand volatility • More people-centric approach to solving Expert Review
problems
• Develop building trust and reliability
• Understanding the sustainability initiative's
importance and benefits
• More popularity of the AI product
Green IoT-driven total automation • Efficient networking Expert Review
• Real-time accurate data
• Better resource optimization
• Efficient transition from a linear economy to
a circular economy
Promoting private partnerships and bilat- • Create an opportunity for the pilot project (Dora et al. 2021; Fraga-Lamas et al. 2021)
eral collaboration • Triggers sustainability
• Create opportunities for bilateral collabora-
tion
• Better policy formulation
• Financial support
Legal and regulatory interventions from • Data privacy and comprehensive security (Akundi et al. 2022; Jain et al. 2021)
the government system
• Setting standards for technology and infra-
structures
• Mitigating bias and building trust
• Transparency
• Reliability and accountability
• Protection and reinforcement of positive
human values
• Right to equality
• Grievance redressal
• Formulate enforcement mechanisms to
adopt AI
Information sharing and communication • Accurate data sharing (Alberti et al. 2021; Devi et al. 2021; Hari et al.
among SC stakeholders • Better trust and transparency 2018)
• Better accountability
• Effective analysis of the data
Build a data integration team • Well established eco-system (Bestelmeyer et al. 2020; Liu et al. 2021;
• Smart resource management Manning et al. 2022)
• Cleaning the dataset for further analysis
• Data standardization

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

Table 2  (continued)
Enablers How it influences the adoption of AI References of the evidence in prior studies

Job security post AI adoption • Increased interest to learn ML and deep (Cadden et al. 2021; Garske et al. 2021; Liu
learning algorithms et al. 2021)
• Enhanced popularity of AI
• Faster diffusion of AI in SC
Collaborative research for responsible AI • Looks for the potential areas of AI imple- (Dora et al. 2021; Fraga-Lamas et al. 2021)
mentation
• Better sustainability initiatives
• Better human-centric solutions
• Smart resource management
• Knowledge transfer and organizational
maturity
Top management’s involvement, support, • Better accountability and trust (Belhadi et al. 2021; Pan and Zhang 2021;
and commitment • Provide resources Taylor and Fearne 2006)
• Smooth flow of information
• Transparency
• Data security
Data-driven organizational culture • Improve diffusion of AI (PWC 2018; Rana et al. 2022)
• Smooth flow of information
• Looks for potential areas of AI implementa-
tion
• Better accountability and trust
• More reliable outcome
Rescaling internal talent; workforce hiring • Improve knowledge and skills of the (Bhat and Huang 2021; Dora et al. 2021;
employees Khanzode et al. 2021)
• Capability of integrating functionality, data,
and processes
• Improve diffusion of AI
• Better popularity of AI products
Implementing prototype projects for check- • High reliability Expert Review
ing viability • Better responsibility
• Provide a better solution
• Smart resource optimization
Digital literacy and awareness building • The popularity of AI products (Bhat and Huang 2021; Mangla et al. 2018)
• Better applicability of AI in sustainability
initiatives
• Enhanced accountability
• More human-centric solutions
• New educational and talent initiatives

4 Case analysis and data collection 4.1 Demographic profile and data collection

The case study technique was employed in this research to After reviewing the available research, 18 AI enablers in
gain a thorough understanding of enablers for the adoption Indian FGSC were found. Through the use of the Delphi
of AI in the Indian FGSC from the CE and Agri 5.0 perspec- method, a standardized questionnaire was developed for
tive. AI-enhanced farming solutions might represent a water- all 18 enablers and distributed to panel experts to receive
shed moment in the history of the FSC. Hence, this study their feedback. A 5-point Likert scale was used in the sur-
attempts to collect the advice of industry professionals and vey to assess their input. 77 professionals from the food SC
conduct a comprehensive literature evaluation to identify the industry and academic institutions were interviewed using
most important factors facilitating the widespread applica- a questionnaire. 72 of the 77 experts responded with 67 of
tion of AI in the field of FGSC. The most important enablers them completing the task. The demographic statistics of the
and the strength of their causal effects have been uncovered analysts are shown in Table 3.
via quantitative analysis of identified enablers.

13
S. Das et al.

Table 3  Demographic statistics of the respondents for F-DEMATEL. Finally, only 5 experts were chosen ran-
Characteristics Profile No of Percentage domly from the final 15 experts for the sensitivity analysis.
respondents
4.2 Delphi study
Sector/profile Executive Director 09 11.68
General Manager 07 9.09
Table 4 displays the results of a Microsoft Excel spreadsheet
Supply Chain Man- 11 14.28
analysis of the replies. The characteristics of central ten-
ager
dency (mean, standard deviation, and median) were used to
Assistant General 12 15.58
Manager finalize the enablers. According to Kumar et al. 2022, to be
Divisional Manager 14 18.18 included in the Delphi research, the mean and median scores
General Manager 14 18.18 must be greater than or equal to 4. As a result, the same cut-
Academician 10 12.98 off point is used for inclusion of enablers in this research.
Experience in Less than 5 years 17 22.07
years Between 5 and 37 48.05
10 years 5 Research methodology
More than 10 years 23 29.87
Qualifications Graduates 30 38.96 5.1 Fuzzy DEMATEL
Postgraduates 32 41.55
PhD 15 19.48 The study adopted the fuzzy DEMATEL technique. This is an
Gender Male 51 66.23 efficient method to deal with a complicated and fuzzy environ-
Female 26 33.76 ment where subjective and imprecise assumptions can influ-
ence any decisions made (Das et al. 2021; Yadav et al. 2020).
The experts communicate their opinions on linguistic expres-
sion in matters linked to complex structures. The guideline
15 experts were selected from a pool of 67 experts to linguistic communication is a typical way for experts to
through stratified sampling (Taherdoost 2016). In this stimulate their judgments; fuzzy figures may convey linguistic
case, the population of experts was first divided into three principles in actuality. Fuzzy set theory has long been used by
groups based on their levels of experience (see Table 3); five scholars to express and handle uncertainty in decision-making
experts were then selected from each group to collect data (Jeng 2015; Yadav and Barve 2018; Zhou et al. 2011). Experts’

Table 4  Selection of enablers for the adoption of AI in Indian FGSC


SN No Enablers Mean Median Standard Accept/ Reject
deviation

1 Development of computing infrastructure 4.494 5 0.641 Accept


2 Coordination and trust among organizations 4.390 5 0.764 Accept
3 Economic incentives from the government and private parties 4.519 5 0.661 Accept
4 Customer satisfaction in demand volatility 4.429 5 0.696 Accept
5 Green IoT-driven total automation 4.481 4 0.503 Accept
6 Promoting private partnerships and bilateral collaboration 4.338 4 0.837 Accept
7 Legal and regulatory interventions from the government 4.455 5 0.680 Accept
8 Information sharing and communication among SC stakeholders 4.260 4 0.657 Accept
9 Build a data integration team 4.468 5 0.640 Accept
10 Job security post AI adoption 4.351 4 0.556 Accept
11 Collaborative research for responsible AI 4.377 4 0.708 Accept
12 Top management’s involvement, support, and commitment 4.455 5 0.597 Accept
13 Data-driven organizational culture 4.312 4 0.799 Accept
14 Rescaling internal talent; workforce hiring 4.416 4 0.593 Accept
15 Implementing prototype projects for checking viability 4.403 4 0.654 Accept
16 Digital literacy and awareness building 4.312 4 0.799 Accept
17 Open communication 3.182 3 1.060 Reject
18 Democratic methods of governance 3.299 3 1.136 Reject

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

Find out the enablers for AI adoption in FGSC


Standardization of fuzzy numbers

Develop an initial direct relation matrix using


Calculate low (ls) and high (hs) normalized values
expert’s assessment

Calculate the total normalized crisp values


Design the fuzzy linguistic variables

Develop fuzzy initial direct relation matrix Find the crisp score of the kth expert’s assessment

Calculate integrated score by averaging the crisp scores of


Find the normalized fuzzy initial direct relation
all experts
matrix

Develop total relation matrix

Calculate the prominence and net effect of every


criterion

Set a threshold value to obtain the diagraph

Develop the causal diagram

Fig. 2  Flow chart of the research

uncertain decisions may be expressed by triangular fuzzy The step-wise F-DEMATEL method is as follows (Das
numbers (TFN); these are widely used for such situations et al. 2021).
(Das et al. 2021). Figure 2 shows the flow chart of the
research methodology. μà (x)
Let  be the TFN, denoted as (e, f, g), where e ≤ f ≤ g and
e, f, and g are real numbers (see Fig. 3, and Eq. (1)). The
membership functions of  in-universe X can be defined as
1.0
follows:

⎧0, x<e

⎪x − e∕f − e, e ≤ x ≤ f
𝜇Â (x) = ⎨ (1) x
0
⎪g − x∕g − f f ≤ x ≤ g
⎪0, x>g

Fig. 3  The membership function for the TFN

13
S. Das et al.

( )
5.1.1 Step 1: set up the initial direct relation matrix (IDRM) TRM = tij = D(I − D)−1 (4)

The expert linguistic judgment via paired comparison of[the]


variables is derived as an initial direct relation matrix X = xij 5.1.6 Step 7: evaluate the direct and indirect effects
using Table 5 based on the influence score. A [ non-negative
]
matrix for each expert can be defined as X = xijk , where k is
k The TRM results may be explored in Ri and Cj in Table 6 by
the number of experts with 1 ≤ k ≤ K, and n is the number of using Eqs. (5–6).
performance variables. ( )� ( ) [∑n ]
Ri = R1 , ..., Ri , ..., Rn = Ri n×1 tij
j=1
(5)
n×1
5.1.2 Step 3: formation of fuzzy initial direct relation
matrix (F‑IDRM) ( )� ( ) ( )� [∑n ]
Cj = C1 , ..., Ci , ..., Cn = Cj n×1 = Cj n = tij
i=1 1×n
The IDRM is converted to the F-IDRM matrix using the lin- (6)
guistic scale described in Table 5 based on the corresponding where Ri indicates the direct and indirect effect of the ena-
fuzzy scores. Let ̃ xijk = (ekij , fijk , gkij ) represent the fuzzy judg- bler i on other factors. Likewise, Cj represents the direct and
ments of expert k (k = 1, 2, 3,…, K) on the degree to which indirect effects of enabler j by other factors. A positive value
criterion i influences the criterion j. of Ri demonstrates the net influence of the parameters on
the system, while a negative value reveals the system's net
5.1.3 Step 4: compute the overall direct matrix (ODM) impact on the parameters.

The ODM is calculated by combining all the experts chosen 5.1.7 Step 8: setting up a threshold value for the diagraph
for the study using Eq. (2).
∑k Usually, it is up to the pool of decision-makers whether they
Xk =
i=0
(xk )∕k (2) want to set a threshold value of the TRM or not. However,
a threshold value must be established to sort out the most
influential variables among the large number of variables
5.1.4 Step 5: compute the normalized direct matrix (NDM) chosen for the analysis (Das et al. 2021). The threshold value
must be obtained by adding one standard deviation (SD) to
NDM can be derived by Eq. (3), i.e., NDM = [dij ] for all the mean of the TRM. Here, the SD and mean are 0.072 and
0 ≤ dij ≤ 1. 0.176. Thus, the threshold value is 0.249. Table 7 highlights
1 all the values equal to or greater than the threshold value of
NDM = ∑n A (3) the TRM.
max j=1
aij
l≤i≤n
5.1.8 Step 9: develop the cause‑effect relationship digraph
5.1.5 Step 6: compute the total relation matrix (TRM)
Scatter plotting the data set ( Ri+Cj , Ri-Cj ) produces a causal
digraph. Figure 4 shows 16 key enablers for AI adoption in
TRM depicts the comprehensive connection between each
FGSC. Cause group variables are above the X-axis, while
pair of variables, with the component tij designating the indi-
effect group variables are below.
rect influence that factor i has on factor j. I is the identity
matrix of n × n (see Eq. (4)).
6 Results and discussion
Table 5  Linguistic assessment scale and corresponding TFNs
A causal diagram (Fig. 4) was developed after confirma-
Linguistic Variable Influence Corresponding
Score Triangular Fuzzy tion from the assessment group and the determination of
Numbers (TFNs) the TRM. The causal diagram simplifies the case's com-
plexity, allowing for more in-depth analysis and better
No Influence (NO) 0 (0,0,0.25)
judgments. The present study's findings are broken down
Very Low Influence (VL) 1 (0,0.25,0.5)
into three categories: an overall ranking of enablers,
Low Influence (L) 2 (0.25,0.5,0.75)
categorization of enablers into cause and effect groups,
High Influence (H) 3 (0.5,0.75,1)
and explanations of the major enablers for managerial
Very High Influence (VH) 4 (0.75,1,1)
decision-making.

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

Table 6  Cause and Effect group enablers for AI adoption in FGSC


Enablers Ri Cj Ri + Cj Rank Ri—Cj Cause/Effect

EN1 Development of computing infrastructure 3.0677 2.0843 5.152 14 0.984 Cause


EN2 Coordination and trust among organizations 3.467 3.0283 6.495 3 0.439 Cause
EN3 Economic incentives from the government and private parties 3.2668 2.155 5.422 8 1.112 Cause
EN4 Customer satisfaction in demand volatility 2.1409 3.175 5.316 10 -1.034 Effect
EN5 Green IoT-driven total automation 3.4069 3.1338 6.541 2 0.273 Cause
EN6 Promoting private partnerships and bilateral collaboration 3.7404 2.1949 5.935 5 1.546 Cause
EN7 Legal and regulatory interventions from the government 4.1294 2.9863 7.116 1 1.143 Cause
EN8 Information sharing and communication among SC stakeholders 3.4088 2.8227 6.232 4 0.586 Cause
EN9 Build a data integration team 1.7997 3.5851 5.385 9 -1.785 Effect
EN10 Job security post AI adoption 1.0334 3.4763 4.510 16 -2.443 Effect
EN11 Collaborative research for responsible AI 2.6939 3.1937 5.888 6 -0.500 Effect
EN12 Top management’s involvement, support, and commitment 4.1154 1.078 5.194 12 3.038 Cause
EN13 Data-driven organizational culture 2.7621 2.3996 5.162 13 0.362 Cause
EN14 Rescaling internal talent; workforce hiring 1.9559 3.7783 5.734 7 -1.823 Effect
EN15 Implementing prototype projects for checking the viability 1.8958 3.0072 4.903 15 -1.112 Effect
EN16 Digital literacy and awareness building 2.2476 3.0032 5.281 11 -0.786 Effect

6.1 Ranking of the enablers based on (Ri+Cj ) score manufacturers explain how AI judgments are made, the
existing method can operate well in most circumstances.
The identified enablers for AI adoption from the CE and Many researchers have highlighted the need for proper reg-
Industry 5.0 perspective in FGSC are ranked based on the ulatory interventions from the government (Manning et al.
( Ri+Cj ) scores (see Table 6). For the current research, ‘Legal 2022; Monteiro and Barata 2021). Next, ‘Green IoT-driven
and regulatory interventions from the government (EN7)’ total automation (EN5)’ has the second highest ( Ri+Cj ) score
is the most significant enabler with the highest ( Ri+Cj ) of 6.541. IoT-driven agriculture can drastically boost the SC
score of 7.116, while the least important is ‘Job certainty industry by using the resources optimally. This can help to
after AI adoption’ (EN10), with ( Ri+Cj ) score of 4.510. AI access real-time data and better forecasting. Similar reports
generates hazards that existing laws and regulations cannot have been noted (Devi et al. 2021; Hari et al. 2018). Next,
address, hence new legislation is required. If AI technology ‘Coordination and trust among the organizations (EN2)’

Table 7  Total relation matrix (TRM)


EN1 EN2 EN3 EN4 EN5 EN6 EN7 EN8 EN9 EN10 EN11 EN12 EN13 EN14 EN15 EN16

EN1 0.104 0.208 0.146 0.199 0.228 0.163 0.190 0.213 0.253 0.246 0.233 0.058 0.190 0.243 0.202 0.191
EN2 0.122 0.176 0.128 0.221 0.250 0.193 0.244 0.232 0.279 0.269 0.254 0.119 0.203 0.289 0.244 0.244
EN3 0.182 0.214 0.116 0.242 0.239 0.185 0.199 0.206 0.247 0.261 0.209 0.060 0.200 0.259 0.215 0.233
EN4 0.108 0.112 0.126 0.113 0.115 0.112 0.158 0.105 0.130 0.197 0.163 0.038 0.151 0.203 0.135 0.175
EN5 0.183 0.240 0.188 0.247 0.176 0.126 0.205 0.229 0.272 0.267 0.233 0.080 0.203 0.283 0.238 0.239
EN6 0.199 0.258 0.202 0.265 0.250 0.134 0.255 0.212 0.292 0.288 0.268 0.086 0.217 0.305 0.255 0.257
EN7 0.205 0.278 0.219 0.288 0.285 0.221 0.205 0.264 0.316 0.309 0.289 0.145 0.220 0.331 0.277 0.279
EN8 0.124 0.246 0.176 0.202 0.251 0.195 0.224 0.162 0.278 0.218 0.239 0.134 0.188 0.285 0.244 0.244
EN9 0.135 0.092 0.118 0.165 0.096 0.089 0.141 0.087 0.105 0.126 0.150 0.032 0.076 0.167 0.109 0.113
EN10 0.038 0.054 0.074 0.106 0.056 0.041 0.103 0.051 0.062 0.063 0.059 0.021 0.043 0.137 0.072 0.054
EN11 0.130 0.191 0.101 0.212 0.192 0.168 0.188 0.179 0.232 0.176 0.143 0.053 0.127 0.225 0.171 0.207
EN12 0.214 0.262 0.218 0.288 0.285 0.221 0.274 0.264 0.316 0.292 0.289 0.073 0.235 0.329 0.276 0.278
EN13 0.131 0.206 0.119 0.217 0.210 0.104 0.192 0.199 0.232 0.227 0.215 0.055 0.109 0.195 0.156 0.194
EN14 0.066 0.151 0.070 0.104 0.154 0.089 0.102 0.157 0.187 0.179 0.172 0.039 0.074 0.125 0.168 0.118
EN15 0.068 0.162 0.070 0.119 0.164 0.070 0.145 0.091 0.182 0.178 0.155 0.038 0.074 0.190 0.095 0.096
EN16 0.077 0.180 0.085 0.185 0.183 0.083 0.162 0.173 0.202 0.183 0.123 0.048 0.090 0.214 0.150 0.111

The bold values represent values equal to or greater than the threshold value of the TRM

13
S. Das et al.

Fig. 4  Causal diagram of the


enablers

has the third highest ( Ri+Cj ) score of 6.495. Management positive ( Ri - Cj ) scores are found to be in the cause group
often has too little understanding of what AI needs. Modern (see Fig. 6), while the other seven enablers with negative
skills and technology are unquestionably required but it's ( Ri - Cj ) scores are found to be in the effect group. To what
also critical to match an industry's culture, organizational extent the AI adoption in Indian FGSC is successful depends
structure, and working practices to promote widespread AI on the performance of the enablers specified in the cause
adoption. However, old attitudes and methods of working group. Therefore, this is a more important area than the vari-
are incompatible with those required for AI at the major- ables influencing the impact group (Das et al. 2021; Kumar
ity of firms that weren't born digital. Hence, coordination et al. 2022).
and trust must be instilled to ensure better transparency and The enabler “Top management involvement, support, and
decision-making; this will facilitate widespread AI adop- commitment (EN12)” has the highest impact power among
tion in FGSC. Similar insights have been explored by other all the enablers in the cause group with (­ Ri—Cj) score of
researchers (Akundi et al. 2022; Dora et al. 2021; Rajput and 3.038. This means that the support of upper management is
Singh 2019). Figure 5 highlights the rank of all enablers. crucial for securing the resources (facility, capital, IT, and
human resource) needed to put various strategies into action
6.2 Discussion on cause group enablers and generate greater rewards. Support from top management
is frequently cited as a critical enabler and a necessity for
Based on their ( Ri - Cj ) scores, the enablers have been cat- technology adoption. The enabler “Promoting private part-
egorized into two groups—the cause group and the effect nerships and bilateral collaboration (EN6)” has the second
group (Table 6). Out of 16 enablers, nine enablers with highest impact power among the other enablers in the cause

Fig. 5  Prioritized enablers Prioritized enablers based on (Ri + Cj) value


based on ­(Ri + ­Cj) value
8.000
7.000
6.000
5.000
Ri + Cj

4.000
3.000
2.000
1.000
0.000
EN7 EN5 EN2 EN8 EN6 EN11 EN14 EN3 EN9 EN4 EN16 EN12 EN13 EN1 EN15 EN10
Ri + Cj 7.116 6.541 6.495 6.232 5.935 5.888 5.734 5.422 5.385 5.316 5.281 5.194 5.162 5.152 4.903 4.510

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

Fig. 6  Prioritized cause group Cause group enablers


enablers
3.500 3.038
3.000
2.500

Ri - Cj
2.000 1.546
1.500 1.143 1.112 0.984
1.000 0.586 0.439 0.362 0.273
0.500
0.000
EN12 EN6 EN7 EN3 EN1 EN8 EN2 EN13 EN5
Enablers

group with ­(Ri—Cj) score of 1.546. This enabler dispatches encouraging initial investment for AI adoption at the farm
direct impact to other enablers such as EN2, EN4, EN5, level (Cadden et al. 2021; Garske et al. 2021).
EN7, EN9, EN10, EN11, EN14, EN15, and EN16. Hence, “Development of computing infrastructure (EN1)” has
to maximize the value of present and prospects, it will be the fifth highest (­ Ri—Cj) score of 0.984 among the cause
necessary to work together to maximize the effect of invest- group enablers. It is evident from previous research that
ments and activities now underway at all levels. India has a currently, India lacks the computational infrastructure
solid research foundation across all of its priority sectors, but which can eliminate many critical challenges of data stor-
the country has to defragment its research infrastructure to age, data accessibility, data privacy, cloud computing, and
establish AI policies that favour the development of useful, data analysis, etc. Development of computing infrastructure
trustworthy, and resilient AI. can be beneficial in collecting information in many forms
“Legal and regulatory intervention from the government (audio, video, picture, text, and digital maps); data may then
(EN7)” has the third highest impact power among the other be gathered and stored centrally on the cloud for further
enablers in the cause group with ­(Ri—Cj) score of 1.143. analysis and to improve the decision-making process (Fraga-
However, the highest Ri score (4.129) indicates that it is Lamas et al. 2021; Maddikunta et al. 2022).
one of the most important cause enablers with a significant The other significant cause group enablers are “Informa-
impact on others. It is evident from the literature that data tion sharing between the stakeholders (EN8)”, “Coordina-
governance, data protection, connectivity, and data privacy tion and trust among organizations (EN2)”, “Data-driven
are all crucial to the future of agriculture's AI adoption. As organizational culture (EN13)”, and “IoT-driven total auto-
the use of AI and other data-centric processes continues mation (EN5)”. The ­(Ri—Cj) scores of EN8, EN2, EN13,
to grow, so are the number of laws governing them. These and EN5 are 0.586, 0.439, 0.362, and 0.273 respectively.
regulations are obligatory for businesses, particularly those These enablers are less impacted by other enablers as seen
in a heavily regulated sector like agriculture. Similar insights by the lower value of ­(Ri—Cj); on the other hand, a com-
have been highlighted by other researchers that legal safe- paratively high ­(Ri + ­Cj) score indicates the significant
guards may need to be updated to account for the unique capability of these enablers to facilitate the process of AI
challenges posed by algorithmic decision-making (Hangl adoption in FGSC (Manning et al. 2021; Ritchie & Brindley
et al. 2022; Manning et al. 2022). Developing a set of guid- 2007). For example, the success of AI implementation is
ing principles and norms is essential to ensure that all parties highly dependent on information sharing between members
involved in the SC have a shared view of what constitutes involved in the SC. Hence, coordination and trust among
appropriate behavior. stakeholders are vital to ensure credible and quality data.
The enabler “Long-term investment and economic incen- Similarly, a data-driven culture guarantees that everyone has
tives from the government and private parties (EN3)” has a data-first approach and uses data in every decision. By
the fourth highest ­(Ri—Cj) score of 1.112. The use of robots introducing enhanced systems and new technology, more
has benefited the agriculture sector in various ways, includ- teams may access previously segregated or hidden informa-
ing increased production yield, but implementing AI from tion, investigate patterns, and execute changes in response
the Agri 5.0 and CE perspective comes with a huge cost to emerging opportunities. Other researchers (Cadden et al.
involved. Hence, a long-term strategic investment plan 2021; Dora et al. 2021) have also found that data-driven
is necessary to facilitate the adoption process in FGSC. organizational culture plays an important role in the process
This study’s results can help to highlight the importance of AI adoption in FSC.
of long-term commercial investment and incentives while

13
S. Das et al.

6.3 Discussion on effect group enablers projects for checking viability (EN15)” has the third low-
est relation value ­(Ri—Cj = -1.112) and does not signifi-
The SC for food grains operates in very turbulent and unpre- cantly interconnect to other enablers ­( R i + ­C j = 4.903).
dictable conditions; this does not imply that the factors in Several respondents highlighted the importance of a pilot
the effect group are not important to the cause group. Each project to explore the scope, develop policies, create data
factor is necessary, but since all key enablers cannot be con- repositories etc.for the adoption of AI in SC.
centrated at once, it is often important to separate them into Next, “Build a data integration team (EN9)” has been
cause and effect group factors. These effect group enablers identified as one of the key enablers for AI adoption with
are categorized based on the negative (­ Ri—Cj) scores. The a ­( Ri—C j) score of -1.785 and ­( R i + ­C j) score of 5.385.
ranking of the effect group enablers is shown in Fig. 7. Thus, this enabler is not strongly interconnected to others
The factor “Collaborative research for responsible AI but is strongly influenced by other enablers. The result is
(EN11)” has a ­(Ri—Cj) score of -0.5 but the fifth highest in line with other research (Cadden et al. 2021; Dora et al.
­(Ri + ­Cj) score of 5.888. These values suggest that while this 2021; PWC 2018) that has highlighted how data integra-
enabler is greatly impacted by other enablers, it also has a tion will continue to be crucial in ensuring that AI can
significant impact on others. Although EN11 is assigned in meet these obstacles and provide useful results. Not only
the effect group, collaborative research work will help find is the dimensionality important, but also the amount of
solutions to real-world complex problems in FGSC and can data. For machine learning algorithms to have a greater
provide a roadmap for successful implementation. Several probability of arriving at the correct results, they need a
researchers have highlighted the importance of research thorough comprehension of all data properties.
for exploring the scope, developing policies, creating data The enabler “Rescaling internal talent (EN14)” signifi-
repositories etc.for the adoption of AI in SC (Hangl et al. cantly contributes to the adoption of AI ­(Ri + ­Cj = 5.734)
2022; Manning et al. 2022). and has the second lowest relational value ­(R i —
The factor “Digital literacy and awareness building Cj = -1.823). This research outcome is in line with previ-
(EN16)” has a strong interconnection ­(Ri + ­Cj = 5.281) and ous studies (Barros et al. 2020; Dora et al. 2021).
significant impact on other enablers ­(Ri—Cj = -0.786) and “Job security post-AI adoption (EN10)” is impacted by
thus plays an important role for AI adoption in FGSC. This other enablers ­(Ri—Cj = -2.443) with the lowest (­ Ri + ­Cj)
result is in line with both Hangl et al. 2022 and Junaid et al. score of 4.510. This indicates that EN10 doesn’t have
2021. These studies emphasized that digital literacy and much influence on the adoption of AI in FGSC but it can
awareness building among farmers will create an opportu- be achievable by approaching other enablers (Cadden et al.
nity and demand for AI and Agri 5.0-based products in SC. 2021; Garske et al. 2021). Indicator checks for both the
The enabler “Customer satisfaction in demand volatility enablers in the cause group and the effect group are shown
(EN4)” has a (Ri + Cj) score of 5.316 and (­ Ri—Cj) score in Tables 8 and 9, respectively.
of -1.034, indicating that EN4 has a significant impact on
other enablers but is also significantly impacted by other 6.4 Correlation between the enablers
enablers. Thus, to cope with the market's ambiguity and
volatility and to meet the demands of the food supply, The TRM may be used to examine the interplay between
businesses must embrace creative technologies that show enablers. A threshold value of 0.249 is applied to each mem-
a high degree of agility as a consequence of the market's ber of the TRM to exclude any associations with insignifi-
wild fluctuations. The enabler “Implementing prototype cant effects. We further categorize the important impacts

Fig. 7  Prioritized effect group Effect group enablers


enablers
3.000
2.443
2.500
2.000 1.823 1.785
Ri - Cj

1.500
1.112 1.034
1.000 0.786
0.500
0.500
0.000
EN10 EN14 EN9 EN15 EN4 EN16 EN11

Enablers

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

Table 8  Indicator check for cause group enablers


Cause group enablers – Causal enablers
SN no Indicator check Description

1 The highest degree of influence power R EN7 – Legal and regulatory intervention from the government has the highest impact power
R score of 4.129; this indicates that it is one of the most important cause enablers with a
significant impact on others
2 Highest ­(Ri—Cj) score in the group EN12 – Top management involvement, support, and commitment have the highest (­ Ri—Cj)
value of 3.038; this indicates that EN12 is the least impacted enabler by other enablers
3 Lowest ­(Ri—Cj) score EN5 – Data-driven organizational culture has the lowest (­ Ri—Cj) value of 0.273; this indi-
cates that although EN5 is a causal factor, it is highly impacted by other enablers
4 Highest ­(Ri + ­Cj) score EN7 – Legal and regulatory intervention from the government has the highest ­(Ri + ­Cj) value
of 7.116; this indicates that management should pay attention to this enabler because of its
potential capacity of enabling AI implementation in FGSC

between the various enablers into three equal-sized value was allocated (see Table 11). We assigned the weights to the
categories. The minimum–maximum difference is divided experts based on their credibility with weights varied between
by 3 to compute the following interaction. Table 10 shows 0 and 1. Three experts had more than 10 years of experience in
that there are modest (green; values from 0.249 to 0.276), the relevant area among the randomly chosen five experts. So,
moderate (yellow; values from 0.277 to 0.304), and substan- we assigned a comparatively higher weighting to those experts
tial (red; values from 0.305 to 0.331) impacts between the (SCE1, SCE2, SCE3) than the others. The ranking order has
enablers. It can be seen from the heatmap that “Legal and also been contrasted with the average matrix of the experts
regulatory interventions from the government (EN7)” has (Scenario 4). Figure 8 replicates the same ranking for the
a substantial effect on the following factors, EN9, EN10, cause and effect group components as presented in Table 12's
and EN14, and has a moderate effect on EN2, EN4, EN5, sensitivity analysis result, suggesting a negligibly small order
EN11, EN15, and EN16. Similarly, EN12 has a significant difference. As a consequence, the findings are reliable and not
effect on EN9 and EN14 and a moderate effect on EN4, susceptible to weight changes.
EN5, EN10, EN11, and EN16. It is also evident that EN9
is mostly impacted by EN7 and EN12, and moderately
impacted by EN2, EN6, and EN8. Consequently, EN14 is 7 Research implications
mostly impacted by EN6, EN7, and EN12 in addition to
the moderate effect from EN2, EN5, and EN8. Similarly, The current research makes original contributions in the
insights can be drawn from the other enablers. form of theoretical and practical suggestions; both are dis-
cussed in the following sub-sections.
6.5 Sensitivity analysis
7.1 Theoretical implications
The link between cause and effect group variables or bias in
the expert data was examined using a sensitivity analysis (SA). While some researchers have studied the factors that encour-
For the experts listed in Table 3, a mix of different weights age the use of new technologies like AI in the SC (Queiroz

Table 9  Indicator check for effect group enablers


Effect group enablers – Impacted enablers
SN no Indicator check Description

1 The highest degree of influence power R EN11 – Collaborative research for responsible AI has the highest impact power R score of
2.694 indicating that EN11 is the most effective enabler to the administration in terms of
turning AI adoption into practice
2 Highest ­(Ri—Cj) score in the group EN11 – Digital literacy and awareness building has the highest ­(Ri—Cj) value of 0.5; this
indicates that EN11 is most impacted by the rest of the enablers
3 Lowest ­(Ri—Cj) score EN10 – Employment stability post-AI adoption has the lowest (­ Ri—Cj) value of -2.443; this
indicates that EN11 is least impacted by the rest of the enablers
4 Highest ­(Ri + ­Cj) score EN11 – Digital literacy and awareness building has the highest ­(Ri + ­Cj) value of 5.888,
indicating that EN11 is highly significant compared to other enablers

13
S. Das et al.

Table 10  Correlation heatmap of the enablers

Table 11  Weights assigned for Supply chain Supply chain Supply chain Supply chain Supply
experts for SA expert 1 expert 2 expert 3 expert 4 chain
(SCE1) (SCE2) (SCE3) (SCE4) expert 5
(SCE5)

Scenario 1 Considering the average matrix only


Scenario 2 0.20 0.20 0.20 0.2 0.20
Scenario 3 0.35 0.30 0.20 0.10 0.05
Scenario 4 0.15 0.15 0.35 0.25 0.10
Scenario 5 0.20 0.40 0.30 0.05 0.05
Scenario 6 0.25 0.35 0.30 0.05 0.05

Fig. 8  Sensitivity analysis

13
Table 12  SA results
Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6
Ri + Cj Ri-Cj Rank Ri + Cj Ri-Cj Rank Ri + Cj Ri-Cj Rank Ri + Cj Ri-Cj Rank Ri + Cj Ri-Cj Rank Ri + Cj Ri-Cj Rank

5.158 0.984 14 5.087 0.960 14 5.091 0.973 14 5.191 0.994 14 5.090 0.957 14 5.092 0.964 14
6.492 0.491 3 6.401 0.475 3 6.410 0.471 3 6.543 0.504 3 6.403 0.478 3 6.411 0.475 3
5.397 1.057 9 5.328 1.054 9 5.337 1.060 9 5.435 1.054 9 5.328 1.051 9 5.336 1.056 9
5.310 -1.028 11 5.227 -1.008 10 5.244 -1.009 10 5.348 -1.037 11 5.226 -1.009 10 5.236 -1.010 10
6.611 0.242 2 6.490 0.261 2 6.499 0.259 2 6.680 0.231 2 6.495 0.260 2 6.500 0.260 2
5.966 1.554 5 5.867 1.519 5 5.875 1.524 5 6.025 1.576 5 5.869 1.519 5 5.877 1.523 5
7.116 1.159 1 7.002 1.164 1 7.014 1.148 1 7.167 1.166 1 7.005 1.169 1 7.011 1.163 1
6.281 0.597 4 6.172 0.576 4 6.180 0.580 4 6.340 0.605 4 6.176 0.576 4 6.181 0.577 4
5.421 -1.793 8 5.329 -1.776 8 5.337 -1.774 8 5.470 -1.807 8 5.333 -1.777 8 5.337 -1.778 8
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

4.510 -2.489 16 4.450 -2.445 16 4.456 -2.446 16 4.539 -2.520 16 4.451 -2.448 16 4.455 -2.450 16
5.937 -0.507 6 5.835 -0.491 6 5.844 -0.494 6 5.991 -0.511 6 5.839 -0.491 6 5.844 -0.491 6
5.223 3.058 12 5.144 3.012 12 5.149 3.014 12 5.269 3.085 12 5.147 3.014 12 5.151 3.016 12
5.188 0.371 13 5.112 0.365 13 5.115 0.366 13 5.233 0.374 13 5.115 0.365 13 5.119 0.366 13
5.777 -1.807 7 5.673 -1.814 7 5.688 -1.811 7 5.825 -1.811 7 5.676 -1.814 7 5.680 -1.816 7
4.875 -1.082 15 4.801 -1.072 15 4.828 -1.081 15 4.890 -1.088 15 4.798 -1.068 15 4.807 -1.076 15
5.3301 -0.8081 10 5.2032 -0.7797 11 5.2238 -0.7814 11 5.39 -0.8148 10 5.2054 -0.7825 11 5.2137 -0.7787 11

13
S. Das et al.

and Wamba 2019; Pillai and Sivathanu 2020), only a very field conditions and be self-sustainable to perform consist-
few have examined these factors with regard to the FSC. We ently without human interference. This requires a system that
believe that it is better to analyze the adoption process of AI can handle a lot of data and work across devices. Agricul-
keeping the area-specific constraints and potential into con- ture AI devices must be networked and compatible across
sideration. Hence, identifying and analyzing the enablers for platforms for hassle-free data collection; this will encourage
AI adoption in FGSC from the Agri 5.0 and CE perspective further distribution of AI technology in SC.
is novel. The study has listed 16 significant enablers, with • “Top management’s involvement and support (EN12)”,
those factors further divided into cause and effect groups. “Development of computing infrastructure (EN1)”, “Coor-
Thus, this research provides a solid foundation for advancing dination and trust among stakeholders (EN2)”, and “Long
the overall understanding of the key enablers of AI adoption term investment and economic incentives from the gov-
in FGSC in India. Insights from this study may speed up the ernment and private parties (EN3)” have been identified
process of AI adoption in FGSC. as some of the significant enablers. Management must
Secondly, this research makes a significant contribution work with openness and provide support to promote bilat-
to current academic literature by elucidating the interplay eral collaboration and coordination among SC members.
and interdependence of several factors that facilitate the The government must focus on creating demand through
adoption of AI. Using the fuzzy DEMATEL method, we an awareness program and provide the necessary financial
have found several relationships between previously unrec- investment for AI implementation in FGSC.
ognized enablers for AI adoption in FGSC. • Expertise in AI and the accompanying field is scarce.
AI demands highly knowledgeable and experienced
personnel. Although companies are recruiting data sci-
7.2 Managerial implications entists and analysts to use AI-based solutions, demand
for trained specialists exceeds supply. Therefore, man-
The research identifies a range of ramifications for both the agement must provide training and resources to rescale
government and business organizations. Thus, it provides the workforce for adoption of AI. Further, the sensitivity
an operational and flexible decision-making model for the analysis helps detect experts' consistency.
management of AI adoption in FGSC.

• Based on the findings of this research, a prioritized list of 8 Conclusion


enabling factors is provided. Policymakers can use this
list to develop strategies for the successful implementa- Even though India's agricultural industry is in a transi-
tion of AI in FGSC. tional phase and is more reliant on technological integra-
• The classification of cause and effect group enablers tion for improved operations, it nevertheless confronts
helps managers and policymakers in determining the several difficulties throughout the value chain. Technology
key elements on which to focus to advance AI adoption. offers remedies to these problems by introducing disrup-
Managers may now readily determine the aspects that tive interferences. Integrating indigenous and traditional
need more focus than others. The cause group variables farming knowledge with game-changing smart farming
should thus be their main concern since they have a sig- methods, such as the use of AI, is necessary for the success
nificant influence on the effect group elements. of the overall system. Adopting AI technology will allow
• According to the results of the study, “Legal and regula- for greater output with more efficient use of resources via
tory interventions from the government (EN7)” is the improvements in areas such as predictive analysis, crop
most important key enabler among all. It is difficult to health management, quality, traceability etc.
regulate AI, as different people have different opinions In this context, this study aimed to identify the key ena-
on how much and what kinds of regulation will work blers for the adoption of AI in Indian FGSC through an
best for any particular applications it will be used for. extensive literature review and consultation with industry
Hence, it is essential to pinpoint the precise function that experts from Agri 5.0 and CE viewpoints. The study has
legislation may have while framing regulatory and legal drawn up sixteen key enablers for AI adoption and prior-
techniques. Furthermore, given the speed at which this itized them based on ­(Ri + ­Cj) scores. Among all the enablers
nascent technology is developing, a continuing assess- “Legal and regulatory interventions from the government
ment of AI ethical concepts and norms will be required. (EN7)”, and “Green IoT driven total automation (EN5)” are
• According to the study, “Green IoT-driven full automation the top two significant enablers for successful implementa-
(EN5)” is the second most significant enabler for responsible tion of AI in FGSC. Hence, decision-makers must concen-
AI adoption. Management must focus on the quality of the trate on these enablers while drafting policies or strategic
equipment used in India. It must be sturdy enough to survive plans for AI implementation.

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

However, the (R + C) value cannot be used to determine Barros MV, Salvador R, de Francisco AC, Piekarski CM (2020) Map-
which enablers should be the focus since certain enablers ping of research lines on circular economy practices in agricul-
ture: From waste to energy. In Renewable and Sustainable Energy
may not have an impact on other enablers and hence may Reviews (Vol. 131). Elsevier Ltd. https://​doi.​org/​10.​1016/j.​rser.​
not be an overall success in terms of the outcome. There- 2020.​109958
fore, using the ­( R i—C j) value, it is essential to catego- Belhadi A, Mani V, Kamble SS, Khan SAR, Verma S (2021) Artificial
rize the enablers into cause and effect categories. Hence, intelligence-driven innovation for enhancing supply chain resil-
ience and performance under the effect of supply chain dynamism:
F-DEMATEL is used to address the question "What is an empirical investigation. Ann Oper Res. https://d​ oi.o​ rg/1​ 0.1​ 007/​
the relevance of various components in a comprehensive s10479-​021-​03956-x
decision framework?", as mentioned in the Solution Meth- Bestelmeyer BT, Marcillo G, McCord SE, Mirsky S, Moglen G,
odology section. Neven LG, Peters D, Sohoulande C, Wakie T (2020) Scaling up
agricultural research with artificial intelligence. IT Professional
Currently, agriculture is going through a catalytic shift. 22(3):33–38. https://​doi.​org/​10.​1109/​MITP.​2020.​29860​62
The industry is absorbing and using cutting-edge technology Bhat SA, Huang NF (2021) Big data and AI revolution in precision
to improve operational efficiency and consequently increase agriculture: survey and challenges. IEEE Access 9:110209–
production. Technological tools work to bring data science 110222. https://​doi.​org/​10.​1109/​ACCESS.​2021.​31022​27
Cadden T, Dennehy D, Mantymaki M, Treacy R (2021) Understanding
and analytics to each element of the agricultural value chain, the influential and mediating role of cultural enablers of AI inte-
ultimately optimizing the effectiveness of its distribution. gration to supply chain. Int J Prod Res. https://​doi.​org/​10.​1080/​
This article also recognizes some limitations that have 00207​543.​2021.​19466​14
emerged. The research relies on subjective expert views. The Chandrasekaran M, Ranganathan R (2017) Modelling and optimisa-
tion of Indian traditional agriculture supply chain to reduce post-
study was also limited to Indian industrial organizations. harvest loss and CO2 emission. Ind Manag Data Syst 117(9):1817–
Expert views were used to analyze one case organization's 1841. https://​doi.​org/​10.​1108/​IMDS-​09-​2016-​0383
enablers. We appreciate that results may vary according to Chauhan A, Kaur H, Yadav S, Jakhar SK (2020) A hybrid model for
country, industry size, domain etc. Future studies may use investigating and selecting a sustainable supply chain for agri-
produce in India. Ann Oper Res 290(1–2):621–642. https://​doi.​
ISM, AHP, fuzzy MICMAC etc. org/​10.​1007/​s10479-​019-​03190-6
All the data has been provided in manuscript. Das S, Barve A, Sahu NC, Muduli K,Kumar A, Luthra S (2023a) Ana-
lysing the challenges to sustainable foodgrain storage manage-
ment: a path to food security in emerging nations. Int J Food Sci
Authors' contributions Sumanta Das, Ideas, Writing– Original draft Technol. https://​doi.​org/​10.​1111/​ijfs.​16437
preparation, Conceptualization, Formal analysis Dr. Akhilesh Barve, Das S, Barve A, Sahu NC, Yadav DK (2021) Selecting enablers for
Formal Analysis, Visualization Supervision, Project Administration sustainable PDS supply chain in the Indian context using fuzzy-
Dr. Naresh Chandra Sahu, Review editing, Formal Analysis Dr DEMATEL approach. J Agribusiness Dev Emerg Econ. https://​
Kamalakanta Muduli, Critical review, Data Curation, Validation, doi.​org/​10.​1108/​JADEE-​01-​2021-​0025
Commentary and Revision. Das S, Myla AY, Barve A, Kumar A, Sahu NC, Muduli K. Luthra S
(2023b) A systematic assessment ofmulti‐dimensional risk factors
Data availability All the data has been provided in manuscript. for sustainable development in food grain supplychains: A busi-
ness strategic prospective analysis. Bus Strategy Environ. https://​
Declarations doi.​org/​10.​1002/​bse.​3435
Devi A, Julie Therese M, Dharanyadevi P, Pravinkumar K (2021) IoT
Ethics approval All authors follow the ethics in the research and pro- Based Food Grain Wastage Monitoring and Controlling System
vide consent to participate in the research. for Warehouse. 2021 International Conference on System, Com-
putation, Automation and Networking (ICSCAN), 1–5. https://​
Consent to participate All authors provide consent for publication. doi.​org/​10.​1109/​ICSCA​N53069.​2021.​95264​00
Dora M, Kumar A, Mangla SK, Pant A, Kamal MM (2021) Critical suc-
cess factors influencing artificial intelligence adoption in food supply
Conflicts of interest/Competing interests The authors declare that they chains. Int J Prod Res. https://​doi.​org/​10.​1080/​00207​543.​2021.​19596​65
have no known competing financial interests or personal relationships Drèze J, Khera R (2015) Understanding leakages in the public dis-
that could have appeared to influence the work reported in this paper. tribution system. Econ Pol Wkly 50(7):39–42
Durrant A, Markovic M, Matthews D, May D, Leontidis G, Enright J
(2021) How might technology rise to the challenge of data shar-
ing in agri-food? Glob Food Sec 28:100493. https://​doi.​org/​10.​
1016/j.​gfs.​2021.​100493
References Faisal MN, Talib F (2016) Implementing traceability in Indian food-
supply chains: An interpretive structural modeling approach.
Akundi A, Euresti D, Luna S, Ankobiah W, Lopes A, Edinbarough I J Foodserv Bus Res 19(2):171–196. https://​doi.​org/​10.​1080/
(2022) State of Industry 5.0—Analysis and identification of cur- ​15378​020.​2016.​11598​94
rent research trends. Appl Syst Innov 5(1):27. https://​doi.​org/​10.​ FAO (2021) State of food security and nutrition in the world. https://​
3390/​asi50​10027 www.​world​bank.​org/​en/​topic/​agric​ulture/​overv​iew
Alberti L, Bhat SA, Huang N-F, Sofi IB, Sultan M (2021) Agriculture- Fraga-Lamas P, Lopes SI, Fernández-Caramés TM (2021) Green
Food Supply Chain Management Based on Blockchain and IoT: iot and edge AI as key technological enablers for a sustainable
a narrative on enterprise blockchain interoperability. https://​doi.​ digital transition towards a smart circular economy: an industry
org/​10.​3390/​agric​ulture 5.0 use case. Sensors 21(17). https://​doi.​org/​10.​3390/​s2117​5745

13
S. Das et al.

Garske B, Bau A, Ekardt F (2021) Digitalization and ai in european Mangla SK, Luthra S, Rich N, Kumar D, Rana NP, Dwivedi YK (2018)
agriculture: a strategy for achieving climate and biodiversity Enablers to implement sustainable initiatives in agri-food supply
targets? Sustain (Switzerland) 13(9). https://​doi.​org/​10.​3390/​ chains. Int J Prod Econ 203:379–393. https://​doi.​org/​10.​1016/j.​
su130​94652 ijpe.​2018.​07.​012
Gunasekera D, Parsons H, Smith M (2017) Post-harvest loss Manning L, Brewer S, Craigon PJ, Frey J, Gutierrez A, Jacobs N,
reduction in Asia-Pacific developing economies. J Agribusi- Kanza S, Munday S, Sacks J, Pearson S (2022) Artificial intel-
ness Dev Emerg Econ 7(3):303–317. https://​doi.​org/​10.​1108/​ ligence and ethics within the food sector: Developing a common
JADEE-​12-​2015-​0058 language for technology adoption across the supply chain. Trends
Haleem A, Javaid M, Khan IH (2019a) Current status and applications of Food Sci Technol 125:33–42. https://​doi.​org/​10.​1016/j.​tifs.​2022.​
Artificial Intelligence (AI) in medical field: An overview. Curr Med 04.​025
Res Pract 9(6):231–237. https://​doi.​org/​10.​1016/j.​cmrp.​2019.​11.​005 Manning L, Morris W, Birchmore I (2021) Organisational forgetting:
Haleem A, Khan S, Khan MI (2019b) Traceability implementation in The food safety risk associated with unintentional knowledge loss.
food supply chain: A grey-DEMATEL approach. Inform Process Trends Food Sci Technol 118:242–251. https://​doi.​org/​10.​1016/j.​
Agric 6(3):335–348. https://​doi.​org/​10.​1016/j.​inpa.​2019.​01.​003 tifs.​2021.​08.​028
Hangl J, Behrens VJ, Krause S (2022) Barriers, drivers, and social con- Mithun Ali S, Moktadir MA, Kabir G, Chakma J, Rumi MJU, Islam
siderations for AI adoption in supply chain management: a tertiary MT (2019) Framework for evaluating risks in food supply chain:
study. Logistics 6(3):63. https://d​ oi.o​ rg/1​ 0.3​ 390/l​ ogist​ ics60​ 30063 Implications in food wastage reduction. J Clean Prod 228:786–
Hari S, Prasadh HR, Jadisha S, Parthiban P, Narainan SN (2018) 800. https://​doi.​org/​10.​1016/j.​jclep​ro.​2019.​04.​322
Enhancement of IoT based smart salvation and monitoring devices Monteiro J, Barata J (2021) Artificial intelligence in extended agri-
for agriculture. Int J Recent Trends Eng Res 204–209. https://​doi.​ food supply chain: a short review based on bibliometric analysis.
org/​10.​23883/​ijrter.​conf.​02180​328.​034.​6fsd3 Procedia Comput Sci 192:3020–3029. https://​doi.​org/​10.​1016/j.​
Holmes W, Porayska-Pomsta K, Holstein K, Sutherland E, Baker T, procs.​2021.​09.​074
Shum SB, Santos OC, Rodrigo MT, Cukurova M, Bittencourt II, Naz F, Kumar A, Majumdar A, Agrawal R (2021) Is artificial intelligence
Koedinger KR (2022) Ethics of AI in education: towards a com- an enabler of supply chain resiliency post COVID-19? An exploratory
munity-wide framework. Int J Artif Intell Educ 32(3):504–526. state-of-the-art review for future research. Oper Manag Res. https://​
https://​doi.​org/​10.​1007/​s40593-​021-​00239-1 doi.​org/​10.​1007/​s12063-​021-​00208-w
Jain V, Tewary T, Gopalakrishnan BN (2021) Unlocking technology adop- Nyamah EY, Jiang Y, Feng Y, Enchill E (2017) Agri-food supply
tion for a robust food supply chain: evidence from indian food pro- chain performance: an empirical impact of risk. Manag Decis
cessing sector. HSE Econ J 25(1):147–164. https://​doi.​org/​10.​17323/​ 55(5):872–891. https://​doi.​org/​10.​1108/​MD-​01-​2016-​0049
1813-​8691-​2021-​25-1-​147-​164 Pan Y, Zhang L (2021) Roles of artificial intelligence in construction
Jeng DJF (2015) Generating a causal model of supply chain collabo- engineering and management: A critical review and future trends.
ration using the fuzzy DEMATEL technique. Comput Ind Eng Autom Constr 122:103517. https://d​ oi.o​ rg/1​ 0.1​ 016/j.a​ utcon.2​ 020.​
87:283–295. https://​doi.​org/​10.​1016/j.​cie.​2015.​05.​007 103517
Junaid M, Shaikh A, Hassan MU, Alghamdi A, Rajab K, al Reshan MS, Parmar PS, Desai TN (2020) Evaluating sustainable lean six Sigma
Alkinani M (2021) Smart Agriculture Cloud Using AI Based Tech- enablers using fuzzy DEMATEL: a case of an indian manufac-
niques. Energies 14(16):5129. https://​doi.​org/​10.​3390/​en141​65129 turing organization. J Clean Prod 265. https://​doi.​org/​10.​1016/j.​
Kamble SS, Gunasekaran A, Gawankar SA (2020) Achieving sustain- jclep​ro.​2020.​121802
able performance in a data-driven agriculture supply chain: A Pillai R, Sivathanu B (2020) Adoption of AI-based chatbots for hospi-
review for research and applications. Int J Prod Econ 219:179– tality and tourism. Int J Contemp Hosp Manag 32(10):3199–3226.
194. https://​doi.​org/​10.​1016/j.​ijpe.​2019.​05.​022 https://​doi.​org/​10.​1108/​IJCHM-​04-​2020-​0259
Khandelwal PM, Chavhan H (2019) Artificial intelligence in agricul- PWC (2018) Sizing the Prize. What’s the Real Value of AI for Your
ture: an emerging era of research article. https://w ​ ww.r​ esear​ chgat​ e.​ Business and How Can You Capitalise?
net/​publi​cation/​33558​2861 Queiroz MM, Fosso Wamba S (2019) Blockchain adoption challenges
Khanzode AG, Sarma PRS, Goswami M (2021) Modelling interactions in supply chain: An empirical investigation of the main drivers in
of select enablers of Lean Six-Sigma considering sustainability India and the USA. Int J Info Manage 46:70–82. https://​doi.​org/​
implications: an integrated circular economy and Industry 40 per- 10.​1016/j.​ijinf​omgt.​2018.​11.​021
spective. Prod Plann Control. https://​doi.​org/​10.​1080/​09537​287.​ Rajput S, Singh SP (2019) Connecting circular economy and indus-
2021.​19809​08 try 4.0. Int J Inf Manag 49:98–113. https://​doi.​org/​10.​1016/j.
Kollia I, Stevenson J, Kollias S (2021) AI-enabled efficient and safe ​ijinf​omgt.​2019.​03.​002
food supply chain. Electronics 10(11):1223. https://​doi.​org/​10.​ Rana NP, Chatterjee S, Dwivedi YK, Akter S (2022) Understanding
3390/​elect​ronic​s1011​1223 dark side of artificial intelligence (AI) integrated business analyt-
Kumar M, Raut RD, Sharma M, Choubey VK, Paul SK (2022) Enablers ics: assessing firm’s operational inefficiency and competitiveness.
for resilience and pandemic preparedness in food supply chain. Eur J Inf Syst 31(3):364–387. https://​doi.​org/​10.​1080/​09600​85X.​
Oper Manag Res. https://​doi.​org/​10.​1007/​s12063-​022-​00272-w 2021.​19556​28
Ling EK, Wahab SN (2020) Integrity of food supply chain: Going Rathore R, Thakkar JJ, Jha JK (2020) Evaluation of risks in foodgrains
beyond food safety and food quality. Int J Prod Qual Manag supply chain using failure mode effect analysis and fuzzy VIKOR.
29(2):216–232. https://​doi.​org/​10.​1504/​IJPQM.​2020.​105963 Int J Qual Reliab Manag 38(2):551–580. https://​doi.​org/​10.​1108/​
Liu Y, Ma X, Shu L, Hancke GP, Abu-Mahfouz AM (2021) From IJQRM-​02-​2019-​0070
industry 4.0 to agriculture 4.0: current status, enabling technolo- Ritchie B, Brindley C (2007) An emergent framework for supply
gies, and research challenges. IEEE Trans Industr Inf 17(6):4322– chain risk management and performance measurement. J Oper
4334. https://​doi.​org/​10.​1109/​TII.​2020.​30039​10 Res Soc 58(11):1398–1411. https://​doi.​org/​10.​1057/​palgr​ave.​
Maddikunta PKR, Pham QV, Deepa BP, Dev N, Gadekallu K, Ruby jors.​26024​12
TR, Liyanage MR (2022) Industry 5.0: a survey on enabling tech- Sharon M, Abirami CVK, Alagusundaram K (2014) Grain storage
nologies and potential applications. J Industrial Inform Integr 26. management in India. J Postharvest Technol 02(01):012–014.
Elsevier B.V. https://​doi.​org/​10.​1016/j.​jii.​2021.​100257 www.​jpht.​info

13
Enabling artificial intelligence for sustainable food grain supply chains: an agri 5.0 and…

Syeda IH, Alam MM, Illahi U, Su’ud MM (2021) Advance control and supply chain management among Malaysian SMEs. Int J
strategies using image processing, UAV and AI in agriculture: a Inf Manag 52:101997. https://​doi.​org/​10.​1016/j.​ijinf​omgt.​2019.
review. World J Eng (Vol 18(4):579–589. Emerald Group Hold- ​08.​005
ings Ltd. https://​doi.​org/​10.​1108/​WJE-​09-​2020-​0459 Yadav DK, Barve A (2018) Segmenting critical success factors of
Taherdoost H (2016) Sampling methods in research methodology; How humanitarian supply chains using fuzzy DEMATEL. Benchmark-
to choose a sampling technique for research. SSRN Electronic ing 25(2):400–425. https://​doi.​org/​10.​1108/​BIJ-​10-​2016-​0154
Journal. https://​doi.​org/​10.​2139/​ssrn.​32050​35 Yadav S, Luthra S, Garg D (2020) Internet of things (IoT) based coor-
Talari G, Cummins E, McNamara C, O’Brien J (2022) State of the dination system in Agri-food supply chain: development of an
art review of Big Data and web-based Decision Support Systems efficient framework using DEMATEL-ISM. Oper Manag Res.
(DSS) for food safety risk assessment with respect to climate https://​doi.​org/​10.​1007/​s12063-​020-​00164-x
change. Trends Food Sci Technol 126:192–204. https://​doi.​org/​ Zhou Q, Huang W, Zhang Y (2011) Identifying critical success factors
10.​1016/j.​tifs.​2021.​08.​032 in emergency management using a fuzzy DEMATEL method.
Taylor DH, Fearne A (2006) Towards a framework for improvement Saf Sci 49(2):243–252. https://d​ oi.o​ rg/1​ 0.1​ 016/j.s​ sci.2​ 010.0​ 8.0​ 05
in the management of demand in agri-food supply chains. Sup-
ply Chain Manag Int J 11(5):379–384. https://​doi.​org/​10.​1108/​ Publisher's Note Springer Nature remains neutral with regard to
13598​54061​06823​81 jurisdictional claims in published maps and institutional affiliations.
UNEP (2021) Food waste index report 2021. https://​www.​unep.​org/​
resou​rces/​report/​unep-​food-​waste-​index-​report-​2021 Springer Nature or its licensor (e.g. a society or other partner) holds
Vincent DR, Deepa N, Elavarasan D, Srinivasan K, Chauhdary SH, exclusive rights to this article under a publishing agreement with the
Iwendi C (2019) Sensors driven ai-based agriculture recommen- author(s) or other rightsholder(s); author self-archiving of the accepted
dation model for assessing land suitability. Sens (Switzerland) manuscript version of this article is solely governed by the terms of
19(17). https://​doi.​org/​10.​3390/​s1917​3667 such publishing agreement and applicable law.
Wong L-W, Leong L-Y, Hew J-J, Tan GW-H, Ooi K-B (2020) Time to
seize the digital evolution: Adoption of blockchain in operations

13

You might also like