AUTHOR

MARTIN RYAN
VP, Retail, EPAM
Connect

CONTRIBUTORS

Introduction

We’ve been helping retailers with data and AI projects for more than 15 years. Based on this expertise, we have developed hypotheses for how retail and consumer packaged goods (CPG) organizations should best adopt the power of AI for their business and operations. The two key hypotheses we’ll examine in this article include:

New Business Models

First, retailers and CPG organizations are searching for new business models that will support revenue and expand margins. Examples are Retail Media, subscriptions, services and marketplace.

Optimizing Long Segments

Second, retailers are rapidly moving to a position where they wish to optimize long segments of the retail value chain rather than making isolated optimizations in specific business functions.

In both cases, we find the role of data and AI critical to success and believe there is a transformational opportunity in this sector.

To support this insight, we conducted a large-scale research study across eight industries, including retail and CPG, assessing AI adoption, the challenges businesses face when doing so and the opportunities AI opens for them. Through this survey, we noted that many retailers and CPGs, despite referring to themselves as mature, are primarily looking to optimize existing processes and, therefore, any value they unlock will be highly constrained.

In our AI Report, we uncovered insights from more than 7,300 participants from enterprises with headcounts of 10,000+ evenly split across the C-Suite and Vice President level as well as engineers and developers from nine countries. This includes 900+ senior leaders and engineers from retail and CPG organizations. In this article, we will focus on their responses, exploring how they align with our hypotheses, what this data means for the industry and what retailers and brands can do to succeed in their AI adoption.

Key Results

In terms of AI maturity, 45% of respondents see themselves as advanced, meaning that they have already achieved consistent results (e.g., productivity improvements, cost savings, etc.), and AI has helped them be competitive in the market, and a further 5% as disruptors, although curiously only about 33% of respondents from the disruptors category report claim to have leveraged AI at scale. Companies in the retail and CPG sector are hugely optimistic about gaining value from AI, and 96% of them plan to hire AI-related roles in 2025.

However, we have seen retailers being quite cautious in AI adoption until now, with use cases in forecasting, inventory optimization, transport planning, personalization of marketing communications, content creation, eCommerce recommendations, and so on. Many organizations still view AI as a tool, not as a creative opportunity, so their goal is often to improve what they already do, without drastic transformations or disruption.

Typical AI use cases have been designed to solve specific moments in the retail value chain, and today these AI capabilities are normally embedded into line-of-business applications, often as black box functionality.

One of the biggest opportunities we see now for retail and CPG companies (even more so for CPG) is increasing consumer satisfaction and engagement with the help of AI. We often talk to our clients about how to not just bring AI into daily operations, but also to use it to get to know consumer needs better and approach them in a more personalized way. Such AI solutions can be quite expensive to build, and there is more risk associated with these more ambitious initiatives, but they also allow companies to catch up with more advanced competition. 

There is a lot more we can do with AI from a creative perspective that hasn’t been explored by most retail and CPG companies, and there will be a lot of progress to report in the coming months. Some top-performing CPGs are already moving in this direction by leveraging AI to boost innovation. For example, they are using GenAI in R&D for concept creation, and consumer twins for in-silico testing, increasing precision and reducing risks. In marketing, AI studios enable creative testing automation and precision targeting, while in supply chain management, predictive AI enhances autonomous planning and agility.

According to our report, almost a third of retail and CPG companies have piloted AI programs with their customers and/or end-consumers, and a similar number with internal teams and processes.

What best describes your organization's deployment of AI?

Our research also showed retail and CPG companies prioritize productivity and operational efficiency for their AI adoption. It is reflected both in the organizations’ top goals and the results they have already achieved from AI projects.

While productivity and operational efficiency are important areas of focus in the short term as businesses adopt a data-driven and AI approach to doing business, they are not the end goal.

Instead, we have developed two advanced data and AI hypotheses with our more sophisticated clients, many of which are some of the U.S. and Europe's leading retailers.

EPAM’s Hypotheses

In both cases above, retail and CPG companies face several challenges and actions, which are also reflected in our AI Report:

1. Protecting data is a top challenge cited by 30% of respondents. When data is locked in business applications, it is subject to the controls and protections afforded by that application. Moving data into a data platform for wider consumption by the business does expose it to more risks. These must be carefully controlled so that the benefits that are sought do not come with data loss or exposure. Respondents particularly highlighted data security, data exfiltration, regulatory compliance and data quality as of greatest concern.

When making an investment in AI, how important are these data-related factors?

2. Training and employee enablement is mentioned as a challenge by 24% of respondents, while inability to hire the right talent is a further 23%. There is a clear need for additional skills that have not been present in the organization up till now. As an example, over half (54%) of respondents agree that their workforce does not have the skills necessary to deploy GenAI effectively. Our research shows that almost 50% of retailers believe their organization will need significant retraining for AI, but interestingly, less than 20% know what these new skills are.

It's important to note that employee enablement here should not just be about reskilling to use new systems and solutions, but there should also be a focus on bringing the workforce on the AI journey, enabling them to feel confident and comfortable in using AI. Sometimes companies make the mistake of ignoring the workforce’s fears of or reluctance towards using AI. Organizational change management and development of new processes and ways of working are just as important as technical training.

What percentage of your company's staff do you estimate will need to be re-trained to acquire AI skills within the next 18 months?

Hiring for AI skills is a consistent theme with 96% of companies surveyed looking to hire AI roles in the near future. Machine Learning Engineer and AI Researcher are mentioned as the most in-demand ones.

3. Governance of data and AI capabilities and data strategy was the third area that retail and CPG companies showed us as needing investment. Only 19% of respondents believed their leadership had a clearly defined strategy for how AI would be used to achieve business goals. Curiously, at the same time, 45% of respondents described themselves as ‘Advanced’ in AI adoption. Action is needed to address this gap, and over 50% plan to roll out a comprehensive AI governance model in less than a year, with only 2% having already achieved this. Such a rollout takes time, because in order to have a functional governance model, there is an inevitable impact on its myriad of data sources, back-end systems and integrations.

50%

of Retail and CPG businesses surveyed expect to roll out a comprehensive AI governance model in less than a year

2%

of respondents have already rolled out a comprehensive AI governance model

19%

strongly agree that their leadership has a clearly defined strategy for how AI will be used to achieve business goals

Next Steps

We recommend retailers and brands pursue a mix of AI adoption projects — some quick wins to start to develop the organizational skills and behaviors necessary, some as longer-term goals that can be achieved in stages. In each case, it is important to ensure the work drives clear business benefits, so that each success can help fuel the next innovation. Where an initiative falls short of expectations, this should result in quick learning and a rapid refocusing of the team to avoid losing time.

It is important to incorporate transparency, fairness and ethical considerations into your AI projects from the very start. Consumers are increasingly aware of, and concerned by, matters related to responsible AI.

This phase of AI is about focusing on use cases with broad impact and prioritizing them correctly. To maximize the potential of both their people and AI, enterprises need a prioritization method that ensures they are working on the right AI projects, at the right time, to generate the right business impact. This phase involves testing solutions with your teams, with your limited data sets and with client focus groups, developing PoCs and MVPs, and taking controlled risks.

90% of our survey respondents have or plan to appoint a Chief AI Officer. This will be a key role during this period of rapid learning and adoption. It probably is not a role that will persist indefinitely, but for now, it can help an organization to rally around the opportunity AI brings. In general, there needs to be a strong people focus — employees, marketers, end-users, buyers and merchandisers should all be on board with the AI journey. This is where change management plays a key role.

It will be fascinating to see whether the results of further research will reinforce the above messages. We are excited by the ambition of many of our clients and are supporting them vigorously with our key differentiators in the market being our domain experts, engineering depth and scale, data and AI tooling, proprietary accelerators, and consulting capabilities.