Focus retail and CPG GenAI investments on the core but leave room for innovation

Every manager tasked with understanding how generative AI (GenAI) can accelerate their business is up against a hard reality: there is no easy blueprint for navigating to the right solution.

When we brought together leaders from the retail and consumer packaged goods (RCPG) sector for a roundtable in London—in partnership with Publicis Sapient—the guidance from around the room was that to make meaningful progress, RCPG leaders should stay focused on core business objectives while maintaining space for innovation.

GenAI offers radical new opportunities—but to succeed, take your time and consider carefully

AI and GenAI have the potential to revolutionize RCPG. Still, there is so much new and so many unknowns that the route to an effective strategy remains largely uncharted. Opportunities for new ways of connecting to customers, radical simplification of supply chain processes, and monetizing data assets seem too good to miss. But how do RCPG leaders navigate to what’s right for their business?

We asked 15 enterprise executives from leading RCPGs (see Exhibit 1) how they should invest to prepare their organizations for GenAI’s coming disruption. Their answers centered on the imperative that organizations be thoughtful about investments while ensuring careful consideration does not hamper innovation.

Exhibit 1: RCPG enterprise executives are treading carefully when it comes to balancing investment plans for GenAI—focusing on core value with one eye and future-facing innovation with the other

Source: HFS Research, 2024

Don’t get waylaid by the pursuit of data perfection

Data, particularly good quality data, is at the heart of every successful GenAI initiative. The imperative to improve data quality is clear. Attendees identified it as one of the most significant barriers to GenAI adoption (see Exhibit 2)—both the training data and the data used to fine-tune the AI models. When organizations seek to define their GenAI strategy, they must start by understanding their available data and how to access it. For RCPG leaders, the data landscape is particularly complex. One CPG executive at our roundtable referenced acquisitions as a key factor in adding complexity to data estate.

With every new addition, “you’re going to get new garbage in all the time; that’s the nature of big business.” And a retail leader described the search for perfect data as “a fool’s errand.” They outlined how “business process is our biggest problem—we need to step back and redesign.”

Exhibit 2: Data quality and strategy rank among the top three barriers to GenAI progress

Source: HFS Pulse, 2024; N=132 RCPG executives

GenAI can be messy—make some progress even if the data is imperfect

Firms need to be wary of getting stuck when faced with data challenges. GenAI will be messy. Fortunately, GenAI enables you to optimize and reinvent—even with imperfect data. For the RCPG industry in particular, the data won’t ever be perfect, but GenAI can give you options to move forward.

A pharma CPG leader asked the room whether they “trust in tech or trust in your data?” Several participants agreed that cleaning all data everywhere is impossible, and organizations should not make this the basis of their thinking. Instead, the recommendation was to focus on individual use cases.

Think about new combinations of structured and unstructured data and experiment your way toward scale

GenAI enables you to optimize and reinvent even with imperfect data. The group agreed that RCPG industry data will never be perfect, but GenAI can provide the best options for moving forward.

Many RCPG stakeholders remain unaware of well-proven GenAI use cases for productivity uplifts. Exhibit 3 shows that the RCPG sector is benefiting from productivity and efficiency upticks but is also targeting competitive advantage and improved customer experience.

An engineering executive cited new value creation from predicting component failure, another described examples in finance processing and responding to procurement questionnaires, and one described a use case of continuous formatting of legal documents in a context in which thousands of changes are made weekly.

The advice from the roundtable was to start small, educate the enterprise, and build trust and proofs of concept (POC).

Exhibit 3: RCPG leaders have ambitions for GenAI beyond productivity

Source: HFS Pulse, 2024; N=132 RCPG executives

One retail leader shared an example from customer targeting. Rather than solely focusing on traditional high-level “customer personas” and optimizing incrementally based on them, they used GenAI to power real-time analysis of in-store and online activity to create new personas. These new data-powered personas are proving to be more refined and powerful than the traditional ones and enable the delivery of more targeted messaging to consumers.

Personas built from historical data can be flawed. One executive cited a company that used four-year-old data (i.e., pandemic data) for its current customer personas. Real-time data is likely to prove more accurate and applicable.

GenAI has a role to play in your RCPG business strategy

Roundtable attendees identified that GenAI is sometimes seen as “a tool looking for a problem.” Focusing only on the potential of the technology means you can take your eye away from the actual burning problems that need solving. When thinking about clients’ GenAI strategies, Julian Skelly, retail lead, EMEA and APAC, Publicis Sapient, advised that “companies need to be wary of only focusing on the technology and should ground their GenAI strategy in use cases that address their business strategy.”

However, GenAI’s promise to reinvent entire business models—even if the platform remains smoldering, not burning—means businesses must not look only at already identified problems and risk missing the value opportunities they can’t yet visualize.

A food and drink company executive said that even though GenAI’s quality was ‘not yet there’—and neither was the quality of their organization’s data—they were investing to “futureproof for deeper insights than the human mind can find.”

Despite this call to think big, 61% of RCPG leaders say their firms have not yet moved beyond the pilot stage with GenAI (see Exhibit 4).

Exhibit 4: RCPG adoption of GenAI reflects a lack of urgency. Laggards may live to regret the lead they have given the 36% that have deployed

Source: HFS Pulse, 2024; N=132 RCPG executives

Is now the time for more POCs or to seize the opportunity and scale up?

A CPG retailer said that POCs are still core to demonstrating value and that “the easiest business case is for improved productivity.” But one of their CPG suppliers in the room came back with:

Don’t be obsessed with a [traditional] business case—make sure you’re future-proof. AI is a new business engine with no one business case.

— CPG legal executive

The engineering executive also stated that no company has an entire business case for reinvention—by definition, it’s not yet proven—and that their CFO had been brave enough to say, “Just go and do it.” They further advocated focusing on using R&D budgets to tap into the potential of GenAI innovation, not only the apparent productivity business cases.

A food and beverage executive explained their business cases—the internal case, the case for customer adoption, and the case throughout the value chain—by saying, “Move through those spheres of influence as you build each proof point.”

Exhibit 5 shows that most RCPG leaders now expect to embed GenAI into employee and customer-facing solutions in the next two years.

Exhibit 5: Most retail and CPG firms expect to embed GenAI in customer and employee-facing solutions within two years

Source: HFS Pulse, 2024; N=132 RCPG executives

Helen Merriott (consumer products lead, EMEA and APAC, Publicis Sapient) said, “We don’t know what we don’t know. Companies will find transformation opportunities by experimenting and innovating with GenAI in areas they might not predict today.”

The Bottom Line: Don’t wait for perfection—improve BAU while innovating toward new sources of value.

Accept that data will never be perfect in an RCPG company, and don’t let your teams use waiting for perfection as an excuse not to act. Instead, embrace ways of improving existing processes while working side-by-side with innovation teams, bringing all parts of the business and IT together. Focus on use cases that will enhance current business operations but create enough space for innovations that may point you toward new business models and value.