Written by Ash White, Managing Partner - Commerce Strategy
As AI begins to reshape how consumers discover and choose products, retail media is entering a new phase. But as our Total Commerce Summit revealed, preparing for the future is as much about connecting what we already have as embracing what's coming next.
"Stop creating new dashboards."
It might not sound like the most futuristic ambition for retail media, but it was one of the more revealing responses from the closing workshop at dentsu's recent Total Commerce Summit.
Throughout the day, we'd explored the past, present and predictive future of commerce media, considering how the industry has evolved and where the next wave of change might take us. In our final session, we turned the conversation towards action, asking attendees to reflect on their own organisations' ambitions, barriers and priorities.
The responses revealed an interesting tension. Alongside ambitions for AI-powered recommendations, automated media planning and seamless shopping experiences, there were repeated calls for better measurement, more connected data, clearer ownership and stronger evidence of commercial impact.
It's a useful reminder that the future of retail media won't be determined by technology alone. The opportunity is substantial, but realising it will depend on how effectively brands and retailers connect their data, capabilities and decision-making.
From responding to intent to anticipating need
For decades, commerce has been organised around making products available and attractive at the moment someone is ready to buy. Physical retail taught brands how to compete for attention on the shelf, while ecommerce extended that challenge to search results, product pages and digital marketplaces.
Retail media has brought those environments closer to the wider marketing ecosystem, connecting commerce data with media investment and creating new opportunities to influence shoppers throughout their journey.
Now, AI is beginning to change the nature of that journey itself.
Rather than waiting for a consumer to search for a product, predictive systems have the potential to anticipate what they might need, using relevant signals to inform recommendations, experiences and offers. Agentic commerce could take this further, with AI assistants increasingly helping consumers research options, compare products and potentially complete purchases on their behalf.
Imagine someone asking an AI assistant to plan their weekly shop around a particular budget, dietary requirement and set of household preferences. The recommendations they receive could be shaped by product attributes, availability, pricing and previous preferences, potentially narrowing their choices before they've visited a retailer's website or encountered a traditional retail media placement.
For brands and retailers, this introduces a significant question: how do you influence a purchase decision when the consumer may never see the shelf?
The answer isn't simply to find new places to advertise. It's to reconsider how product information, customer understanding, brand preference and media work together to influence decisions.
Brand still matters. But it needs to work harder
There is a temptation with every major technological shift to declare that established marketing principles have become obsolete. The emergence of agentic commerce has prompted some familiar predictions about the diminishing importance of brand.
But that overlooks something fundamental about consumer behaviour. People will continue to develop preferences, form emotional connections and place their trust in brands. Those preferences may influence the instructions they give an AI assistant, the recommendations they're willing to accept and the purchases they ultimately approve.
What changes is how those preferences translate into a purchase.
An AI agent evaluating products isn't responding to brand storytelling in the same way a human does. It needs information it can interpret and compare: product specifications, accurate pricing and availability, credible reviews, clear descriptions and evidence to support relevant claims.
This gives renewed strategic importance to work that commerce teams have been doing for years. Better product pages, cleaner feeds, consistent taxonomy, richer product attributes and more connected data have traditionally been treated as the fundamentals of ecommerce optimisation. Increasingly, they also help determine whether products can be discovered, understood and evaluated by AI-powered systems.
In that sense, the work of becoming retail-ready, then digital shelf-ready, has been preparing businesses for the next challenge: becoming agent-ready.
This doesn't mean replacing creativity or brand-building with technical optimisation. It means recognising that brands increasingly need to succeed on two fronts: creating preference among people while ensuring that the information AI systems use to inform recommendations accurately represents what makes their products valuable.
A brand that people love but machines struggle to understand risks missing opportunities for consideration. Equally, a perfectly structured product feed is no substitute for giving consumers a compelling reason to choose one brand over another.
The opportunity lies in connecting the two.
The industry's ambitions are running ahead of its foundations
Perhaps the most interesting insight from our Summit workshop was how closely the barriers to future growth resembled the challenges retail media has been wrestling with for some time.
Measurement and incrementality emerged repeatedly in attendees' responses, including calls for consistent reporting across platforms, better attribution of in-store sales and more robust evidence of retail media's contribution to business growth.
Others wanted to connect customer, ecommerce and loyalty data, or bring sales performance, media investment and content effectiveness into a more unified view. One attendee envisaged being able to understand, in near real time, not just what was happening across their business but why.
These aren't simply reporting improvements. They're foundational to a more predictive commerce model. If businesses can't reliably connect the signals they already have, their ability to anticipate demand, optimise investment or demonstrate the value of AI-enabled decision-making will remain constrained.
Organisational challenges were just as prominent. Fragmented ownership, competing departmental priorities, limited resources and the difficulty of securing investment all featured in the responses.
Yet the ambitions were anything but modest. Attendees described futures in which AI agents automate media planning and budget recommendations, customer needs are predicted before shopping begins, and experiences feel connected across online and physical environments.
The gap between those ambitions and today's operating reality is where much of the next phase of retail media transformation will take place.
Retail media has already challenged the traditional boundaries between marketing, sales, commerce and technology. Retailers have become media owners, commerce data has become a source of audience intelligence, and investment decisions increasingly need to account for outcomes across the wider customer journey.
Predictive commerce will accelerate that convergence. It will become harder to optimise media independently of product availability, customer experience, creative content or commercial performance.
That makes connected ways of working a competitive capability, not simply an organisational aspiration.
Prediction is only valuable if it improves the experience
There's another important consideration in this shift: just because technology allows brands and retailers to anticipate a customer's next move doesn't mean every opportunity should be acted upon.
There’s a tension between generating retail media revenue and protecting the customer experience. It's an important challenge, particularly as retailers look to scale their media networks and brands explore increasingly automated approaches to engagement.
The promise of predictive commerce is greater relevance and convenience. Done well, it could help consumers discover useful products sooner, reduce friction in the buying process and make interactions with brands feel more responsive to their needs.
Done badly, it risks creating more noise, intrusive recommendations and a shopping experience shaped primarily by the commercial priorities of whoever controls the technology.
For retail media networks, this reinforces the importance of moving beyond simply expanding available inventory towards demonstrating the value they create for advertisers and shoppers alike.
And for brands, it means applying the same rigour to customer experience and creative relevance as to data infrastructure and automation.
The objective shouldn't be to predict more for the sake of it, but rather to use prediction to make commerce genuinely more useful.
Building for what comes next
If there was one message to take from our closing workshop, it was that businesses don't necessarily need another ambitious vision of the future. They need a clearer path towards it.
That might begin with improving the quality and accessibility of product and customer data, establishing shared measures of success, or identifying where fragmented responsibilities are slowing progress. For some, it will mean focused experimentation with AI-enabled planning or more predictive customer experiences. For others, the priority will be demonstrating incrementality and building the investment case to scale what already works.
The starting points will differ, but the underlying principle is consistent: the businesses best placed to embrace predictive commerce will be those that connect their media, data, creative and customer experiences rather than treating each emerging capability as another standalone initiative.
Retail media has already shown us how powerful those connections can be. Its next chapter is about extending that intelligence further, from understanding what customers have done to anticipating what they may need, and from influencing a purchase at the shelf to shaping consideration much earlier in the journey.
Brand won't disappear in that future. But how brands become visible, relevant and chosen will continue to evolve – and the groundwork for winning in that world starts with the decisions businesses make today.