By James Taylor, CEO, Particular Audience
When the headlines are full of warnings and relentless hype about AI – which we all know is either set to be disastrously successful or else is a debt-laden stock market bubble about to burst – it can be easy to forget the technology’s genuine potential to make things work a lot better.
In our own retail media field, which is still a sector undergoing rapid development, AI has an opportunity to enhance customer experiences and optimise revenue – but we still have to build it, and a lot depends on the next steps we take.
As a specialist in retail media, search and personalisation, we see three very clear areas in which AI and machine learning can supercharge this space – helping customers find what they want, growing baskets and driving faster and more frequent conversion.
Model Context Protocol prepares RMNs for AI-driven shopping
AI has dramatically eaten into search, but commercial integration is lagging far behind. According to the most recent OpenAI figures, ChatGPT handles more than 2.5 billion prompts per day – roughly 15-18% of Google’s daily search volume – but directly accounts for only tiny revenues. There is a gap that needs closing between engagement and conversion.
Clearly, that is going to change with agentic commerce. And for brands hoping to capitalise on this imminent wave, here is the key message: The most scalable and high-optionality investment you can make right now is in Retail-MCP.
MCP (Model Context Protocol) is an open-source standard that enables AI models to seamlessly connect with external tools and data. It allows retailers to build ChatGPT integrations – known as ‘apps’ – that customers can open through ads. In this way, ads in AI search become functional shopping experiences.
When AI assistants become the new storefront, the entire purchase journey can play out in one place. Commerce and advertising finally collapse into a single system, and brand chat moves from ‘here’s some information’ to ‘here’s the outcome you were trying to achieve’.
Ads stop being messages and become real-time interfaces; checkout becomes a tool, not a redirect; loyalty, order status and support are all there too. Even negotiation becomes feasible. ‘Functional Ad Units’ are critical for any retailer to facilitate.
For brands and retailers, achieving this requires a dual-layer stack, made up of a Transaction Layer, which standardises the handshake between the AI agent and the checkout (for example Universal Commerce Protocol or ‘UCP’, and Agentic Commerce Protocol or ‘ACP’), and an Intelligence Layer, which not only ingests your inventory but infers meaning, vectorises it, and enables business objectives to compete within the ranking cocktail that ultimately governs relevance.
In other words, when an agent requests “waterproof gear for high-humidity climates”, it is able to retrieve semantically accurate SKUs, rather than zero-result keyword failures or obscure token matched results.
The MCP handshake – the process that establishes secure, standardized connections between AI models and external data sources or tools – is now being adopted widely across the AI industry as a standard for AI-tool communication. It is vital that retailers, not platforms, own this layer – because if they don’t control the experience, someone else will.
The good news is, MCP is a free-to-use standard, and retailers and brands only need to build their MCP tooling once. They can then traffic the intelligence into ChatGPT, Claude and others, solving the fragmentation problem in the process.
AI has no tolerance for irrelevance: MCP is only as capable as its underlying infrastructure
For retailers living with legacy keyword-first discovery technology, they have finite days to sort out their tech stacks before such an investment in MCP is otherwise wasted. Most retail platforms still operate two separate systems: one for organic search and another for sponsored products. These are typically stitched together late in the process using rules, boosts, or rank overrides. That approach may blend results, but it does not unify or govern relevance.
The majority of search technology remains keyword-first, usually reverting to rules-based ranking. This works for simple queries, but breaks down when intent is nuanced, seasonal, or contextual, which it often is.
Where humans have a tolerance to facet and filter their way to results, AI will value absolute relevance through mathematical scores, and discard anything below a threshold. Think about it along the lines of the “computer says ‘no’” skit.
The concepts of keyword bidding cannot operate in such a world, and instead advanced Transformer-based search technology must replace it. With relevance curves generated by Adaptive Transformer Search technology, boosting items relative to their organic position achieves an advertising opportunity that is indistinguishable from organic relevant results. This is the golden goose of true relevance – understanding what a customer means and selecting the right product in that moment.
Retail media’s modular revolution
Underpinning MCP with strong technology is mission critical. And AI also offers huge benefits to Retail Media Network (RMN) operators seeking to augment and scale their networks. Any retailer – even those stuck on legacy search technology – needs to be able to rapidly deploy state-of-the-art, AI-driven ad formats and personalisation across these new channels. And so, as retail media evolves, flexibility and intelligence are emerging as the key differentiators not to be stifled by incumbent technology on their legacy web storefronts.
A key development here is in modular retail media tools, which allow RMNs to build and grow commerce monetisation feature by feature on any interface, rather than by adopting a single closed platform meant only for websites.
This development towards modular technology brings retail media into line with the broader trend towards MACH (Microservices-based, API-first, cloud-native and headless) principles for composable enterprise architecture, which are increasingly adopted across global industries, with huge implications for the ROI of AI projects.
This is retail media’s own MACH moment, and this shift into open and transformative technologies is essential for those who are serious about making retail media work to its full potential by cutting costs and speeding up innovation.
