By Laurens Van Wiele, Chief Product Officer at Talon.One
As AI becomes embedded in everyday decision-making, it will transform how consumers discover, evaluate, and choose products – disrupting the commerce landscape as we know it, including retail media.
Research from Adobe in 2025 found that 25% of generative AI users have already used AI platforms, like ChatGPT or Gemini, for shopping and price comparison, while 18% rely on them for tailored product recommendations. Additionally, discussion has increasingly centred on the potential for AI agents to handle transactions entirely on the user’s behalf (known as agentic commerce), with major payment providers such as Stripe and Visa developing protocols to enable seamless agentic shopping.
Agentic commerce creates new opportunities to offer a more personalised, tailored experience for shoppers, but it also presents structural and commercial challenges. After all, when AI is doing the shopping, how will brands ensure they are the ones being chosen?
It’s clear marketers will need to understand and adapt to this new commerce channel – using all the tools in their arsenal to drive discoverability and differentiation.
An inflection point in discoverability and personalisation
Discoverability will be a defining challenge for brands to navigate.
There are some concerns that AI agents will precipitate a race to the bottom, as brands lower prices to get cut through. But, this simplistic approach to determining value would ultimately alienate the many shoppers that place value elsewhere.
Shoppers’ perception of value is influenced by multiple factors, such as quality, delivery, returns, loyalty benefits and promotions – so it’s critical that AI shopping assistants evaluate multiple value drivers. Here, incentives such as loyalty points, tiers, and discounts could play an important role in driving discoverability and consideration in agentic channels, beyond price alone.
There is also huge potential to leverage these incentives to create much more highly personalised online shopping experiences, enabled by identity linking. As the capability evolves, AI agents may be able to “sign in” to a business’s loyalty programme on a customer’s behalf, allowing the agent and customer to factor this into purchase decisions – and earn and redeem loyalty value as part of the transaction.
For example, we can envision an AI shopping assistant recommending Brand X over Brand Y, not just on price, but because the purchase would earn the shopper 200 loyalty points and unlock Gold tier status.
Supercharging personalised retail media
Retail media is already moving in this direction. Last year, Google introduced new loyalty offerings to help retailers retain long-term relationships and surface their programmes more prominently within free and sponsored retail listings.
It’s an approach that’s already delivering results for some brands. Sephora used personalised annotations on listings to show signed-in shoppers specific discounts based on their loyalty tier, driving a 20% increase in click-through rate for personalised ads shown to loyalty customers. (Note: Sephora is a client of my employer, Talon.One).
This integration of personalised incentives into retail media will grow more powerful as agentic commerce matures and shoppers become able to link loyalty memberships to AI shopping agents.
However, the early wave of AI shopping assistants we’ve seen so far has only scratched the surface of what’s possible. In most AI-assisted shopping journeys that exist right now, these value drivers remain largely invisible as platforms haven’t yet factored them into AI decision-making. Meanwhile, many brands have fragmented data and tech stacks, which create a technical barrier to making this possible.
Clearer standards are needed to bridge the gap, better connecting brands’ value drivers to AI platforms and giving brands a technical pathway to discoverability by AI agents.
Baking incentives into agentic commerce
One approach to making this future possible is a set of platform-agnostic standards designed to surface promotions and loyalty incentives across AI agent-based shopping experiences. Talon.One has been developing one such set of standards with the Unified Incentives Protocol (UIP).
The goal is to make incentives an integral part of an agent’s decision-making logic. By enabling brands to surface smart, differentiated incentive strategies, UIP is designed to enable both discoverability and personalisation within AI-driven shopping experiences.
Ultimately, agentic commerce will work best for all industry players if it operates as an open ecosystem, not a walled garden. It has been encouraging to see that Google’s Universal Commerce Protocol is designed as an agnostic set of standards, signalling that even the largest platforms recognise the need for shared frameworks over tech siloes.
If the industry is to unlock the full economic potential of AI-assisted shopping and deliver a consistent, value-aware experience to consumers, this will depend on brands being set up to communicate seamlessly with a multitude of platforms in a consistent way.
Laying the groundwork early
The opportunity to shape the future of agentic commerce is now, while the ecosystem is still in its earliest and most mouldable stage.
The brands that succeed will be those that bring together commerce, marketing, loyalty, ecommerce, data, and engineering, with legal and privacy teams involved early to support identity and consent. Like any major digital transformation, this will take time and require early alignment, clear priorities, and decisive action.
As AI takes on more transactional decision-making, we will see significant changes in how people discover and experience brands, presenting both risks and opportunities. As we sit on the cusp of a new age of commerce, marketers will need to ensure their brands stand out or risk being lost in the noise.
Read more opinion on essential topics from industry thinkers on Retail Media Age.
