From search to sale: How AI agents will rewrite discovery
Opinion
Gatekeepers such as OpenAI’s Atlas will sit between brands and audiences, controlling discovery, influencing consideration, and in many cases making the purchase decision. Don’t get filtered out says Mindshare UK’s head of futures and innovation.
OpenAI unveiled Atlas last month, an AI-powered web browser built around ChatGPT, designed to rethink how we search, browse and act online.
Atlas’s USP is that it can operate in (a paid) agent mode, autonomously conducting searches, interacting with e-commerce and travel-booking sites. But beyond the headlines, Atlas points to a much nearer-term reality for brands: we’re entering the age of the AI gatekeepers.
These gatekeepers aren’t just smart assistants. They are autonomous actors capable of making decisions and taking action on our behalf – a market projected to generate up to $450bn in economic value by 2028. As agents like Atlas become part of consumers’ daily digital lives, they will fundamentally reshape how brands are discovered, considered, and purchased.
So how is that shift unfolding across our ever-evolving media landscape? And what does it mean for the brands now being gatekept by these AI-first interfaces?
From assistant to actor
Until now, most AI in consumer life has been assertive. It helps us search, summarise and recommend. But the next generation of AI agents will be capable of acting independently once given a goal – a shift already possible via the use of ChatGPT’s “agent mode.”
Ask for a weekend away and your AI could book flights, find the best hotel, arrange transfers and send the itinerary – all without opening a single tab. Many of OpenAI’s partners, including DoorDash and Booking.com, are exploring how agentic integrations could redefine travel and on-demand experiences.
In this kind of world, the traditional purchase funnel – awareness, consideration, conversion – could collapse into a single step. Discovery becomes proactive, with agents identifying products or services based on the user’s needs and context.
Consideration will increasingly happen invisibly, as the agent compares features, scans reviews, checks prices and delivery times, and even negotiates offers.
Conversion becomes automated, with payment processed in the background. Post-purchase, agents could handle returns, warranty claims or reorders without human involvement.
Advertising, meanwhile, will effectively have two focuses: one stream focused on direct-to-machine optimisation, ensuring agents surface a brand in their automated decision-making, and another stream aimed at influencing humans to prompt their agents. Think: “Do my regular food shop, but remember to get Coke this time.”
We’re already seeing early moves that could apply across a broad range of purchases: PayPal, Visa and Mastercard are developing ‘agent-ready’ payment platforms; Amazon’s ‘Buy For Me’ completes purchases from other sites when its stock runs out.
And this trend, preparing for an agentic AI-enabled future, is also reflected on the agency side.
At WPP Media, we are already including generative search optimisation and AI readiness in our total search and commerce frameworks and providing tools to help leverage AI to gain greater creative control, ensuring clients are embracing and adapting to this evolving landscape.
As AI agents become the gatekeepers to choice itself, brands will increasingly have to win over the machine in addition to reaching the consumer.
The first audience might not be human
In an AI-mediated world, agents will filter, rank and prioritise on behalf of the user. If you can’t pass the algorithm’s criteria, you may never reach a human at all.
That means media planning is no longer just about capturing human attention. It’s about passing machine-level scrutiny.
The agent may decide which brands make the shortlist, which creative is relevant enough to present and which offers are worth surfacing.
Take Google’s AI Mode, launched in the UK a few months ago, which delivers a more in-depth conversational response to a query rather than a clickable list. Similar to Google’s AI Overviews, it reduces opportunities for organic discovery. But it’s where this is going next that is more interesting, with the addition of agentic commerce capabilities currently being tested in the US.
You can, for example, search for a bag and ask it to track that bag across multiple sites until it drops to the price you are willing to pay. When this happens, a notification then gives you the option to ‘buy it for me’, there and then, simply using Google Pay. This demonstrates how the journey from search to sale could bypass the human entirely, fundamentally changing the search landscape.
Aside from what this means for the web as we know it (goodbye, clicks, conversion rates and page ranks), from a marketing perspective, this is not a simple optimisation task – it’s a complete rewiring of media and creative strategy.
AI agents operate on logic, efficiency and structured data. They need product information, pricing and offers presented in machine-readable formats, via APIs and standardised markup. Humans, by contrast, are moved by brand storytelling, emotional resonance and cultural relevance.
To thrive, brands will need dual-audience strategies: factual clarity for AI, and emotional connection for humans.
A product page, for example, might deliver its full dataset in a way that an agent can instantly parse, while still needing to carry compelling imagery, copy and brand cues to create desire in the person behind the decision. It’s not too hard to imagine a world where the majority of a brand’s price or offer-led advertising will be targeted at agents rather than humans.
Trust will decide the pace
While the technology is accelerating, adoption will depend on trust. Our 12 years of trends research tell us that people’s concerns around AI remain pretty consistent – recurring themes centre around bias, accuracy, data privacy and loss of control.
When we spoke to people about commerce agents, specifically in our latest round of research for our ‘Trends 2026’ report, we found that many liked the idea of AI doing the research element for them; however, most would still like oversight of the purchase.
In the near term, we’re likely to see hybrid behaviour, with agents handling the “legwork” (comparing prices, reading reviews, checking terms). However, humans still approve final purchases, especially for high-value items. Over time, as confidence grows, more transactions will be delegated entirely.
Our research also shows that people’s emotional attachment to shopping remains strong, with 65% saying they enjoy finding the best deals themselves. As AI evolves to have enough contextual understanding of the purchase situation, consumers may even maintain multiple agents – one for travel, another for insurance, and another for retail – each with a different level of access to personal data. This could instil a sense of control while still reaping the benefits of automation.
For brands, transparency will be critical: communicating when and how AI agents are involved and ensuring that consumers understand the criteria being applied.
Overall, our research shows that commerce agents are a complex topic for consumers. We need to remain mindful that just because the tech is here now doesn’t mean everyone will be using it overnight.
Having said that, as marketers, it’s important to stay ahead of the curve and start radically re-thinking our marketing strategies in preparation for what lies ahead.
Adapting before the AI gate closes
With the Atlas announcement and the integration of agentic commerce into Google’s AI mode, it’s clear that AI agents are not just talk. These types of agents will sit between brands and audiences, controlling discovery, influencing consideration, and in many cases making the purchase decision.
The brands that adapt now – rethinking their creative, data and media strategies for this AI-mediated reality – will be ready when the gatekeepers become mainstream. Those who don’t may be filtered out long before the human even has a chance to decide.
NB – Watch this space for more details on this topic with The WPP Media UK 2026 Trends report launching on 11 December.
Sophie Harding is the head of futures and innovation at Mindshare UK