Preparing your brand’s data for the era of AI-powered search
Opinion – Week in Focus
Brands are no longer competing for clicks; they are competing to be selected, interpreted, and recommended by AI systems. LiveRamp’s Luke Fenney offers advice for an agentic future.
The way people discover information online is undergoing its most significant shift since the rise of search engines in the 1990s. For almost thirty years, the search bar has been the front door to the internet. Users typed keywords, received a list of blue links, and did the heavy lifting of clicking, reading, and synthesising information.
Today, that interface is rapidly becoming a relic. The traditional search experience is being transformed by conversational, AI-driven experiences, with current industry projections suggesting AI search will drive more visitors to websites than traditional search by 2028.
The large language models (LLMs) behind this shift have transformed how consumers converse with technology. The next phase of this behaviour is being driven by AI agents: autonomous software programs capable of evaluating, making decisions and taking actions on behalf of a person.
Rather than browsing manually, users can delegate tasks such as “find the best running shoe for a marathon under £150” or “plan a family trip to Italy.” The agent then evaluates options and presents the user with a single, best-fit recommendation rather than a page of options. It can then go a step further and make the purchase.
In this new paradigm, brands are no longer competing for clicks; they are competing to be selected, interpreted, and recommended by AI systems.
Search ranking algorithms matter less to brands than relevance, trust, and contextual understanding; it’s about being understood and trusted by the agents acting on behalf of consumers.
Marketers need to rethink their strategy if they’re going to succeed.
Visibility is key
As AI agents increasingly act as the interface between consumers and brands, advertisers need to rethink what visibility actually means. Visibility is no longer something brands can simply buy; it’s something they have to earn through the quality of their data.
A brand’s visibility now depends on how effectively its data can be discovered, understood, and trusted by external AI systems. Brands need to be made visible to the external AI agents that their audiences use. This relies on signals. Agents assess relevance, credibility, and usefulness based on structured inputs that describe entities, relationships, behaviours, and outcomes.
If data signals such as product descriptions, prices, and availability are fragmented, inaccessible, or unreliable, even the most advanced AI tools will struggle to accurately represent them. The risk isn’t just lower performance – it’s invisibility. If your data can’t be interpreted by AI agents, your brand may never even enter the consideration set.
For marketers hoping to find new customers online, they should prioritise the quality and consistency of these signals that AI systems can interpret and trust.
Structured, permissioned and connected data: the new foundation
High-fidelity signals share three defining characteristics. They are structured in ways that machines can easily interpret, are permissioned for responsible use, and are connected across environments to provide context rather than isolated facts. Without these qualities, AI systems default to assumptions that may be incomplete or misleading.
This challenge is particularly acute in marketing. Purchase decisions are influenced by a confluence of factors and clarified by high-quality data. When signals are locked within organisational silos, AI agents lose the ability to connect cause and effect.
This is the core problem: modern consumer journeys span platforms, publishers, and devices. No single organisation can supply sufficient context on its own. They need data collaboration.
When data is safely connected across organisations, underpinned by shared standards and governance, AI systems gain a richer, more accurate picture. Organisations can participate in shared data collaboration networks that ensure responsible use and mutual value creation among the various parties.
For brands, this represents a shift from isolated data ownership to collaborative data ecosystems, where standardisation directly impacts discoverability and AI recommendations.
For example, permissioned connectivity delivers scale without sacrificing privacy, ensuring insights are both compliant and actionable. Moreover, an interoperable data layer allows signals to persist across use cases, from incremental measurement to advanced personalisation and forecasting.
It enables AI systems to reconcile identity, behaviour, and outcomes without relying on invasive tracking or probabilistic shortcuts. This consistency improves both accuracy and trust.
Re-architecting your data infrastructure for AI legibility represents a shift in mindset. Instead of optimising each activation in isolation, organisations should focus on building trustworthy and up-to-date intelligence that compounds. This elevates AI from simple automation to informed decision-making at scale.
Building for an agentic future
Brands are no longer just speaking to consumers. They are also communicating with the AI systems that advise them, and they need to be able to talk back.
Connected, permissioned data provides the foundation for this dialogue. It enables internal AI to make smarter decisions and deliver relevant experiences without compromising privacy. By contrast, investing in AI without addressing these data foundations is a strategic dead end.
As AI agents increasingly dominate search, your visibility depends as much on data architecture as it does on creative output. Re-architecting around trusted signals is an investment in relevance for an agentic world, where success is defined not by who shouts the loudest, but by who is understood the best.
In an agent-led ecosystem, brands won’t win by just being the most visible – they’ll win by being the most intelligible to machines.
Luke Fenney, SVP, publishers and platforms International at LiveRamp