Marketers have spent the last decade getting very good at collecting consumer data. They have been perhaps less successful at working out how to consistently get the best out of it. AI is changing that calculation, though not necessarily in the direction you would assume.
A fundamental change is underway in how consumers discover, research, and decide, driven by increasingly complex and fragmented marketing touchpoints, and accelerated by the shift towards AI. As consumers constantly move between search engines, websites, apps, AI assistants, retail platforms, and social media, an unprecedented volume of behavioural data is being generated.
For marketers, that creates a conundrum. We have more information about consumers than ever before, yet understanding what somebody actually wants in a given moment, and connecting these dots to insightful strategy, effective targeting and campaign activation, is becoming harder.
Conventional wisdom assumes that more data leads to better advertising. If that’s the case, why do so many of today’s briefs feel more at home with the legacy media terminology of linear channels like broadcast TV? We talk a good game around data, but why as an industry do we still struggle to evolve beyond simple demographics in practice
With exponential growth in data volumes, especially in the age of Gen AI, the question is no longer where to get more data, it is how to identify and leverage the signals that lead to better business outcomes.
The volume trap
Web, mobile, CTV, DOOH, first-party, third-party, behavioural, transactional — each new source has arrived promising to help brands reach the right person at the right moment. The waste tells a different story. eMarketer’s Attention Metrics Ecosystem work puts the share of viewable digital ads that are actually looked at at 30%. Media money is still going to the wrong people at the wrong time, and it is going there with more data behind it than ever.
That is not an argument for another dataset. It is what happens when you trade clarity for fragmentation — more dots, less ability to connect them.
The answer to this problem lies in understanding consumers in all their complexity. If a person is reading an article about spa resorts in Spain, are they in the market for a holiday? Or researching destinations for an intimate wedding? If the latter a wedding photography ad would drive better outcomes than a packaged travel deal.
Consumers are becoming harder to “read”
Modern consumers are second-screeners, multi-taskers, and experience-seekers. Someone who was looking for hiking gear this week may be obsessed with baking videos the next. Historical data rapidly becomes irrelevant, reinforcing the increasing need for real-time contextual signals.
Recently, industry focus has shifted to ‘AI visibility’. While understanding the influential signals from LLMs is certainly valuable, it misses the simple requirement of starting with the consumer. Most of today’s solutions employ mass API-scraping to capture LLM responses to predetermined lists of questions. This may provide some insight into how AI models respond in some situations – but it certainly doesn’t help us to understand the root changes in consumer behaviour.
And neither is there a single, simplistic pattern for brands to understand and quantify. It is tempting to see AI interaction as a single upper funnel discovery stage, but the reality is that consumers chop and change across platforms as they move through their purchase journey. Even the starting point of this journey is inconsistent. Our recent Prompt to Purchase report showed how just over a fifth (22%) of in-market auto consumers start with an LLM prompt, while the 78% majority still begin their research by curating a brand shortlist across the open web.
The customer journey has become more fluid, more dynamic, and harder to interpret as demand has proliferated across channels and environments. The journey has not just become longer. It has become harder to read, because intent is now formed across environments that do not talk to each other.
So where does honest intent actually live?
Not, on the current evidence, where most of the money is. Consumers spend the majority of their online time on the open web, yet eMarketer’s Share of Advertising Budget analysis still puts over 60% of advertiser spend inside walled gardens.In today’s landscape, marketers need to be more selective about the data sources they trust their budgets with.
So, are marketers investing in environments that capture consumers’ genuine intent? Or just those that are the easiest to buy?
Environments like search and conversational AI provide a more honest, accurate, and timely read on what people care about, what they actually need, and what they really want versus passive behavioural signals or past engagement. People rarely lie to their search bar – or, it turns out, their AI assistants.
Consumers increasingly rely on platforms such as ChatGPT, Gemini and Claude not just for automation, but for everyday tasks from meal planning to fashion advice. Adding a conversational AI layer to a marketer’s data toolkit helps create a far richer picture of consumer intent, revealing not just what product someone wants to buy, but why they need it, their budget, personal context, and how they’re planning to use it.
Combining signals beats collecting them
Search intent, contextual engagement and conversational AI each reveal different dimensions of consumer behaviour, and answer different questions.
Search captures explicit demand — what someone is looking for. Context reveals the environment and the interests sitting around it. LLM conversations expose the part we have never had access to: the reasoning. The evaluation, the trade-offs, the constraints people talk themselves through before they buy.
Someone asks an assistant whether an Airbnb or a hotel works better in Tuscany with young children. Then for a five-day itinerary around school holidays. Then for a week of high-protein meals on a fixed grocery budget. Individually, three queries. Together: priorities, constraints, budget, household, timing. No single channel gets you that.
Scale still matters. It just is not the differentiator any more. Confidence in a signal, understanding how signals interact, and knowing which combinations actually predict something — that is where the advantage sits now.
Where is this heading?
No single platform owns the customer journey anymore, so advertisers need to evolve more effective intelligence. That’s why the industry is seeing a broader shift from siloed tools and identifier-based targeting towards integrated systems that connect different data points into an actionable dataset.
Signals from AI conversational interfaces and traditional search environments create a more realistic view of human behaviour. That enables marketers to build audiences that evolve in real time, just like people, rather than relying on static campaign segments built solely around demographics or location. As intent becomes increasingly distributed, understanding how these diverse signals interact will shape the next generation of marketing performance.
The challenge facing marketers is evolving. In the AI era, success will come from sharper signals, not bigger pools of data. Success will depend less on collecting the largest volume of information and more on transforming fragmented signals into predictable outcomes.
This is the problem we have built for at Captify — search intent from the open web and LLM conversation signals read against the same user, so demand and reasoning sit in one place rather than two. Captify’s approach to unified intent creates a unique proposition, allowing marketers access to the real-time rich intent data signals that will deliver more meaningful outcomes for their brands.
We have had a decade of collecting. The next ten years should be about interpreting — and the brands that win will not be the ones with the most data. They will be the ones who know which signals deserve their money.
Sam Coates
Sam Coates is VP Data and Insights at Captify (part of Verve) where he leads global insight, analytics, and measurement functions – transforming consumer intent data into powerful strategic, marketing, and commercial value.
