How is AI changing consumer behaviour? At heart, in three places: search, the discovery of new products, and trust. For agencies, that has an uncomfortable consequence. Insight methods built on annual studies and demographic segments lag behind behaviour that keeps on shifting.
At the same time, AI is changing strategists themselves. According to WARC's ‘The Future of Strategy 2026’ report, AI adoption is the biggest change to the strategy role. Some 56% of the strategists surveyed, most of them agency-side, name AI replacing strategy roles as their biggest expected threat – up 15 percentage points on the previous year. And yet 66% see clear reasons for optimism.
How is AI changing the way consumers behave?
Strategists should keep an eye on three shifts – not as laws of nature, but as hypotheses that play out differently in every category and that you should test against your own data.
1. From search query to question
Instead of typing three keywords and comparing ten links, people increasingly ask AI assistants whole questions – and get an answer with a few recommendations. The shortlisting moves from the consumer to the machine. Brands that don't appear in those answers drop out of the consideration set before the comparison has even begun.
2. From searching to being served
Feeds on TikTok, Instagram or YouTube are curated by recommendation algorithms, not by friends. Products are discovered without anyone looking for them. Purchase impulses arise in comment sections and creator videos – in public, and therefore observable.
3. From trusting content to trusting people
The more slick, generic content AI produces, the more weight recognisably real voices carry: reviews, first-hand accounts, creators with a point of view. Brand communication that sounds like everything else gets lost in the background noise.
For each discipline, that means something different. Media agencies need to understand where the shortlisting happens, before an ad exposure is even possible. Creative agencies need to deliver work that recognisably sounds like people, not like the average. CRM agencies are dealing with customers spoilt by algorithms, who bin generic newsletters faster than ever. PR and influencer agencies gain weight because credible voices become more valuable – provided they choose those voices by genuine resonance, not reach alone.
Why aren't classic insight methods enough any more?
Because they are too slow and too far removed. An annual study takes a snapshot of behaviour that is already changing during fieldwork. Demographic segments describe who someone is, but not how that person decides today – with a chatbot, a creator and three comment sections as advisers.
Then there is a question of method. Surveys measure what people say about their behaviour when asked. Public conversations show what they express when nobody is asking: in comments, reviews, forums. Both have their place, but the second sits closer to the moment of decision.
The data illusion: why more data doesn't mean more understanding
Here lies a trap that grows with AI. The so-called data illusion effect describes the tendency to infer deep understanding from a large volume of data. The more curves, clusters and dashboards on the table, the more confident the team feels – regardless of whether it truly understands the people behind them.
AI amplifies the effect. It sums up vast amounts of data in three convincing sentences. Whether those sentences are right often can't be traced any more. But a pattern is not an explanation. Data reveals patterns, but it cannot define humanity.
WARC draws the consequence for the discipline: the value lies in judgement, in knowing which questions are worth asking, and in understanding the people behind the datasets. No summary can do that for you.
Old and new insight work compared
| Dimension | Classic insight work | Insight work in the AI era |
|---|---|---|
| Source | Surveys, panels, annual studies | Public conversations, content and competitor ads, supplemented by surveys |
| Rhythm | Annual or per project | Continuous, with fixed checkpoints |
| Unit | Demographic segments | Personas with motives and tensions |
| Evidence | Aggregated percentages | Every claim traceable to real posts |
| Role of AI | Barely present | Collecting, clustering, summarising |
| Role of the strategist | Sourcing and preparing data | Asking questions, checking, interpreting, deciding |
How do agencies rebuild their insight work?
- Audit your sources. What data are your current personas and positionings built on? How old are they? Who last checked them?
- Make public conversations a standing source. Comments, reviews, creator content and competitor ads from the category belong in every insight foundation.
- Test the three shifts. Ask specifically: how do people in this category discover products? What role do AI assistants and creators play? Whom do they trust?
- Introduce a burden of proof. No AI conclusion goes into the strategy document unless someone has seen the posts behind it.
- Change the rhythm. Don't scrap the annual study, but add a monthly insight check with clear questions.
- Free strategists from data sourcing. AI can take over the collecting, sorting and summarising. The time gained belongs to interpretation and to conversations with the client.
Will AI replace strategists?
The worry is real, as WARC's figures show. The counter-argument is still stronger: AI replaces data sourcing, not judgement. When consumer behaviour changes faster, the need grows for people who can interpret patterns, live with contradictions and derive a direction from them.
Independent agencies can even turn this into an advantage. They usually rebuild processes faster than big networks with entrenched structures – provided they have access to comparable data. That is what the Defy platform is built for: in Defy Monitor, every AI conclusion can be traced back to the real published posts behind it, and strategy comes from what audiences and competitors actually publish, not from surveys.
What does this mean for your next client presentation?
Replace a line like ‘The target audience is 25 to 45, urban and digitally savvy’ with three real quotes from public conversations that show what is on these people's minds right now. Add where they discover products and whom they trust along the way. The client won't ask about the sample size, but about the next step. And that is exactly where strategy begins.
