Highlights
- Traditional search is evolving rather than disappearing.
- AI overviews mean visibility rises, but traffic declines – and this is likely to be a permanent change.
- Marketing teams need to be developing content that addresses questions their customers are asking at each stage of their purchasing journey
AI is reshaping how customers research, compare and buy. The latest Inflexion Marketing Exchange explored how this is shaping marketing strategies as AI overviews and natural language search impact the economics of customer acquisition.
“There is no shortage of ideas; just a shortage of evidence on what works,” says Dave Kirby, CEO of Coppett Hill, a firm which works with investor-backed businesses to improve marketing and sales performance. “We see what companies are actually doing, what vendors are trying to sell them, and importantly, what is working and what isn’t.”
The opportunities and risks for marketing and sales teams are significant as AI reshapes how customers discover and evaluate businesses online. “Traditional search still matters enormously; the things businesses have relied on for a long time remain in the mix,” he says. “But change is happening fast, and leadership teams need to understand where to focus.”
AI is changing how buyers research – but not replacing search
Much of the conversation around AI suggests traditional search is in terminal decline. In reality, while large language models (LLMs) such as ChatGPT are increasingly being used for research, they remain just one part of a broader shift in online behaviour.
According to Coppett Hill’s analysis, LLMs accounted for around 17% of informational searches in the UK in March, up from 10% a year ago. Meanwhile, the biggest disruption has occurred inside search engines themselves, as Google increasingly serves AI-generated overviews alongside traditional results.
“LLMs appear to be additive rather than substitutive,” says Dave. “People are using them to think through problems and shortlist providers – but they are still using search.” This is particularly the case in B2B, where prospective customers may use AI to build evaluation criteria and pressure-test vendor lists before continuing their selection and purchase choices elsewhere.
Why AI search is reshaping digital economics
AI overviews – the AI-generated summaries increasingly appearing at the top of Google results – are changing the relationship between search impressions and website clicks: visibility rises, but traffic declines. Coppett Hill data reveals an average 50% increase in search impressions, but an 11% decline in non-brand organic traffic.
“We think this is a permanent structural change rather than something temporary,” Dave says.
The commercial impact is significant. As traffic falls, businesses often respond by increasing paid search budgets to maintain lead volumes, but this raises customer acquisition costs: “Businesses are compensating for lower organic performance by spending more on PPC,” Dave says.
Efficiency of spend becomes more important than ever. One overlooked lever is Google’s quality score: a 1-point improvement can improve spend efficiency by 10-15%.
What you need to start doing differently
AI overviews need to be treated as a separate performance channel, requiring dedicated measurement and optimisation.
Companies already benefiting from strong SEO foundations tend to perform better in AI overviews, but Generative-AI Engine Optimisation (GEO) relies more heavily on relevant content, citations and off-page mentions.
Management teams need to be capturing the data on where they appear and where they don’t. The most effective results tend to be highly specific. “A useful framework is publishing the exact answer to the exact question your buyer is asking ChatGPT.” Rather than broad thought leadership, this often means practical FAQs, comparison pages and problem-solving content structured around customer intent.
What works also varies by sector. For some businesses, AI tools primarily cite aggregators and review platforms. For others, especially in specialist B2B markets, they rely more heavily on expert and vendor content, meaning that technical documentation is increasingly a marketing asset, not just a sales-enablement one.
“This is still frontier territory,” Dave cautions. “There isn’t much hard evidence, so reverse engineering and testing are really the best tools available.”
Where AI is genuinely creating value in marketing teams
While the external shift is structural, the internal adoption story is more mixed — despite the hype, AI is delivering modest results in how marketing teams operate. The strongest use cases tend to involve high-volume, repeatable processes rather than executive judgement. Examples include building target account lists, classifying leads against ideal customer profiles, drafting email follow-ups and improving sales enablement processes.
“The implementation is normally more important than the tool,” Dave says. “Many businesses are buying software without thinking about changing how teams actually work.”
Concluding, Dave recommends thinking like your customer:
