When the Answer Becomes the Marketplace

For more than two decades, one of the central assumptions of digital marketing was remarkably simple:

People search.

Search engines provide links.

People click.

Websites persuade.

Customers convert.

An enormous industry was constructed around that sequence.

Search engine optimisation attempted to move organisations higher in the results. Content marketing created material capable of attracting searches. Paid search bought visibility around valuable queries. Analytics systems measured impressions, rankings, clicks, sessions and conversions.

Artificial intelligence is beginning to disturb the sequence itself.

Increasingly, the first meaningful encounter between a person and an organisation may not take place on the organisation’s website.

It may take place inside an answer.

Someone asks an AI assistant:

Which organisations could help me develop a community arts project?

What is the best printing process for a limited-edition digital artwork?

How should I turn an early-stage business idea into a credible 90-day roadmap?

Which consultants specialise in cultural strategy and artificial intelligence?

What should I consider before applying for arts funding?

The AI system does not necessarily return ten blue links and ask the user to conduct the research.

It may conduct part of the research itself.

It can interpret the question, compare alternatives, summarise evidence, explain trade-offs, introduce organisations and help the user decide what to investigate next.

The consequence is profound.

The competition is gradually moving from:

Can people find us?

towards:

Will intelligent systems understand us well enough to include us in the answer?

And beyond that lies an even more important question:

When someone encounters us through an AI-generated answer, what evidence will make them trust us enough to continue?

That is the emerging frontier of digital discovery.

Search Is Becoming a Decision Environment

Traditional search is fundamentally a retrieval system.

Someone enters a query and receives possible destinations.

AI-assisted search increasingly behaves differently.

It attempts to understand the intention behind the question and construct an answer.

That distinction matters enormously.

Consider someone searching traditionally for:

business strategy consultant Nottingham

The search engine might provide directories, advertisements, map listings and websites.

The individual must investigate.

Now imagine that person asking:

I run a small cultural organisation in Nottingham. We have a promising community project but no clear business model. I need somebody who understands culture, community participation, funding and AI. What kind of consultancy should I look for and how should I evaluate them?

This is no longer simply a keyword.

It is context.

The system can identify multiple requirements simultaneously: location, sector, organisational stage, commercial need, cultural sensitivity, technology, funding and decision criteria.

That changes what successful digital visibility means.

The organisation with the largest website or highest ranking for one keyword is not automatically the organisation most relevant to the complete question.

AI systems can potentially assemble evidence from many sources and construct an answer around the user’s actual situation.

Search therefore begins to move from retrieving information towards interpreting intention.

The Strange New Economics of AI Traffic

One of the most interesting developments concerns something marketers have traditionally valued above almost everything else:

traffic.

AI search creates an apparent contradiction.

It can reduce clicks while simultaneously increasing the potential value of some of the clicks that remain.

Research from Pew Research Center illustrates the first side of this equation.

Its analysis of Google searches conducted by 900 U.S. adults found that when an AI-generated summary appeared, users clicked a conventional search result during 8% of visits. When no AI summary appeared, the figure was 15%.

Clicks on sources contained directly inside the AI summary were rarer still: around 1% of visits.

Users encountering an AI summary were also more likely to end their browsing session without progressing elsewhere.

The implication is uncomfortable for publishers and marketers:

being visible no longer guarantees receiving the visit.

The answer itself can satisfy the search.

But something different happens among some of the people who do continue.

They may be considerably better informed.

Before reaching the website, the AI may already have explained what the company does, how its product works, how it compares with alternatives, what problems it solves, who it is appropriate for and what evidence supports its claims.

The website is therefore no longer necessarily educating a completely cold visitor.

It may be receiving someone who has already completed part of the evaluation process.

Fewer Visitors Can Sometimes Mean Better Visitors

This is where some recent commercial evidence becomes particularly interesting.

Adobe reported that during the 2025 U.S. holiday season, traffic arriving at retail websites from generative AI sources increased dramatically.

More importantly, Adobe found that AI-referred retail visitors converted 31% more than other traffic sources during the period.

They were also less likely to leave immediately, while revenue per visit from AI-referred traffic had increased substantially during the year.

Earlier Adobe research showed the transition happening almost in real time.

In July 2024, AI-referred retail traffic was substantially less likely to convert than other traffic. By February 2025, much of that difference had disappeared.

AI-referred visitors were also showing lower bounce rates, more page views and longer sessions.

Research-intensive categories appeared particularly suited to AI-assisted discovery.

Why?

Because artificial intelligence can compress research.

The customer does not necessarily arrive saying:

Who are you?

They may arrive saying:

I think you might be the organisation I need. Prove it.

That is a very different visitor.

But There Is an Important Warning

The evidence is not uniform.

And that matters.

Ahrefs reported a striking result from its own website: AI search represented only around 0.5% of its traffic during the period studied but accounted for 12.1% of sign-ups.

For Ahrefs, AI-search visitors converted at a dramatically higher rate than conventional organic-search visitors.

However, BrightEdge reported very different results from research across a wider group of websites.

Its analysis found AI referrals represented less than 1% of referral traffic and that conventional organic search continued to produce substantially more conversions. BrightEdge characterised AI search primarily as a research and discovery channel rather than a direct conversion channel in the data it observed.

These findings are not necessarily contradictory.

They may reveal something more important:

There is no universal AI-search conversion rate.

Performance depends on industry, product complexity, purchase value, user intention, brand recognition, type of query, stage of decision-making and how accurately attribution systems capture AI influence.

A software buyer investigating an expensive technical platform behaves differently from someone looking up tomorrow’s weather.

Someone commissioning a consultant behaves differently from someone purchasing groceries.

Someone researching an artist for a significant acquisition behaves differently from someone casually browsing images.

AI discovery therefore needs to be understood at the level of intent, not simply traffic.

AI May Influence Purchases It Never Receives Credit For

There is another complication.

Much of AI’s commercial influence may be invisible.

Imagine this journey.

A potential client asks an AI assistant:

Which consultancies specialise in helping creative organisations develop commercially sustainable projects?

The AI introduces several organisations.

One sounds interesting.

The user remembers the name.

Twenty minutes later they open Google and search directly for that organisation.

They visit the website.

They make an enquiry.

Traditional analytics may record:

Organic search → enquiry.

But that is incomplete.

The actual journey was:

AI discovery → brand recognition → Google verification → website → enquiry.

The AI influenced the decision without receiving attribution.

This creates an emerging AI attribution gap.

Conventional analytics can record the final visible interaction while missing the conversation that shaped the customer’s decision.

The last click is not necessarily the source of the decision.

AI Search Does Not Mean SEO Is Dead

It would be a serious strategic mistake to interpret these developments as evidence that organisations should abandon traditional search optimisation.

The emerging system is interconnected.

Research into ChatGPT referral behaviour indicates that users frequently move between AI systems, conventional search engines and websites rather than simply replacing one with another.

Google itself argues that AI-enhanced search is generating new kinds of queries and increasingly complex forms of exploration.

At the same time, independent research demonstrates that AI summaries can suppress conventional clicks for some searches.

Both developments can coexist.

Some informational searches may generate fewer clicks.

From Ranking to Recommendation

Traditional SEO asks:

Where do we rank?

AI discovery introduces another question:

When does the system mention us?

And another:

How does it describe us?

And another:

What evidence is it using?

These are fundamentally different questions.

Imagine a consultancy appearing first in Google for:

creative business consultancy

but being absent when someone asks an AI:

Which organisations combine cultural intelligence, business strategy, creative development and responsible AI?

The organisation may possess search visibility while lacking machine-mediated relevance.

Conversely, an organisation that is not number one for a broad keyword might repeatedly appear in highly specific AI answers because its expertise is unusually well documented.

That could become extremely valuable.

The future competition may therefore be less about owning individual keywords and increasingly about owning conceptual territory.

Your Website Is Becoming Evidence

This may be one of the most important strategic changes.

For years, businesses were encouraged to think of websites primarily as destinations.

AI search encourages us to think of them as something else as well:

evidence repositories.

Every strong page helps establish what an organisation knows.

Every case study establishes what it has done.

Every biography helps establish who is responsible.

Every methodology explains how it works.

Every original research project establishes intellectual authority.

Every transparent service page clarifies what it sells.

Every well-documented project provides evidence.

Every useful article strengthens the relationship between the organisation and particular areas of knowledge.

That information serves humans directly.

But it can also help machines understand the organisation.

The website therefore becomes both:

a place people visit

and

a body of evidence from which systems construct understanding.

That is a significant conceptual shift.

Generic Content Is Becoming Less Valuable

Artificial intelligence can produce enormous quantities of competent generic information.

That makes generic information less strategically distinctive.

An organisation publishing another article called:

Ten Ways AI Can Improve Your Marketing

is contributing to an already saturated information environment.

An organisation publishing:

What We Learned From 47 AI-Assisted Marketing Audits of Independent Cultural Organisations

is contributing something different.

It possesses evidence that does not exist elsewhere.

The same applies to original research, proprietary datasets, case studies, experiments, methodologies, field observations, interviews, first-person experience, local knowledge, community knowledge, technical testing and carefully documented failures.

AI can summarise existing knowledge extraordinarily well.

What it cannot legitimately replace is experience that has not previously been documented.

This changes the economics of content.

The advantage increasingly belongs not to whoever publishes the most material, but to whoever contributes the most useful new knowledge.

Experience Is Becoming a Search Asset

There is an interesting paradox here.

The more synthetic information becomes available, the more valuable genuine experience may become.

Consider the difference between these two statements:

Community engagement is important when designing cultural programmes.

and:

During our community programme, the first consultation failed because we invited participation only after the central concept had already been decided. We redesigned the second session so participants could influence the premise itself.

The first is information.

The second is evidence.

It contains experience, failure, learning, specificity and judgement.

Those characteristics are difficult to commoditise.

Authority Will Increasingly Exist Outside Your Website

There is another consequence.

Organisations cannot completely control what AI systems learn about them.

Their own website matters.

But so does the wider information environment.

That can include journalism, industry publications, professional directories, research papers, interviews, podcasts, videos, reviews, public databases, conference programmes, community discussions and partnership websites.

This creates a distinction between:

what an organisation says about itself

and

what the wider information ecosystem says about it.

The second can become extremely powerful.

AI visibility therefore cannot simply become another exercise in manipulating webpages.

It is partly a reputation problem.

Real authority must increasingly exist beyond the organisation’s own claims.

The Rise of Citation-Worthy Organisations

This leads to a useful strategic idea.

Businesses traditionally asked:

How do we make this page rank?

Increasingly they should also ask:

Why would anyone cite this?

That question produces better content.

A citation-worthy page normally contains something worth referencing:

a useful definition;

a distinctive framework;

original research;

credible statistics;

a documented methodology;

a case study;

an expert explanation;

a unique dataset;

a strong visualisation;

a first-hand observation;

or a genuinely useful answer.

The objective is not to trick an AI system into mentioning the organisation.

It is to become the kind of source that deserves to be mentioned.

That distinction is critical.

Marketing Is Moving from Traffic Acquisition to Trust Architecture

For much of digital marketing history, the website visitor was treated as the scarce resource.

More visitors usually meant more opportunities.

AI discovery complicates that equation.

Suppose one organisation receives 100,000 poorly qualified visitors.

Another receives 10,000 highly informed visitors.

Traffic alone tells us very little about commercial value.

This means marketing measurement increasingly needs to consider qualified enquiries, conversion rates, customer acquisition, revenue, brand searches, AI referrals, assisted conversions, repeat visits, citation visibility, brand mentions and eventually AI-influenced revenue.

The central metric changes from:

How many people arrived?

towards:

How much qualified demand did our information create?

That is a much more useful question.

The Website After the Answer

If AI increasingly handles the early stages of explanation, websites must become better at what happens next.

A visitor arriving from an AI recommendation may not need another 2,000 words explaining the basics.

They may need confirmation.

That means the website must answer:

Are these people credible?

Can they solve my particular problem?

Have they done something similar before?

What is their methodology?

What does working with them involve?

What evidence supports their claims?

What does it cost?

What happens next?

The role of the website gradually shifts.

From:

Tell me who you are.

Towards:

Prove that the recommendation I received was justified.

That makes case studies, methodology pages, testimonials, transparent offers, diagnostics, demonstrations, portfolios and evidence increasingly important.

The Cultural Intelligence Question

There is another dimension that conventional marketing analysis can easily miss.

AI systems do not simply organise commercial information.

They increasingly mediate cultural representation.

When someone asks:

Who are the important contemporary Black British artists?

or:

Which cultural organisations are doing meaningful community-led work in Nottingham?

or:

What are examples of innovative African diaspora creative businesses in Britain?

the answer is not culturally neutral.

Someone or something is included.

Someone is absent.

Some sources are considered authoritative.

Others are invisible.

Some histories have abundant digital documentation.

Others survive through local, oral or community knowledge that may be poorly represented online.

AI discovery therefore raises questions about:

representation;

visibility;

authority;

language;

power;

participation;

and whose knowledge becomes machine-readable.

This is particularly important for cultural organisations.

An organisation may possess enormous cultural importance within a community while having relatively little structured digital evidence describing that importance.

AI systems cannot reliably surface knowledge that has never been documented or that exists in sources they cannot access.

The emerging AI-discovery economy therefore creates both an opportunity and a risk.

The opportunity is to make overlooked cultural knowledge more visible.

The risk is that existing digital inequalities become reproduced inside AI-generated answers.

From Search Engine Optimisation to Knowledge Infrastructure

The deeper strategic change may therefore be larger than SEO.

Organisations increasingly need to build what could be called knowledge infrastructure.

That means deliberately documenting:

who they are;

what they believe;

what they know;

what they have learned;

what they have created;

who they serve;

how they work;

what evidence exists;

where their knowledge came from;

and what makes their perspective distinctive.

This knowledge should not exist only for machines.

It should primarily exist because it is useful to people.

But when organised clearly, accurately and consistently, it becomes useful to both.

The organisation becomes easier to:

discover;

understand;

verify;

cite;

recommend;

and trust.

What Should Organisations Do Now?

The answer is not to abandon everything and chase the latest acronym.

The fundamentals remain remarkably familiar.

Create genuinely useful things.

Document them properly.

Make expertise visible.

Build relationships.

Earn independent recognition.

Answer real questions.

Provide evidence.

Maintain technical accessibility.

Understand your audiences.

Measure outcomes rather than vanity metrics.

But several additional disciplines now matter.

Businesses should begin auditing the questions customers might ask AI systems before discovering them.

They should investigate whether their organisation appears in relevant AI-generated answers.

They should examine how accurately those systems describe them.

They should identify which sources appear repeatedly in their category.

They should develop original, citation-worthy knowledge rather than endlessly rewriting generic information.

They should strengthen the relationship between articles, services, case studies, research and expertise.

And they should improve attribution so AI-assisted journeys do not disappear into categories such as direct or organic traffic.

Most importantly, organisations should stop treating AI visibility as a technical trick.

It is increasingly an authority problem.

The New Customer Journey

The old model looked something like:

SEARCH → CLICK → WEBSITE → RESEARCH → DECISION

The emerging model can look very different:

QUESTION → AI INTERPRETATION → COMPARISON → RECOMMENDATION → VERIFICATION → WEBSITE → DECISION

And sometimes:

QUESTION → AI ANSWER → DECISION

The number of visible website visits may therefore decline for some kinds of information even while digital influence continues to grow.

That requires marketers to distinguish between:

traffic

and

influence.

The two are no longer synonymous.

The Most Important Strategic Shift

The great mistake would be to reduce this transformation to another marketing acronym.

SEO.

AEO.

GEO.

LLMO.

Whatever terminology eventually survives matters less than the underlying behavioural change.

People are learning to ask machines to help them think.

They are asking them to:

research;

compare;

interpret;

shortlist;

explain;

evaluate;

and recommend.

That means organisations are entering an environment in which they are increasingly discovered through machine-mediated judgement.

The response should not be to manufacture content designed only for machines.

It should be almost the opposite.

Create information so useful, distinctive, credible and well evidenced that both humans and machines can understand why it matters.

The Answer Is Becoming the Marketplace

For twenty years, digital businesses fought for positions on pages of search results.

The next competition may be considerably more consequential.

It is the competition to become part of the answer.

Not merely visible.

Relevant.

Not merely ranked.

Recommended.

Not merely visited.

Trusted.

And not merely capable of attracting attention, but capable of providing enough evidence that when an intelligent system attempts to answer someone’s question, the organisation belongs naturally within that answer.

The organisations that understand this will stop asking only:

How do we get more traffic?

They will begin asking:

What knowledge do we possess that deserves to travel?

What evidence makes us credible?

What questions should we be known for answering?

Where does independent evidence of our authority exist?

What happens when someone asks an AI about the problem we solve?

And perhaps most importantly:

If the answer becomes the marketplace, have we given the marketplace enough reason to remember us?

That is the challenge AI search places before marketers.

But it is also the opportunity.

Because the emerging contest for visibility may ultimately reward something considerably more valuable than keyword optimisation:

real knowledge, demonstrated experience, cultural relevance, independent authority and earned trust.