Bots Now Outnumber Humans: What Marketers Need to Measure in the AI Search Era

For years, digital marketing measurement has been built around a simple assumption: a person searches, clicks, lands on a website, browses, converts, or leaves.

That model is broken, sort of…

Cloudflare data now shows that automated traffic accounts for the majority of web HTTP requests. In practical terms, more of the web is being requested by bots, crawlers, AI systems, assistants and agents than by traditional human browsing sessions. This post not about arguing what is right or wrong, but exposing this reality.

That does not mean humans no longer matter. It does not mean every bot is bad. And it does not mean website analytics are suddenly useless. But it does mean marketers need to rethink what their numbers are actually showing.

A pageview may not be a person.
A session may not represent human interest.
A drop in organic clicks may not mean a drop in influence.
A brand mention in an AI answer may matter even when no visit happens.
A paid search click may arrive later in the journey, after an AI assistant has already shaped the shortlist.

This is the quiet earthquake for marketing. The web is no longer only a place where humans browse. It is becoming a place where machines research, summarise, compare and recommend on behalf of humans.

That changes what we need to measure.

The web is becoming machine-readable before it is human-clicked

The rise of AI agents changes the role of the website.

Historically, a website was the destination. The goal of SEO, PPC, content and social was to bring people to the site, where they would read, compare, trust and convert.

That still happens. But increasingly, the website is also an input. AI systems read it, extract information from it, compare it with other sources and use it to generate answers somewhere else.

The user may never visit the page. They may ask ChatGPT, Gemini, Perplexity, Claude, Google AI Mode or another AI assistant to compare providers, explain a topic, shortlist vendors, summarise reviews or recommend what to do next.

In that journey, the brand may still influence the decision, but not through the traditional route of impression, click, session and conversion.

The question for marketers is no longer only:

Are people visiting our website?

It is also (and most importantly):

Are AI systems able to understand, trust and use our content when people ask relevant questions?

That is where SEO and GEO start to overlap.

What is GEO, and how is it different from SEO?

GEO usually stands for Generative Engine Optimisation. It means improving the chances that your brand, content, products, services or expertise appear in AI-generated answers.

SEO focuses on visibility in search engines.

GEO focuses on visibility inside AI-generated responses. You could argue it’s the new SEO. Noone will sell you SEO services anymore, it has evolved.

The two are connected, but they are not identical.

Traditional SEO asks:

• Are we ranking for relevant queries?
• Are we earning impressions and clicks?
• Are our pages indexed and technically healthy?
• Are users engaging with our content?
• Are organic visitors converting?

GEO adds a different set of questions:

• Are we mentioned in AI-generated answers?
• Are we cited as a source?
• Are our claims being extracted accurately?
• Are AI systems associating us with the right topics?
• Are we visible in comparison and recommendation-style queries?
• Are we present when users ask for advice, not just when they search for keywords?

This is why marketers need to move beyond “rankings and traffic” as the only measure of organic performance.

In AI search, visibility can happen without a click.

Why traditional analytics are becoming less reliable on their own

Website analytics were designed around human behaviour.

Sessions, users, bounce rates, time on site, pageviews and conversion paths all assume that a visitor is meaningfully interacting with the website.

That assumption is weaker now.

Some automated traffic is easy to filter. Some is not. AI agents may fetch content, render pages, behave like browsers, use residential IPs, or act on behalf of real users. Some bots are useful. Some are harmful. Some are somewhere in between.

This creates three measurement problems.

First, traffic quality becomes harder to interpret. A rise in sessions may not mean a rise in human demand.

Second, engagement metrics become noisier. Bounce rate, time on page and page depth are less useful if a growing share of visits are non-human or machine-assisted.

Third, influence becomes harder to see. AI systems may use your content in an answer, but the user may not click through to your website.

This does not mean marketers should ignore analytics. It means analytics need to be segmented, cleaned and combined with other signals.

The old dashboard is not enough.

What marketers should measure now

The measurement model needs to shift from traffic-first to influence-and-outcome-first.

That does not mean abandoning sessions, clicks or rankings. It means putting them in context.

A more useful measurement framework has five layers:

1. Human traffic quality
2. Bot and AI crawler activity
3. Traditional SEO performance
4. GEO and AI answer visibility
5. Commercial impact

Each layer answers a different question.

Together, they give a more realistic view of performance in the AI search era.

1. Measure human traffic quality, not just traffic volume

The first step is to separate traffic from valuable traffic.

Marketers should stop treating all sessions as equal. A session from a qualified prospect, a returning buyer, a sales-ready lead or a target account is not the same as a bot request, accidental visit or low-intent click.

Useful metrics include:

• Engaged sessions
• Conversion rate by source
• Qualified leads by source
• Revenue or pipeline by landing page
• Returning visitors from target segments
• Assisted conversions
• CRM-matched visitors where available
• Target account engagement for B2B brands

The goal is to understand whether real demand is growing, not just whether traffic is moving.

This matters because AI search may reduce some informational clicks. If users get simple answers directly from AI tools, fewer people may visit basic educational pages. But the people who do click may be more informed, more qualified or further along in the journey.

A decline in organic sessions is not always a decline in value.

The better question is:

Are the right people still finding us, trusting us and converting?

2. Measure bot and AI crawler activity separately

Marketers should also start treating bot traffic as its own reporting category.

This does not mean every marketing team needs to become a cybersecurity team. But if automated traffic is now a large share of web activity, it cannot be ignored.

At a practical level, teams should work with analytics, web, SEO or security colleagues to understand:

• How much traffic is clearly automated?
• Which known crawlers are visiting the site?
• Which AI crawlers are requesting content?
• Are important pages being crawled?
• Are bots creating server load or analytics noise?
• Are bad bots being blocked without blocking useful crawlers?
• Are robots.txt and CDN rules aligned with the marketing strategy?

This is where marketing and technical teams need to work more closely.

Blocking all bots may protect content, but it may also reduce visibility in AI systems. Allowing all bots may increase exposure, but it can also increase scraping, cost and risk.

The answer is not “block everything” or “allow everything”.

The answer is governance.

Brands need a clear view of which bots they want to allow, which they want to restrict, and which pages should be accessible to AI systems.

3. Measure SEO performance beyond clicks

SEO measurement is changing because clicks no longer tell the full story.

A page can be highly influential in search without receiving the same number of visits it would have received five years ago. AI Overviews, featured snippets, knowledge panels, zero-click results and AI Mode can all satisfy part of the user journey before a website visit happens.

That means SEO performance should include both visibility and outcomes.

Core SEO metrics still matter:

• Impressions
• Clicks
• Click-through rate
• Average position
• Indexed pages
• Ranking keywords
• Organic conversions
• Organic revenue or pipeline
• Landing page performance
• Technical health
• Internal linking strength

But marketers should interpret them differently.

If impressions rise while clicks fall, the answer may not be that SEO is failing. It may mean the content is appearing in more answer-led search experiences where fewer users click.

If clicks fall but conversions stay stable, organic traffic may be becoming smaller but more qualified.

If rankings are stable but leads decline, the issue may be demand, messaging, offer, landing page quality or AI answer displacement.

SEO reporting needs more diagnosis and less dashboard reading.

The key question is:

Is organic search still creating discoverability, trust and commercial value, even when the click path changes?

4. Measure GEO and AI answer visibility

GEO measurement is still emerging, but marketers should not wait for perfect tools.

The most practical approach is to create a repeatable set of AI search tests.

Start with the questions your audience is likely to ask AI tools:

• “What is the best solution for [problem]?”
• “Compare [category] providers”
• “Best [service] for [audience]”
• “How do I solve [specific pain point]?”
• “What should I consider before choosing [product/service]?”
• “Alternatives to [competitor]”
• “Who are the leading companies in [category]?”
• “What are the risks of [approach]?”
• “How much does [solution] cost?”
• “What questions should I ask before buying [service]?”

Then test those questions across the AI environments your audience is likely to use, such as Google AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini, Claude and Copilot.

Track:

• Whether your brand is mentioned
• Whether competitors are mentioned
• Whether your website is cited
• Which pages are cited
• Whether the answer describes your offer accurately
• Which topics are associated with your brand
• Whether your differentiators appear
• Whether AI tools recommend you, ignore you or misrepresent you
• How answers change over time
• How answers differ by query wording

This should be done regularly, not once.

AI answers are dynamic. They can change by model, location, freshness, phrasing, source availability and user context. The goal is not to create a perfect ranking report. The goal is to understand whether your brand is visible and accurately represented in the places where decisions are increasingly being shaped.

A simple monthly GEO scorecard could include:

• Share of AI answers mentioning the brand
• Share of AI answers citing the website
• Share of answers where competitors appear
• Accuracy of brand description
• Presence of key messages
• Presence in comparison queries
• Presence in commercial-intent queries
• Pages most often cited
• Gaps where competitors appear but you do not

This gives marketing teams a practical starting point.

5. Measure brand demand, not just website demand

If AI search reduces clicks, brand demand becomes even more important.

A user may discover your brand through an AI answer, a LinkedIn post, a podcast, a review, a community discussion or a comparison summary, then later search your brand directly.

That journey may not show up neatly in analytics.

This is why marketers should track signals such as:

• Branded search volume
• Direct traffic quality
• Returning visitor growth
• Brand plus category searches
• Brand plus competitor searches
• CRM source notes
• Sales call mentions
• Self-reported attribution
• Review site visibility
• Social mentions
• Community mentions
• AI answer mentions

None of these is perfect on its own. Together, they show whether the brand is becoming more present in the market.

In an AI-mediated journey, the brand that gets remembered may win even if it does not win the first click.

How to understand GEO and SEO performance together

SEO and GEO should not be managed as separate silos.

A useful way to think about it is this:

SEO helps your content become discoverable. GEO helps your content become usable by AI systems.

That means your content needs to work on two levels.

For humans, it needs to be clear, useful, credible and persuasive.

For machines, it needs to be structured, consistent, specific and easy to extract.

The same page should ideally help both.

That means marketers should create content that includes:

• Clear definitions
• Direct answers to common questions
• Specific claims supported by evidence
• Comparison sections
• Pricing or decision criteria where relevant
• Author or company expertise
• FAQs
• Internal links to related topics
• Structured data where appropriate
• Consistent terminology
• Clear product or service descriptions
• Original insight, not generic summaries

This is not about writing for robots at the expense of humans.

It is about removing ambiguity.

If your website does not clearly explain what you do, who you help, why you are different, what problems you solve and when someone should choose you, AI systems will struggle to represent you accurately.

So will humans.

What this means for content strategy

The old content playbook was often built around keyword volume.

Find the keyword. Write the article. Rank the page. Win the click.

That model still has value, but it is incomplete.

AI search rewards content that answers complex, multi-step, comparison-led and decision-led questions. It is less about isolated keywords and more about topical usefulness.

Instead of only asking, “What keywords should we rank for?”, marketers should ask:

• What questions does our audience ask before they know the category?
• What questions do they ask when comparing options?
• What misconceptions do they have?
• What information would help an AI system explain our value accurately?
• What proof do we have that competitors do not?
• What should we be known for?
• What should AI systems confidently associate with our brand?

The best content will not just chase traffic.

It will shape understanding.

That is a different job.

Implications for paid search ads

Paid search is not going away.

But its role is changing.

If AI systems increasingly answer informational queries, compare vendors and summarise options, then paid search may become less about capturing every research click and more about capturing high-intent moments, defending brand demand and reinforcing trust after AI-led discovery.

There are several implications.

1. Informational clicks may become less predictable

Queries that once generated blog traffic may now be answered directly inside AI experiences.

This could reduce paid and organic opportunities around broad informational searches, especially where the answer is simple.

For paid search, this means advertisers should be careful about paying for clicks that AI experiences have already partially satisfied.

The focus should shift towards queries where the user still needs to act, compare, buy, book, request, evaluate or speak to someone.

2. Brand search becomes more strategically important

If users discover brands through AI answers, they may later search directly for the names they remember.

That makes brand search more valuable, not less.

CMOs and paid media teams should monitor:

• Branded search volume
• Brand search conversion rate
• Competitor bidding on brand terms
• Brand plus review searches
• Brand plus pricing searches
• Brand plus alternative searches
• Brand plus competitor searches

Paid search may increasingly act as the bridge between AI-led discovery and conversion.

If AI tools put your brand into the consideration set, paid search needs to help capture that demand when it appears.

3. Search campaigns need to reflect questions, not just keywords

AI search is training users to ask longer, more specific and more conversational questions.

That has implications for search query analysis.

Paid search teams should pay attention to the language users use when they are closer to a decision. Search terms may reveal new problems, objections, comparison points and use cases that should feed both ad copy and content strategy.

This is where paid search becomes a research tool.

Search query data can help identify:

• New customer pains
• Competitor comparisons
• Objections
• Feature demand
• Pricing questions
• Category confusion
• Emerging use cases

Those insights should not stay inside Google Ads.

They should feed SEO, GEO, landing pages, sales enablement and product messaging.

4. Landing pages need to answer more of the decision

If a user arrives after using an AI assistant, they may already have a basic understanding of the category.

They may not need another generic introduction.

They may need proof, clarity and confidence.

That means landing pages should become more decision-focused. They should answer:

• Who is this for?
• What problem does it solve?
• How is it different?
• What proof supports the claim?
• How does it compare with alternatives?
• What happens next?
• What are the risks of doing nothing?
• What does pricing depend on?
• Why should someone trust this provider?

Paid search traffic becomes more valuable when the page continues the decision journey instead of restarting it.

5. Measurement needs to connect paid search with AI visibility

Paid search teams should not measure performance in isolation from SEO and GEO.

If AI answers are increasing brand awareness, paid search may see more branded demand. If organic clicks decline but brand search rises, that may indicate that discovery is happening elsewhere. If paid search conversion rates improve while top-of-funnel traffic declines, the journey may be becoming more compressed.

This is why paid media reporting should include context from:

• Google Search Console
• AI visibility tracking
• Brand search trends
• CRM data
• Self-reported attribution
• Search term reports
• Landing page conversion quality
• Sales feedback

The question is not just whether paid search is converting.

The question is what role paid search plays in a journey where AI may have influenced the user before the click.

A practical measurement framework for the AI search era

Marketers need a new reporting rhythm.

A practical monthly dashboard could include:

Human demand

• Qualified organic traffic
• Engaged paid search traffic
• Direct traffic quality
• Returning users
• Conversion rate by source
• Revenue or pipeline by source

SEO visibility

• Search impressions
• Organic clicks
• CTR by query group
• Ranking changes
• Indexed pages
• Pages gaining or losing visibility
• Non-brand vs brand performance

GEO visibility

• Brand mentions in AI answers
• Website citations in AI answers
• Competitor mentions
• Accuracy of AI-generated descriptions
• Share of answers where the brand appears
• Pages most often cited
• Topics associated with the brand

Bot and crawler activity

• Known crawler visits
• AI crawler activity
• Blocked vs allowed bots
• Server load from automated traffic
• Important pages crawled
• Analytics noise from bots

Paid search impact

• Brand search volume
• Non-brand conversion quality
• Search terms showing comparison intent
• CPA or ROAS by intent group
• Landing page conversion quality
• Assisted conversions
• CRM lead quality by campaign

Commercial outcomes

• Qualified leads
• Opportunities
• Pipeline
• Revenue
• Customer acquisition cost
• Payback period
• Sales feedback

This framework recognises that marketing impact is no longer contained inside one platform.

SEO, GEO, paid search, brand and website analytics need to be read together.

The strategic shift: from traffic acquisition to answer influence

The biggest change is not technical. It is strategic.

For years, marketers have been trained to think in terms of traffic acquisition.

More impressions. More clicks. More sessions. More visitors.

Those metrics still matter, but they are no longer enough.

The new challenge is answer influence.

When an AI system responds to a buyer’s question, does your brand appear?
Is your expertise represented accurately?
Are your differentiators clear?
Are competitors being recommended instead?
Is your content good enough to be cited, summarised and trusted?
Does paid search capture the demand created by that AI-led discovery?

This is where the advantage will be built.

Not by abandoning SEO.
Not by overreacting to bots.
Not by chasing every new AI tool.

But by making marketing assets clear enough for machines to understand and strong enough for humans to trust.

Final thought

The rise of bots and AI agents does not mean human marketing is over.

It means human attention is being mediated differently.

People still have needs, problems, budgets, preferences and objections. They still choose brands. They still buy from companies they trust.

But increasingly, their first layer of research may be done by an AI system.

That changes the job of marketing.

Your website still needs to persuade people.

But it also needs to be understood by machines.

Your SEO still needs to drive visibility.

But it also needs to support AI answer inclusion.

Your paid search still needs to capture demand.

But it also needs to respond to journeys that may have started inside an AI assistant.

The brands that win will not be the ones that optimise only for humans or only for machines.

They will be the ones that do both.

Clear for AI.
Useful for humans.
Measurable against business outcomes.

That is the new search challenge.