AI Search Sends Fewer Clicks: What's the Business Case for Optimizing for It Anyway?

The business case in one view

Admit the click loss first

Any credible business case starts by conceding the cost. The best independent data comes from Pew Research Center, which tracked the real browsing of 900 U.S. adults across 68,879 Google searches in March 2025. When an AI summary appeared, users clicked a traditional result on 8% of visits. Without a summary, they clicked on 15% (Pew Research Center). Links inside the summary itself were clicked on only 1% of visits.

Users also ended their browsing session more often after seeing an AI summary: 26% of the time, compared with 16% on pages with only traditional results.

Two qualifications matter when you repeat these numbers.

• The study is observational. AI summaries tend to appear on longer, question-style queries, which may have had lower click rates anyway. "Pages with AI summaries saw about half the click rate" is accurate. "AI Overviews cut clicks in half" overstates what was measured.
• The rates are per visit, not per user, and they are dated. They cover Google in the U.S. in March 2025, before later AI Overview and AI Mode expansion.

Google disputes the findings. It calls Pew's methodology flawed and says it sends "billions of clicks" to websites daily, but it has not published counter-data. Search Console does not separate AI Overview clicks from other clicks, so site owners cannot check the claim directly (Search Engine Land). Google's related claim that AI Overviews drive more queries is about Google's usage. It says nothing about traffic to your site.

Why fewer clicks does not mean less value

Clicks were never the whole value of search. In the same Pew study, around two-thirds of all Google searches ended without a click on any result, with or without an AI summary. Much of search's value was already being delivered on the results page, through what people saw and remembered.

The useful way to frame the business case is as an equation. It is an editorial framework, not a research finding:

Search value = influence (being seen and named) + clicks.

The click component is shrinking. The evidence below suggests the influence component is growing, and it is concentrated at the moment that matters most for revenue: when a buyer decides who to consider.

Where AI answers influence revenue: the shortlist

Buyers now use AI at the comparison stage, before they ever visit a vendor site. Three independent B2B surveys point in the same direction.

AI is the top named influence on vendor shortlists. G2's survey of 1,100 B2B decision-makers ranked GenAI chatbots first, at 17.1%. Review sites came second at 15.1%, and vendor websites third at 12.8% (G2). This is a plurality, not a majority. It still means AI now edges out every other single source.

Nearly all business buyers use AI somewhere in the purchase. Forrester reports that 94% of business buyers use AI in the buying process. Twice as many buyers as a year earlier named generative AI or conversational search as more meaningful than any other source. Forrester's advice to marketers is to shift "from driving traffic to driving visibility" (Forrester).

The shortlist decides the winner. 6sense surveyed about 4,000 B2B buyers. On average they evaluated 5.1 vendors, and they already knew 3.9 of them. The vendor contacted first won 77% of the time (6sense). Buyers mainly used LLMs mid-journey, to compare and synthesize options.

Taken together, the stakes are simple. If buyers compare vendors inside AI answers, and the vendors on the day-one list usually win, then absence from those answers means absence from the deal. That happens whether or not anyone clicks.

One caution applies: each survey sponsor has a commercial interest. G2 benefits when companies invest in review sites. 6sense sells tools for tracking anonymous buyer research. Forrester sells the underlying report. The findings deserve weight because they converge, but name the sponsors when you present them.

What the evidence does not support

The strong case for AI visibility is easy to oversell. Three popular claims go further than the data.

"SEO is dead." 6sense found that buyers interacted with the winning vendor about 16 times, a figure unchanged from 2023 despite 94% LLM use. G2's SMB respondents still rely most on Google (47%). AI answers also draw on search-indexed pages, so traditional search work still feeds AI visibility. The overlap is explored in Is AEO Really a New Discipline?.

"Lost traffic equals lost pipeline." 6sense argues that the traffic going missing is probably the "professionally curious" and future buyers rather than people in market now. That is plausible, and useful in a CFO conversation. It is the researcher's interpretation, not a measured result, and future buyers still matter for long-term demand.

"AI referral traffic converts several times better." Vendor multiples like this circulate widely, but the evidence here does not verify them. They are also prone to selection effects, since people who click through from an AI answer may already intend to act. Leave them out of a business case unless your own data shows them.

Decide whether AI visibility deserves budget for your business

AI visibility matters most where AI answers actually appear and where buyers compare options. Pew found AI summaries on 18% of Google searches in March 2025. The share varied sharply by query type: 8% of one- or two-word searches versus 53% of searches with ten or more words, and 60% of queries starting with who, what, when, or why.

The grid below turns the evidence into a decision guide. The factors come from the sources above. Their combination into a grid is an editorial judgment.

Incumbents still have a reason to act, just a different one. If buyers already know you, the job is defense. Check that AI answers describe you accurately and do not recommend a competitor in your place for your core use cases.

The consumer and local case is less proven. The buyer surveys cover B2B, mostly software and high-value deals. For consumer or local businesses, rely on your own query data rather than borrowing these percentages.

Put part of the budget off your own site

AI answers lean heavily on third-party sources. In Pew's data, Wikipedia, YouTube, and Reddit together made up 15% of the sources cited in Google AI summaries, and 88% of summaries cited three or more sources. Government sites made up 6% of AI summary sources, compared with 2% in standard results.

Pair that with G2's finding that review sites are the second-largest shortlist influence, and a budget implication follows. Some AI visibility spend belongs in PR, review profiles, community participation, and inclusion in the comparison articles AI already cites, not only in content on your own domain. Pew's citation data covers Google only, so patterns in ChatGPT or Perplexity may differ. Check which sources your target engines actually cite before committing spend.

Whether paying for placement in third-party roundups counts as legitimate optimization is its own debate, covered in Are Third-Party Listicles Good GEO or Just AI-Era Link Building?.

Measure AI visibility as four separate signals

AI influence mostly leaves no referral trail, so a click-based ROI model will always undercount it. A single "visibility score" has the opposite problem: it can look like progress without proving any business result. The honest approach is to track four signals separately and be clear about what each one proves.

These distinctions are not pedantry. A bot request does not establish indexing, citation, or a conversion, and referral tracking only works when the visit carries identifiable source information (Tideflow AI; Tideflow AI). Treat identifiable AI referrals as the floor of AI's value, not the ceiling.

Presence data also has limits. AI answers vary between runs, which is why sampling repeated prompts beats a single check. And Forrester found that 61% of buyers use private AI tools provided by their organization. Monitoring public engines captures a sample of the buyer experience, not all of it.

A test you can run this week: write down 10 questions a buyer asks when comparing solutions in your category, such as "What's the best [category] tool for [use case]?" or "[Your brand] vs [competitor]." Ask each one in ChatGPT, Perplexity, and Google on three separate days. Record whether you appear, which competitors appear instead, and which sources get cited. If competitors show up consistently and you don't, you have found a concrete shortlist gap. If the answers keep citing the same review site or article, you have found where to spend first.

Present the case to a CFO without overclaiming

A finance leader will trust a case that concedes the downside. Build it in this order:

1. Concede the click loss. Pages with Google AI summaries saw roughly half the click rate in Pew's March 2025 data. Don't hide it.
2. Reframe the value. Around two-thirds of searches were already zero-click. The question is whether buyers see and consider you, not how many visit.
3. Tie it to revenue mechanics. Buyers rank AI as the top shortlist influence (G2). The vendor contacted first wins 77% of the time (6sense). Absence from AI answers is a shortlist risk.
4. Name the sources' interests. G2, 6sense, and Forrester each benefit commercially from these conclusions, but their findings agree.
5. Commit to layered measurement. Report presence, discovery, AI referrals, and outcomes separately, and say plainly that referral numbers are a lower bound.
6. Scope the bet. Start with the query types and buyer segments where the grid above says the case is strongest, and expand only if the signals move.

Where a connected platform helps

The hard part of this business case is not any single signal. It is connecting them. Knowing that a competitor appears in answers where you don't is only useful if you can see which sources the AI cited, decide what content or third-party placement would change that, and then check whether crawler visits, referrals, and conversions moved afterward. Many teams stitch that together from a monitoring tool, spreadsheets, a CMS, and web analytics, and lose the thread between steps.

Tideflow AI is built to keep that chain in one place. It monitors prompts daily across ChatGPT, Gemini, Perplexity, Google AI Mode, and Google AI Overview, showing where a brand is mentioned, cited, and ranked, and which competitors and sources appear instead. It turns those gaps into plans to improve a page, create content, or get included in cited third-party sources. It then logs AI crawler visits and ties identifiable AI referrals to conversions by landing page, keeping the four signals distinct rather than blending them into one score.

Two limits are worth stating plainly. Tideflow AI is a newer entrant in early access without independent reviews yet, and no tool, including this one, can guarantee that an AI system will cite or recommend a brand. For how it compares with established alternatives, see Tideflow AI vs Profound and Tideflow AI vs Peec AI.

Frequently Asked Questions

Can you trust AI visibility scores?

AI visibility scores are reliable only as a sample, and only when built from repeated prompts across several engines. AI answers vary between runs, and a large share of enterprise buyers use private AI tools that public monitoring cannot see. A score also says nothing about clicks or revenue. Use it as the presence signal, alongside crawler, referral, and outcome data.

Should you cut traditional SEO budget to fund AI visibility?

Cutting traditional SEO to fund AI visibility is usually premature. AI answers draw on search-indexed and third-party sources, SMB buyers still rely most on Google, and vendor interactions have held steady despite near-universal AI use. For most teams the better move is reallocating within existing search and content budgets toward comparison-style content and third-party presence, rather than abandoning SEO.

How long before AI visibility work shows results?

The supplied evidence does not establish a reliable timeline for AI visibility results. The practical approach is to set a baseline for all four signals before making changes, then check whether presence and discovery move first, followed by referrals and outcomes. Crawler visits can confirm that AI systems have found new content, but they do not confirm citation.

Your next step

Start with the 10-question test above to see whether your brand appears where buyers compare options in your category. If it doesn't, and your buyers fit the "invest more" side of the grid, you have the core of your business case. The next step is closing specific gaps, which How to Get Mentioned by ChatGPT: A Tested, Step-by-Step Playbook walks through step by step. Judge the result by whether presence, discovery, AI referrals, and outcomes each move, not by clicks alone.

Sources

• Pew Research Center, Do people click on links in Google AI summaries? (July 2025)
• Search Engine Land, Google's AI Overviews are hurting clicks: Pew study (July 2025)
• G2, CMOs 2025 Buyer Behavior Report
• Forrester, B2B Buyers Make Zero-Click Buying Number One (January 2026)
• 6sense, The Impact of GenAI and LLMs on B2B Buyer Research
• Tideflow AI, Detect AI Bots; Add analytics and track views and conversions; AI Visibility & Content Platform