crontent

Google exposed AI search impressions. Founders still can’t prove what they did.

Crontent

Google finally gave Search Console an AI report, and it still ducks the question founders actually have. You can now see whether AI Overviews and AI Mode showed your pages. You still can’t see the queries, clicks, CTR, citation position, or pull the report through the API.

That matters because small SaaS teams don’t fund content with impressions. They fund content that helps revenue. If measurement stays this incomplete, the safest content budget is not generic traffic bait. It’s proof.

How do I measure AI search traffic and ROI?

You can only measure part of it today, and Google’s own setup leaves the most useful parts hidden. According to Try Aivo, Search Console’s Generative AI features report launched on 3 June 2026 in beta and shows impressions by page, country, device, and date. That’s the whole report.

What’s missing is the part a founder needs to defend budget. Try Aivo says there are no queries, no clicks, no CTR, no average position, no citation placement, and no API access. Data starts on 18 May 2026 with no backfill, so nobody gets year-over-year context until May 2027.

Google did solve one small problem. You no longer have to guess whether your site appears in AI Overviews or AI Mode. Spilt Media notes the report expanded from a UK-only beta to the US, India, and Switzerland and gives page, country, and device splits. But that still doesn’t answer the business question. Showing up is not the same as proving impact.

Generic top-of-funnel content gets harder to defend when the clicks disappear into blended reports

Google counts AI Overview and AI Mode clicks, but it blends them into the main Performance report instead of breaking them out. Try Aivo is blunt about the result: you can prove Google’s AI showed your page, but you can’t cleanly see what happened next.

GA4 doesn’t really rescue you. In GA4, those Google AI clicks arrive as google / organic, per Try Aivo. At the same time, GA4’s broader AI traffic handling is messy enough that teams often need custom channel rules just to see traffic from ChatGPT, Claude, Perplexity, Copilot, or Gemini. Terminus says GA4’s default channel grouping does not separate AI assistants from regular Referral or Organic Social, so the traffic stays invisible unless you build your own channel grouping.

That changes the math on fluffy awareness content. If you publish a generic explainer and the only clear number you get is AI impressions, you’re left arguing from vibes. Maybe it influenced pipeline. Maybe it helped branded search later. Maybe it assisted a deal. Good luck defending that to yourself, let alone a cofounder.

Proof-heavy pages survive bad attribution because they do more than chase search visibility

Pages tied to real buyer questions keep paying off even when tracking stays fuzzy. A comparison page can help a prospect choose you. An implementation writeup can shorten a sales call. A benchmark post can give your team something concrete to send in follow-ups. Docs can cut support load and answer objections before a demo.

Those pages still have search value. Search Console’s AI report breaks impressions down by page, which means you can at least see which specific assets show up in AI surfaces, per Spilt Media. That’s useful. But the stronger reason to fund them is that they also work outside search.

For a tiny SaaS team, the best bets are usually:

  • comparison pages tied to real alternatives buyers mention
  • implementation notes that show how the product works in practice
  • customer problem breakdowns with screenshots, workflows, or hard numbers
  • benchmark or methodology posts you can cite in sales and onboarding
  • docs and FAQs that answer the questions support keeps hearing

If Google never gives you clean AI attribution, these assets still earn their keep.

Judge content work by page-level business use, not AI impression growth

A content budget should buy reusable sales proof, not just screenshots from a half-finished Google report. The lack of query, click, CTR, and API data in Search Console’s AI report means agencies and freelancers can too easily hide behind rising impressions on AI surfaces, per Try Aivo.

Ask better questions instead:

  1. Which pages can sales send to active prospects?
  2. Which pages answer objections before a call?
  3. Which pages reduce support or onboarding time?
  4. Which pages can rank, get cited, and still matter if attribution stays fuzzy?

That standard is harder to fake. It also fits the reality of GA4 and Search Console right now. Terminus shows teams still need custom channel definitions to even separate some AI assistant traffic, and Try Aivo shows Google’s own AI surfaces remain blended into google / organic.

If you run a small SaaS, assume measurement will stay incomplete longer than you want. Build content that can justify itself without a perfect dashboard. Publish pages that help buyers trust you, help customers succeed, and give your team proof to reuse everywhere.

Sources