Organic sessions rose, 12 new page-one terms appeared since spring — and pipeline fell 18% for the second consecutive quarter. That gap, from a founder's dashboard reviewed by growth consultancy Winston Francois, is the core problem with traditional SEO reporting in the era of AI assistants.

When Winston Francois checked the founder's top five buyer prompts across four AI assistants, the company appeared in only three of 20 answers. All 12 page-one rankings were for informational searches — questions people ask when learning about a category, not when deciding which company to buy from. "We'd gained visibility on questions people ask when they're learning about a category. We were barely showing up when they asked which company to buy from."

Winston Francois works with venture- and PE-backed companies on growth, including AI search. Rankings and organic traffic still help the firm understand performance, but they miss buyers who use an assistant to compare vendors, then go straight to a company's website. The consultancy now uses five measures to assess whether AI search work reaches buyers and produces opportunities.

Brand presence on buyer prompts

The first measure starts with five questions buyers would ask when choosing a vendor, run across four assistants, producing 20 saved answers. Marketers record whether the company is named, linked, or absent. The percentage of answers naming the company is the brand presence rate, tracked separately from linked citations.

Prompt specificity matters. "Best contract management software for mid-market legal teams" reveals more about buying intent than "What is contract management?" The consultancy recommends repeating identical prompts weekly and waiting several weeks before weighting any change.

Brand presence has been the first measure to improve in the firm's engagements, ahead of pipeline. It serves as an early check on whether the work reaches the right searches — but mentions should be tied to qualified opportunities before budgets increase.

Accuracy of the answers

Being named is not enough if the assistant describes the company wrong. One Series A client appeared in 11 of 20 answers, but only four described the company accurately. Assistants positioned it as an enterprise product while it sold to businesses with fewer than 200 employees — meaning a 100-employee company could rule it out before visiting the website.

After the firm made the company description consistent across sources the models could read, 10 of 11 mentions were accurate six weeks later. Demo-to-opportunity conversion recovered the following quarter.

Conversion from AI referrals

In Google Analytics 4, the firm groups sessions attributed to AI sources such as ChatGPT.com, perplexity.ai, gemini.google.com, copilot.microsoft.com, and claude.ai, using Session source for consistency. Across clients, these sessions convert at three to five times the rate of organic sessions.

The consultant's read: some evaluation has already happened before the buyer clicks through, because the assistant has explained and recommended the product. He still tracks how many visits become qualified opportunities before putting more money behind them. The report has limits — visits without a usable referral source go uncounted, and Google AI Mode traffic can mix into Google organic.

Branded search

A buyer can see a company name in an AI answer and then search it on Google. The consultancy tracks branded impressions and clicks in Search Console weekly, alongside direct traffic to the homepage and pricing page as supporting context.

Branded search has increased four to eight weeks after brand presence improves in the firm's engagements. Before attributing that lift to AI, the consultant checks what else ran during the period, including paid campaigns and PR. He would increase spend only if brand presence, answer accuracy, and branded search improved together and sales heard buyers mention AI — branded search alone would not be enough.

What buyers say brought them in

The fifth measure adds "ChatGPT or another AI assistant" to the "How did you hear about us?" field on inbound forms, with sales recording responses on opportunities in HubSpot or Salesforce. One client added the field in April; by July, nine of 54 new opportunities had selected the AI option. Those deals closed 30% faster than the rest of the inbound group.

Timing and setup

The firm's engagements show a pattern: brand presence improves two to six weeks after company information cleanup and first content changes, answer accuracy improves around the same time, AI referral conversions appear in weeks four to eight, and branded search rises around weeks eight to 12. Self-reported attribution can lag by a full sales cycle.

The consultant uses those intervals to decide when to review the work but does not promise clients pipeline will follow the schedule. If presence improves and nothing else moves after 90 days, he re-reads the answers to check accuracy and whether the prompts reflect questions real buyers ask.

Operationally, the system needs only a spreadsheet updated every Monday by whoever owns AI search — roughly 40 minutes for the 20 answers and accuracy review, 15 minutes to build the GA4 segment, and five for a form edit. He advises giving the sheet eight weeks before building a dashboard.

The founder whose dashboard opened this story now reports his presence in buyer answers alongside pipeline, and both have started improving — a better basis for the next budget decision than a report that stopped at rankings.