Kanata Labs builds a product-analytics platform that engineers genuinely prefer to the incumbents. Their problem wasn't the product. It was that their buyers had quietly changed how they shop — and Kanata was invisible in the new aisle.
The background
Kanata's buyers are technical: product managers and engineers who research tools carefully before booking a demo. Increasingly, that research started with a question typed into ChatGPT or Perplexity — "best product analytics tools for a small team," "alternatives to [incumbent]." The assistants would return a tidy list of three or four names. Kanata was never one of them.
They were losing deals at the very first step of the funnel, to a competitor set the AI had decided to trust — and they had no visibility into any of it.
The challenge
Getting recommended by an AI assistant isn't something you can buy, and it isn't classic SEO. Our audit found the root causes:
- A blurry entity. Kanata described itself differently on its homepage, its LinkedIn, its G2 profile, and in its press. To a language model assembling a picture from scattered mentions, they were fuzzy — and fuzzy brands don't get named.
- Content models couldn't quote. Their best material was locked in dense feature pages and gated PDFs, not the clean question-and-answer format answer engines lift from.
- Thin presence in trusted sources. The comparison articles, listicles, and community threads that AI models lean on barely mentioned them.
Our approach
We ran a focused Search Everywhere Optimization program over four months, measuring against a simple scoreboard: when we ask the assistants what we ask, do they say Kanata?
Phase 1 Entity foundations
We wrote one canonical description of Kanata and enforced it everywhere — site, profiles, directories, structured data — and published an llms.txt file giving AI crawlers a clean, unambiguous summary. We also confirmed their robots.txt actually welcomed GPTBot, ClaudeBot, and PerplexityBot (it hadn't).
Phase 2 Answer-ready content
We rebuilt their cornerstone content into the shape answer engines quote: a clear question as a heading, a direct answer beneath it, supporting detail below — with FAQ schema so the structure was machine-readable. Comparison and "how to choose" pages targeted the exact prompts their buyers use.
Phase 3 Trusted-source presence
Through digital PR, we earned honest mentions in the industry publications, comparison roundups, and community threads that ChatGPT and Perplexity cite — the sources that actually move an AI's recommendation.
Phase 4 Measurement & iteration
Every month we re-ran the category's key prompts across ChatGPT, Perplexity, and Gemini, logged whether Kanata appeared and how it was described, and doubled down on what moved the needle.
The results
By month four, Kanata Labs was the named recommendation in ChatGPT and Perplexity answers for its core category — appearing in the majority of the buying-intent prompts we tracked. That visibility fed the rest of the funnel: branded search rose 180% as people who heard the name went looking, and inbound demo requests tripled — at zero incremental ad spend.
Most importantly, it created a moat. The trust signals that make an AI recommend you — consistent identity, quotable content, credible third-party mentions — compound over time and are genuinely hard for a competitor to copy quickly.
"The marketers, developers, and SEO team are the same brain. Everything ships right the first time."— David Kim, CEO, Kanata Labs
The takeaway
AI assistants have become a discovery channel as important as Google was a decade ago — and almost nobody is optimizing for it yet. Kanata's four-month head start is exactly the kind of advantage that's cheap to earn today and expensive to catch tomorrow.
Services used on this project: