How it works

Find the gap, draft the fix, ship it.

GetIntel puts your buyers' real questions to the AI engines every day, shows you which ones name a competitor instead of you, and drafts the specific artifact that closes each gap. You approve everything.

The loop

Three steps, running daily.

01

Connect your domain

Point GetIntel at your site and name two or three competitors. It reads your homepage and pricing page to work out what you sell and who you sell it to, which is what the buyer questions get built from.

02

We probe the engines daily

A fixed set of buyer questions goes to ChatGPT, Perplexity, Gemini and Google AI Overviews every day, with Claude on the Growth plan. The set stays fixed on purpose, because a changed question makes your own history incomparable.

03

You approve the fix

Every gap comes back with a drafted artifact attached. It ships through Claude Code or Cursor as a reviewable diff, or straight to WordPress, Ghost or Webflow if you have connected one. Nothing goes out without your sign-off.

In detail

What each step actually does

What a "buyer question" actually is

Not a keyword. A question a real buyer types into an assistant — "best AI visibility tool for a small SaaS team", "Profound vs Otterly for agencies". Those are the moments where an engine names two or three products, and either you are one of them or you are not. Starter tracks 30 of them, Pro 60, Growth 120.

Why each engine is measured separately

Because they cite very different sources. In our own 100-question run across three Google surfaces, 1,011 of 1,364 cited domains appeared on exactly one surface. A blended score hides which engine actually moved, so we never collapse them into one number.

What gets drafted

The artifact that closes the specific gap found: llms.txt entries and Schema.org markup for technical gaps, Wikidata and directory entries when an engine does not recognise your brand, counter-article briefs when a competitor is cited instead of you, and outreach emails when the citation sits on someone else’s domain.

Where the fix lands

In your repo, not our dashboard. GetIntel hands the draft to the coding agent you already use as a diff you read before merging. That is the part most monitoring tools leave to you, and it is the reason the loop closes at all.

More depth on each piece: how the tracking works, how fixes ship, or the full FAQ.

FAQ

Frequently asked questions

A few minutes. You add your domain and two or three competitors, and GetIntel reads your homepage and pricing page to build the first set of buyer questions. The first scan runs immediately; daily probing starts from the next cycle.

Daily, on every paid plan. The same fixed set of buyer questions goes to each engine each day, so movement in your results reflects the engines changing rather than the questions changing.

ChatGPT, Perplexity, Gemini and Google AI Overviews on every paid plan, with Claude added on the $99 Growth plan. Each engine is scored separately rather than blended, because engines cite very different sources from one another.

No. Every draft is a proposal you approve. Fixes ship through your own coding agent — Claude Code or Cursor — as a reviewable diff, or to a connected CMS such as WordPress, Ghost or Webflow. Nothing is published without your sign-off.

Not to see the results. The tracking, scoring and drafts all work without touching code. Shipping a fix through a coding agent assumes you or someone on your team can review a diff, but the CMS publishing path does not require that.

See where AI names you

Seven days free, no credit card. Your first scan runs in minutes.

Start 7-day free trial