How it works
See why AI picks your competitors. Then change it.
We ask AI the high-intent questions your buyers ask right before they choose, find where you're losing and why, and give you a fix for each one. Then we re-measure every month to show what worked, whoever made the fix.
- High-intent questions only
- Prioritized by severity
- Traced to the source
- Re-measured monthly
- ChatGPT
- Gemini
- Google AI Mode
- Perplexity
- Google AI Overviews
- Claude
- DeepSeek
- Grok
A score tells you there's a problem. This tells you what to do about it.
Every AI visibility tool can count how often you're mentioned. That number doesn't tell your team what to change on Monday. LLM Authority Index is built around the next step: what's wrong, where it came from, and what fixes it, written so anyone can act on it: your team, your agency, or ours.
Why we only track high-intent questions
A brand can show up in hundreds of general answers ("what is a personal loan?") and still lose every answer that leads to a sale ("best personal loan for bad credit"). We measure the second kind, so every number in your report is tied to revenue.
8 AI platforms · 10 buying moments · Re-measured monthly with the same questions and rules · Every finding linked to the prompt, platform and page behind it
01 · Response Conflicts
Catch the wrong answers AI gives about you, and where they came from.
We run each buyer question through 8 AI platforms and compare every answer against the others. When AI gets your pricing, features or reputation wrong, we trace the error to the exact page and sentence it was pulled from. Each conflict is rated Critical, High, Medium or Low and tagged by where it lives: your own website, review and comparison sites, publishers, or Reddit, forums and social. If AI didn’t cite your website at all, we flag that too, because it usually means a competitor’s page is answering for you.
You'll know
- Which facts AI gets wrong about you
- Which platforms repeat the error
- Which source page and sentence caused it
You can
- Ask the publisher to correct the source
- Update your own pages so AI picks up the right fact
- Brief legal or PR before a buyer acts on it
Example
Claude and Gemini say the brand has no free plan. ChatGPT says it does. Source: a 2024 comparison article, paragraph 3. Fix: request an update from the publisher, and add a plan comparison table to your pricing page.
02 · Recommendation tracking
Where AI recommends a competitor instead of you.
We group real buying questions by moment (best-of, comparisons, pricing, alternatives, trust, reviews and more) and show every one where AI recommends a competitor instead of you.
You'll know
- How often you’re mentioned vs. actually recommended
- How often you make the top 3, and how often you’re the #1 pick
- Which competitor wins each buying moment, on which platform
You can
- See what the winner has that you don’t
- Decide which comparison, pricing or proof page to build next
- Stop guessing which content will move the needle
Example
“Best CRM for small business”: Competitor A wins on 6 of 8 platforms, and its own comparison page is one of the sources AI cites.
03 · Sources
The sources AI listens to about you, and which ones are hurting you.
AI builds its answers from a small set of sources. We show every site and page it cites about you, by type (your site, competitors, official, reviews, community, publishers), with your share of citations and the sources causing conflicting answers.
You'll know
- The domains and pages AI cites most about you
- How much of it comes from competitors’ own pages
- Which sources are causing the wrong answers
You can
- Hand PR a prioritized target list
- Pitch the reviewers and publishers that actually move AI
- Check whether new coverage gets cited
Citation frequency shows which sources AI uses. It does not prove a source’s quality or that it caused a recommendation.
04 · Recommended fixes and Action Plan
A fix for every conflict, and one plan to get them done.
Every page causing a conflict gets its own fix, down to the detail: the page, the sentence, what it should say instead, who should own it (content, web, PR, social or legal) and how long it should take. Update your page, request a correction from the site, or post one from your official account. Tell us which answer is correct and the fixes rewrite themselves. Each fix is marked as something your team or agency can do, or something our team can handle for you.
You'll know
- What to do first, and why
- Which team should own each fix
- How much effort each fix takes
You can
- Assign it in your next standup
- Send the fix list to your agency as a brief
- Copy the fix list into your team’s tickets
- Hand the fixes you want done to our team in one click
05 · Proof Tracking
Prove the fixes worked, whoever made them.
Every month we re-run the same questions under the same rules, so you can see which fixes changed AI’s answers, and which didn’t. Fixes made before your next scan show up in your next report. Because the measurement doesn’t depend on who did the work, it works as an independent scorecard for your in-house team or your agency.
You'll know
- Your recommendation share, month over month
- Which competitor is gaining on you fastest
- Whether a specific fix changed a specific answer
You can
- Report real progress to leadership
- Check whether your agency’s work is moving AI answers
- Stop work that isn’t moving AI
- Double down on what is
Independent scorecard
Already have a team or agency? Measure their results independently.
The same monthly measurement works as a neutral scorecard: same questions, same platforms, same rules, whoever does the work. Export CSV on Pro ($500/mo) and share it with your agency or leadership.
06 · Done-For-You
Want it fixed, not just found? We'll do the work.
Our team does the work for you every month: technical SEO, page fixes, new pages, comparison pages and case studies, correction requests to the publishers AI quotes, and posts from your official accounts. 10, 20 or 30 deliverables a month from $1,500, with everything in Pro included, so next month's report shows exactly what changed. One plan, one login, one team. Public rankings are never affected.
Measured across the AI your buyers actually use.
- ChatGPT
- Gemini
- Google AI Mode
- Perplexity
- Google AI Overviews
- Claude
- DeepSeek
- Grok
See your own conflicts and fixes before you pay a cent.
Start free and get your first report within 48 hours: three answers you're losing, where each one came from, and the fix for each. No call, no credit card. Every month you wait, the competitors already fixing their answers pull further ahead.
Get my independent scorecard · Get it fixed for me · Need Enterprise? Talk to us
- No credit card for Free
- Cancel anytime
- First report in 48 hours
- Public rankings can't be bought