SHERLOCK AI · OUR ENGINE

We don't guess what AI thinks of you.
We ask it.

Sherlock AI runs 50+ real buying questions across ChatGPT, Perplexity, Gemini, Claude, and Google AI Mode — every week. It sees exactly what your customer sees.

sherlock.ai · v0.9 · live query
> "best magnesium supplement for sleep"
ChatGPT named Momentous · BiOptimizers · Thorne
Perplexity named BiOptimizers · Momentous
Gemini no clear brand named
Claude named Thorne · Natural Vitality
Google AI named BiOptimizers · Momentous · Thorne
Your brand: not named share of voice: 0 / 5
Also tested this week: "best mattress for side sleepers under $2000" · "cold brew coffee subscription with best value" · "electric toothbrush for sensitive gums" · +47 more
HOW IT WORKS

Same input a shopper gives. Every answer AI gives back.

01
50+ real buying questions

Written the way your customer would ask — category, price, feature, comparison. Not keywords.

02
Five engines, in parallel

ChatGPT, Perplexity, Gemini, Claude, and Google AI Mode. Real accounts, real answers, no scraping.

03
Named / not named

For every answer we record which brands the engine actually recommends — and where you rank.

04
Weekly. Every week.

Engines shift. Competitors move. Drift shows up in days, not months. Sherlock catches it.

WHAT SHERLOCK PRODUCES

The Sherlock AI Report.

The deliverable. What lands in your inbox when you request a free audit — and what a client sees on week one, week four, and every quarter after.

01
Named-status across five engines

For every buyer question, whether the engine named you — and where you rank when it did.

02
Competitor comparison

Which brands the engine names instead of you. Not a guess — the exact string it printed.

03
Buyer questions we asked

The full 50+ question set, phrased the way your customer would ask. Category, feature, price, comparison.

04
Personas the engine assumes

The buyer segments AI reads your category through — the ones you win, and the ones you should.

05
What to fix, in priority order

The prescription. Content, structure, entities. What moves share of voice first.

Every finding is verifiable by you, on any AI engine — not just our tools.

SHERLOCK IN THE PROGRAM

Runs before. Runs throughout. Runs after.

Week 1
Sets the baseline. Share of voice, citation rate, named responses across all five engines — the targets we hold the program to.
Every week
Verifies every prescribed change once it ships. Named or not named — not an opinion.
Throughout
Catches drift. Engines shift. Competitors move. Sherlock sees it in days, not months.
Quarterly
Re-runs after the program ends. Bounded. Quarterly, not ongoing. You keep the visibility.
COVERAGE

Five engines. Every week.

ChatGPT
Category leader by usage. Cites sources sometimes; the pass we watch first.
Perplexity
Cites almost every claim. The one where a working citation strategy shows up fastest.
Gemini
Google’s conversational surface. Behaves differently than AI Mode — both matter.
Claude
Rarely cites, high-intent audience. Named-response count is the signal here.
Google AI Mode
The generative overlay on Google. Bridges classic search and the answer layer.
HONEST ABOUT WHAT IT IS

What Sherlock does not do.

Sherlock measures what AI says about you. It does not predict revenue. It does not attribute sales. It does not promise a specific customer count. It tells you the truth about your presence in AI answers — where you are named, where you are not, and who is named instead. What you do with that truth is the program.

Try Sherlock AI on your brand.

Your Sherlock AI Report runs 50+ real buyer questions across all five engines. You see exactly where you're named — and where a competitor is named instead. Free. Delivered in 3 business days.

Get Your Free Audit
Request-based. We produce it manually. Usually within 3 business days.