The AI search visibility platform for SMBs

Turn AI mentions
into customers

Your customers are asking ChatGPT, Claude, and Google AI who to hire and what to buy. We measure whether you get named, diagnose why you don’t, and fix the evidence behind the answer.

No guaranteed placements. Reproducible measurement, a named gap for every omission.

Prompt panel · run 12
Who’s the best emergency plumber in Leeds?

Three firms come up most often:

  • Your business — named, position 1, cited from your own site
  • Competitor A — mentioned, no citation
  • Competitor B — mentioned from a directory

64%

Mention rate

1.8

Position

92%

Accuracy

Measured separately on every engine your customers use

ClaudeChatGPTGoogle AI
The shift

You can lose visits while gaining exposure

When an AI summary appears, people click through roughly half as often — and only 1% of summary visits produce a click on a link inside the summary. Your traffic report won’t show what the answer said about you.

2.5B

monthly users of Google AI Overviews

Google I/O, May 2026

700M

weekly ChatGPT users

OpenAI, January 2026

8% vs 15%

click-through with an AI summary vs. without

Pew Research Center, 2025

40–59%

of cited domains change within a month

Profound, June–July 2025

SEO asks

Does this page rank for this query, and did it earn a visit?

1.
2.
3.

AI search optimization also asks

  • Did the answer mention the brand at all?
  • Was it recommended, merely listed, or left out?
  • Which claim and which source supported that?
  • Was the description accurate and current?
  • Did it vary by engine, follow-up, location, or week?

Some people call this GEO or AEO. It’s the same work.

The method

Five phases, run on a schedule

Phases one and two are the audit. Three and four are the ongoing work. Five is how you know what actually changed.

01

Measure

A versioned prompt panel built from real customer language, run repeatedly across engines to capture a baseline.

02

Diagnose

A topic–source matrix that maps every important omission to a named gap: no content, weak evidence, missing validation, wrong source type, technical exclusion, entity ambiguity, or inaccuracy.

03

Build

Specific, attributable, current evidence published where the engines actually look — not more blog posts.

04

Distribute

Correct listings and entity data, earn third-party coverage, and participate honestly in the communities that get cited.

05

Learn

One controlled change per cycle, re-measured, kept only if it holds up against model variance.

What gets recorded

Every run captures the verbatim answer, date, engine and model, location and account state, every cited URL, brand mentions and their position, sentiment, and a human accuracy review.

Reporting

Eight metrics, never rolled into one score

A business can be widely mentioned and rarely recommended. A single “visibility score” hides exactly the thing you need to fix.

Mention rate

Are we present in the answer at all?

Position

Are we ranked highly, not just named?

Citation rate

Does our own site supply the evidence?

Earned-source share

Is outside corroboration present?

Accuracy rate

Is the answer safe and current?

Source diversity

Resilient, or dependent on one page?

AI referral quality

What happens after a click?

Assisted demand

Did branded search and direct visits lift?

What actually moves visibility

Graded by the evidence behind it

  • Third-party brand mentionsCorrelation
  • Topical authorityIndirect
  • Community and review contentObserved
  • Original research and statisticsBenchmark
  • Clear, structured writingBenchmark
  • Citing authoritative sourcesBenchmark
  • Entity and schema consistencyWeak

A controlled test

The same pages, rewritten different ways, scored on how prominently each version was used. Attributable evidence moved the score. Keyword stuffing did nothing.

Adding quotations27.1
Adding statistics25.5
Citing sources25.3
Improving fluency24.4
No change (baseline)19.8
Keyword stuffing19.8

Source: Aggarwal et al., GEO: Generative Engine Optimization, KDD 2024. Treat it as a reason to test, not a ranking law.

Who this is for

Small and mid-sized businesses whose buyers make decisions inside an AI answer before they ever reach a website.

Local services

Plumbers, dentists, lawyers, and salons whose customers now ask an assistant instead of scrolling a map.

SaaS & tech

Software brands fighting to survive the comparison and best-of prompts buyers run before a demo.

E-commerce

Stores where product, price, and merchant data consistency decides whether you get listed at all.

Professional services

Consultants and clinics in high-consideration categories where an inaccurate answer costs real money.

Engagement

The first 90 days

A deliverable at every stage. This work crosses teams — content, PR, analytics, and whoever owns your product and location data.

Days 1–30

Prompt research, engine selection, competitor set, crawler audit, baseline capture.

Deliverable

Versioned prompt library and baseline dashboard.

Days 31–60

Topic–source gap analysis, entity cleanup, priority content briefs.

Deliverable

Opportunity map and the first citable assets.

Days 61–90

Earned-source outreach, community and review process, repeated measurement, accuracy fixes.

Deliverable

First test readout and next-quarter roadmap.

Pricing

Start with the audit

The audit stands on its own. Most businesses move to a retainer because cited domains keep rotating after you stop looking.

One-off

Visibility Audit

For a business that has never measured this. Prompt panel built and run across ChatGPT, Claude, and Google AI.

From $900
  • Custom prompt panel from real customer language
  • Baseline run across ChatGPT, Claude, and Google AI
  • Eight-metric baseline scorecard
  • Topic–source matrix
  • Prioritised gap list
Book an audit
Most clients

Monthly retainer

Managed Program

For most businesses. Continuous measurement plus the work that changes it.

$1,500/month
  • Scheduled repeated measurement
  • Citable content production
  • Entity and listings cleanup
  • One controlled test per cycle
  • Monthly readout with confidence intervals
Talk to us

Monthly retainer

Larger Program

For regulated, high-consideration, or multi-location categories that need more hands on.

$4,000+/month
  • Everything in Managed
  • Larger prompt base
  • Multi-location and multi-product coverage
  • Accuracy and risk review
  • Named senior lead
Request a scope

What we promise, and what we won’t

Third-party mentions are valuable because they corroborate. Manufacturing them removes the only thing that made them worth having.

We promise

  • Reproducible measurement across engines and time
  • A named gap for every important omission
  • Specific, attributable, current published evidence
  • Accurate entity, product, and location data
  • Honest confidence intervals on every reported movement

We avoid

  • Guaranteed placement in any AI answer
  • A secret formula or proprietary ranking factor
  • Seeded forum posts, fake personas, synthetic reviews
  • Mass directory placements bought by volume
  • One opaque visibility score standing in for the scorecard

The usual objections

Have you asked ChatGPT what it says about you?

Almost nobody has. Leave your email and we’ll run a short prompt panel on your business and send back what the engines actually said.