Skip to content

AI & search recommendation readiness

Find out what is stopping AI from recommending your business.

We audit your website, your Google presence, and the third-party evidence AI assistants rely on, then show you exactly what's weakening your recommendation readiness.

Free preview to start · Full audit £199, one-time, no subscription

No guaranteed rankings. No secret AI tricks. Every finding is linked to evidence.

Example finding

Three sources agree. One doesn't.

Illustrative example, not a real client.

Website says

“Commercial property solicitor in London.”

Google Profile says

“General legal services, Croydon.”

Structured data says

No legal-service category present.

Citeworthy finding: Specialism and location are not consistently supported across the website, Google Profile and structured data.

Action: Align the homepage, the service page, the Google Profile category and the schema so that the real specialism and the real location are the same on all four surfaces.

What's different

Anyone can ask an AI what it thinks. Citeworthy proves why.

Illustrative example, not a real client.

What AI told you just now

Single response · unsourced · no evidence attached

“I believe this firm mainly handles general legal services out of Croydon.”

This is exactly what you'd get asking an AI chatbot yourself — no source list, no verification, and it can say something different the next time you ask.

What Citeworthy verified

Sourced · dated · independently checkable

Specialism not consistently supported

Website says commercial property, Google Profile says general legal services.

Checked 16 Aug 2026 · CW-0417

This is a preview of the evidence layer. The full audit checks every source, not just one or two.

What we audit

What Citeworthy looks at

Each area produces observable findings that are linked to the underlying evidence in your report, not opinions, and not a black-box score.

[1] 71% of consumers expect generative AI to influence at least half their spending decisions in the next 12 months.

First-party evidence

Your website and structured data -- name, services, locations and technical accessibility. Does it say clearly what you do, where, and for whom, and can a crawler read it at all?

Public evidence

Your Google Profile and third-party sources -- category, service area, opening hours, reviews and directory listings, cross-checked against your website.

Healthcare & care providers

Your CQC registration and rating, checked against your stated identity.

AI observation

How Perplexity, Gemini, ChatGPT and Claude respond to representative questions about your service, including when a system doesn't recognise you or returns something contradictory. Citeworthy is not affiliated with or endorsed by any of these providers.

Evidence ledger

[1]Accenture, Consumer Pulse Research 2026 (25,590 consumers surveyed across 16 countries).

How it works

Four steps, no long engagement required.

  1. 01

    Confirm your business

    You tell us who you are and link the sources we should audit (website, Google Profile).

  2. 02

    Audit runs

    Deterministic technical checks plus AI-assisted interpretation of your public evidence.

  3. 03

    Review evidence & fixes

    You receive a report with every finding linked to the evidence and a prioritised action list.

  4. 04

    Improve & re-audit

    You apply the changes; a re-audit confirms which findings have closed.

Methodology

What is measured, how, and by whom.

We score the public evidence available to machines. We do not claim to score secret AI algorithms.

Report preview

What the report contains.

Every audit produces the same structured report -- a finding, the evidence behind it, the certainty taxonomy, the AI recommendation observation, and the scoring breakdown. The blocks below are from the real report format; numbers in the scoring and AI blocks are illustrative.

Summary

3 blockers, 5 improvements identified

Finding

Specialism not consistently supported

Evidence / source

Website, Google Profile, structured data (JSON-LD)

Recommendation

Align specialism and location across all four surfaces

Timestamp / severity

16 Aug 2026 · Blocker

Verification method

Deterministic checks + AI-assisted interpretation

Certainty taxonomy

Every finding is tagged with one of these evidence classes. The labels are the same as in the report's technical appendix.

DETERMINISTIC

Rule-based check read from a source. Reproducible.

CROSS-PROVIDER

The same result seen across different AI providers in one audit.

SINGLE SNAPSHOT

One AI response captured once. Not stable on re-run.

INFERRED

A Citeworthy conclusion drawn from two or more evidence items.

AI recommendation observation

A branded recognition check across the major generative-AI providers. The report stores the full response for each system.

4 systems checked, 2 recognised the business, 1 currently recommends it.

Illustrative result, not a specific business. The real report lists each provider and its response.

Section III of the report.

Scoring breakdown

The report shows the arithmetic behind the final score: baseline evidence, confirmed deductions, and the current score.

Baseline evidence score

Verified presence, citations, reviews and structural signals.

72

Confirmed deduction

Specialism not consistently supported.

-12

Current score

Baseline evidence less the confirmed deduction.

60

Find out what's stopping AI from recommending you.

Free preview to start · Full audit £199, one-time, no subscription