The AI-Mediated Web

Your next customer is asking AI about you — not Google.

Before they ever reach your site, buyers ask ChatGPT, Claude, Gemini and Perplexity for a recommendation. Prisma measures whether your brand appears in those answers, whether it's cited as a source, and how often a competitor is named instead — then shows you exactly what to do about it.

Request an audit See how it works

Evidence-backed · Repeatable · A Spectras product

The mark is a citation matrix. Saturated cells are the sources AI cites often — faint cells are where your brand is absent.

The shift

Search used to return links. Now it returns an answer.

When the answer names one brand and not another, the buyer never sees a list to choose from. The recommendation has already been made. If your brand isn’t in that answer, you were never in the running — and nothing in your analytics will tell you why.

  • GEOGenerative engines. ChatGPT, Claude, Gemini, Perplexity — a real question, asked conversationally. Being named in the answer is the win.
  • AEOAnswer features. Google AI Overviews, featured snippets, People Also Ask — the box shows its sources, so being the cited source is the win.
  • Off-siteEarned authority. AI cites sources you don’t own. We name the exact ones winning the questions you lose.

The pattern we find most

A brand can be everywhere when named — and invisible when it isn't.

Ask the assistant about a brand by name and it speaks glowingly. Ask the way a real buyer asks — by describing the problem — and the same brand vanishes. Only the second number is the business risk, and it's the one no other tool measures.

Asked by name0%

“What does [brand] offer?” — named in nearly every answer.

Asked by need0%

“Which option is best for someone like me?” — absent, while rivals are named.

That gap is the whole game. Prisma measures it, explains it, and closes it.

What we measure

Three questions, each answered with evidence.

01

Mention

Does the answer name your brand at all? Matched on name and domain, so a paraphrase still counts and a look-alike doesn’t.

02

Citation

Did the engine cite your site as a source? Being named and being sourced are different problems — with different fixes.

03

Share of answer

Of every brand named in the answer, how much of it was you. A competitor above you here is being recommended where you are not.

The deliverable

Seven reports. One of them is the point.

Six reports are evidence. The seventh — the Gap Bridge — is the answer. Select any report to see what it delivers.

Beyond your own site

AI cites the surfaces you don’t control. We tell you which ones.

The answer isn’t only your website. Assistants lean on third-party sites, video and community platforms. Where those win the questions you lose, the work is earned — and Prisma names the exact targets from real citation data, not generic advice.

The authority gap

The precise third-party domains AI cites in answers where your brand appears nowhere — your earned-media target list.

Video & community

Where YouTube, forums and reference sites are already being cited for your category, and how to be present in them.

Signed off, not guessed

Recommendations that change a regulated brand’s pages carry their evidence and a human review — never an unchecked machine verdict.

How it works

Five steps. Repeatable by construction.

An audit is only useful if you can run it again next quarter and trust the difference. Follow the line — hover any step to see what it does.

Define the brandYour domain, competitors, and the categories and audiences that matter to you.
1
Define the brand
Your domain, competitors, and the categories and audiences that matter.
Draft the questionsBuilt the way real buyers actually ask — across your categories, audiences and buying intents.
2
Draft the questions
Built the way real buyers actually ask — across categories, audiences and intents.
Lock the setA locked question set is fixed, so an audit run today still means the same thing next quarter.
3
Lock the set
Fixed once locked, so today's audit still means the same thing next quarter.
Measure across enginesFanned across every major assistant, with repeats, and every answer kept as evidence — forever.
4
Measure across engines
Every major assistant, with repeats, and every answer kept as evidence.
Score & deliverScored the same way every time, out to a scorecard, a written report and a prioritised roadmap you can act on.
5
Score & deliver
A scorecard, a written report and a prioritised roadmap you can act on.
0
AI engines measured
0
Reports per audit
0
Answers stored as evidence
0
Re-measured every quarter

Why the numbers hold

Evidence, not assertions.

Raw answers kept forever

Every response is stored immutably — every finding traces back to the exact answer it came from.

Models frozen per run

A re-measurement reuses the same models, so score movement is never confused with model drift.

Locale-accurate

Answers are measured for your market, not a generic one, so the citations reflect the buyers you actually serve.

Find out where you stand

See how AI answers for your brand.

Melbourne, Australia — serving clients globally.

Request an audit