Step one · Fixed scope
Brand data and GEO audit
A fixed-scope review of one brand, one product category and your priority markets. It produces an evidence-based baseline and a prioritised implementation plan with owners and costs.
- Duration
- Six weeks
- Scope
- One brand, one category, priority markets
- Output
- Baseline, backlog, read-out
What we assess
Four questions, in order.
Ordered deliberately: there is no point measuring how you appear in AI answers before checking that the pages behind those answers are reachable and correct.
Accessible
- Whether important pages and product information are reachable by the search and AI services you intend to reach.
- How crawler controls, robots rules, bot categories and CDN configuration combine in practice — which is often not what anyone specified.
Accurate
- Whether product facts, claims, prices, availability and identifiers match your authoritative source.
- Where incorrect or unsupported claims about you originate, and which source is propagating them.
Consistent
- Whether your website, structured data, feeds, retailer pages and other sources describe the same products the same way.
- Which contradictions an AI system has to resolve on your behalf, and how it currently resolves them.
Observed
- How your brand appears across a governed set of AI prompts for your category and markets.
- Which sources AI answers cite when they discuss you or your competitors.
How it runs
Six weeks, three stages.
- Weeks 1–2
Scope and baseline
Agree the brand, product category, priority markets and the prompt set. Collect the authoritative data, map the sources exposing it, and read the current tool configuration.
- Weeks 3–4
Assessment
Run the accessibility, accuracy and consistency checks. Run the prompt set. Reconcile what the tools are configured to do against what they actually do.
- Weeks 5–6
Read-out
A senior read-out of the baseline with explicit unknowns, the prioritised backlog with owners and costs, and separate implementation and monitoring proposals.
What you get
A costed backlog, not a report.
The deliverable is a baseline you can act on: what is true today, what is wrong, what it costs to fix, and who owns each item. Unknowns are stated as unknowns rather than estimated.
The audit deliberately identifies incorrect tool configuration as well as data and content problems. In most engagements a meaningful share of the backlog is not new work at all — it is existing tooling that was never configured against the brand's actual policy.
You are not obliged to buy the implementation. The backlog is written so that your own team, or your existing agency, could execute it.
Book an audit.
Tell us the brand, the product category and the markets that matter, and we will come back with a scope and a fixed price.