Generative Engine Optimisation
See what AI tools say about your organisation today, then improve it. An engagement that makes generative search visibility measurable.
AI tools are answering questions about you before anyone reaches your website
Generative AI tools such as ChatGPT, Google AI Overviews, Microsoft Copilot, Perplexity and Gemini are increasingly answering people's questions before they ever visit an organisation's website.
For many organisations, this creates a new visibility and trust challenge. The answers these tools generate may shape public understanding, influence decision-making, and determine which sources are cited or ignored. Yet in most cases, organisations are not yet measuring what these tools say, how accurate the answers are, or whether their own content is being used as a trusted source.
This engagement makes generative search visibility measurable, then gives you a practical plan to improve it.
What the engagement produces
The engagement produces two things: a measurement instrument, and an action plan.
We build a library of approximately 80 to 120 natural-language prompts, based on the real questions people are likely to ask AI tools about your organisation, your category, your services and products, the issues you work on, and your areas of expertise.
These prompts are run across the major generative AI engines and scored for:
- Accuracy
- Completeness
- Framing and balance
- Source visibility
- Citation quality
- Gaps, risks, or misleading answers
- Whether your content is being retrieved, cited, or ignored
Key outputs
Four deliverables across the engagement, each building on the one before it.
Prompt Library
A reusable measurement instrument, made up of the questions people are likely to ask generative AI tools. It is used for the first benchmark, and for every future re-measurement cycle.
Measurement Report
An engine-by-engine view of what generative AI tools are saying today. It identifies where you are visible, where you are absent, which sources AI tools rely on, and where there are accuracy, framing, or content gaps.
Action Plan
A prioritised set of recommendations to improve generative engine visibility and answer quality.
Recommendations may span
- Website content
- Information architecture
- Structured data & schema
- Answer-led content & FAQs
- Authoritative external sources
- Third-party directories & publications
- Entity & brand-source signals
Re-measurement
A follow-up measurement cycle that shows whether the recommended changes have shifted generative engine behaviour, with a quantified view of improvement across priority topics, prompts, and engines.
The process
Three steps, from building the prompt library to delivering a prioritised action plan.
Generate a prompt library
We build a library of natural-language prompts based on the real questions people ask AI tools about your organisation, your category, and your areas of expertise. The library is reviewed with you before it is run.
Run and analyse responses
The prompts are run across the major generative AI engines. Every response is scored for accuracy, completeness, framing, source visibility, and citation quality, and the patterns are analysed.
Create an action plan
We turn the findings into a prioritised action plan: the gaps, risks, and opportunities the analysis surfaced, and the specific changes that will improve how AI engines answer you.
Execution and re-measurement
Execution begins once the action plan is approved. Data Story produces drafts and specifications for the new and refreshed content, structured data, and third-party source corrections set out in the action plan; your web and content team publishes. Re-measurement of priority topics follows once engines have re-crawled the updated sources.
What's required from you
A single project lead as point of contact across the engagement.
Access to source systems and research, provided at the start of the engagement:
- RequiredGoogle Search Console and GA4 for your website (read-only)
- High valueExisting audience and attitudinal research, customer journey maps, content inventory or sitemap
- UsefulBrand and content guidelines, prior LLM exploration if any has been done
An LLM Refs measurement-platform account, set up before the run-and-analyse stage.
A subject-matter reviewer available for one session while the prompt library is being built.
Timely sign-off at two points: the prompt library, and the action plan.
Web and content team availability for the execution phase, scoped against the approved action plan.
Recommended next steps
- 1Confirm the scope, priority topics, and preferred kick-off window.
- 2Provision access to Google Search Console, GA4, and any relevant research.
- 3Schedule the kick-off session.
- 4Schedule the two review meetings:
- Prompt library review
- Action plan review
A clear benchmark of where you stand, and a plan to improve it
Generative Engine Optimisation helps your organisation understand and improve how it appears in AI-generated answers.
The engagement gives you a clear benchmark of current visibility, identifies where AI tools are getting their information, highlights gaps and risks, and gives your team a practical plan to improve your presence across the generative search landscape.