Skip to main content

Generative search

Building authority a system can verify

A GEO audit tells you what is missing. A GEO strategy decides what to build, what to state publicly, and who is prepared to put their name to it — which turns out to be the difficult part.

Most of the structural work in GEO is quickly done. Naming services properly, describing the organisation in machine-readable form, connecting the site to profiles it controls elsewhere — these are weeks of work, not years, and once done they largely stay done.

What comes after is harder, and it is where a strategy earns its value. Generative systems favour sources that can be relied on to be specific and correct. Becoming that source means publishing information that is genuinely yours, attributing it to people who genuinely know it, and making claims specific enough that they could in principle be shown to be wrong.

Every one of those requirements is a business decision before it is a content decision. That is why this work stalls in organisations that approach it as a marketing exercise, and why the strategy addresses those decisions explicitly rather than assuming them.

Approach

How the strategy is developed

  1. Establish the questions that matter

    The questions a prospective buyer would ask an assistant on the way to choosing a supplier like you. Not keywords — questions, phrased as people phrase them when they are talking rather than searching.

  2. Establish what a good answer requires

    For each question, what a system would need to be able to state, and what it would need in order to state it confidently. This produces a specific list of facts your public presence must support.

  3. Establish what you already have

    Which of those facts are already stated, which are implied, and which are absent. Also which are stated inconsistently, because contradiction is worse than silence.

  4. Establish what only you can say

    Where the organisation holds information, data or experience that is not already in circulation. This is the material that makes a site worth retrieving from rather than one of many restating the same thing.

  5. Decide who speaks

    Which people will be named, what expertise will be claimed for them, and how that will be substantiated. Organisations that skip this end up with well-structured anonymous content, which is a weak position.

  6. Sequence and specify

    Structural work first because it is fast and unlocks everything else, then the publishing programme, with a realistic view of what your organisation can sustain rather than what would be ideal.

Being retrievable is a structural problem and takes weeks. Being worth retrieving from is an editorial problem and takes considerably longer.

Workstreams

What a GEO programme typically contains

Entity foundation

Formal description of the organisation, consistent naming, connection to controlled external profiles, and accurate structured data. Done once, maintained thereafter.

Service architecture

A page per service, named as buyers name them, stating who it is for and what it involves. The single highest-yield change on most sites.

Definitional content

Clear, direct answers to the questions in your field that currently receive vague ones. Written to be quotable rather than to be long.

Attributed expertise

Named authors with described and substantiated expertise, applied to the content where judgement genuinely matters.

Original information

Data, methods, findings or experience that exist nowhere else. The hardest workstream to start and the one that compounds.

Consistency maintenance

A mechanism for keeping the website, the external profiles and the structured data in agreement as things change. Without it, drift returns within a year.

Well suited to this work

  • Organisations with genuine specialist expertise
  • Businesses whose buyers research before enquiring
  • Fields where accuracy matters and errors are costly
  • Categories where the same generic advice is endlessly repeated
  • Firms willing to name the people behind the work

Poorly suited to this work

  • Businesses competing purely on price
  • Organisations unwilling to state anything specific publicly
  • Sites intending to publish at volume rather than with substance
  • Anyone expecting an effect within weeks

Frequently asked questions

Is a separate GEO strategy necessary, or can this sit inside SEO planning?
For most organisations it sits inside the search strategy, and separating them would be artificial. A dedicated GEO strategy is worth it where generative discovery is commercially significant in its own right — typically businesses selling something people research through conversation rather than through a list of links, or businesses whose category is being actively reshaped by assistants.
What does this ask of the business beyond website changes?
More than most search work does. Attributed expertise requires someone prepared to be named. Verifiable claims require claims specific enough to be checked, which means deciding what you are prepared to state publicly. Original information requires having something to say that is not already in circulation. These are business decisions, and a GEO strategy that avoids them produces a well-structured site with nothing distinctive in it.
How do you decide what to publish?
By identifying where the organisation holds information nobody else is publishing, and where that information is relevant to a decision a buyer is making. That intersection is usually smaller than clients expect and considerably more valuable than another overview of a well-covered topic.
Will AI-written content help or hurt?
Content that restates what is already widely available adds nothing regardless of how it was produced, and publishing it at volume is actively harmful to a site's standing. Where a model assists a genuine expert in expressing something they know, that is a production question rather than a strategic one. The distinction that matters is whether the content contains anything that was not already available.
How is progress reported?
Against the observable signals rather than against a position figure. Whether services are individually identifiable, whether the organisation is described consistently, whether claims are attributable, whether external sources corroborate the site. These can be measured directly and improved deliberately. Where AI referral traffic is distinguishable in your analytics, it is reported alongside, with its limitations stated.

Decide what your organisation is prepared to state publicly

That question sits underneath every GEO programme. A conversation is usually a better starting point than a proposal, because the answer shapes the scope more than anything else does.