Flagship service

Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) Services

Answer engine optimization is the work of getting a firm cited, referenced and recommended by AI answer engines rather than merely ranked by Google. The two are no longer the same job. It covers four things: making your content reachable and parseable by the crawlers that feed ChatGPT, Claude and Perplexity; establishing your firm as a resolvable entity through structured data and consistent identity signals; restructuring pages so a retrieval system can lift a clean, self-contained answer from them; and measuring citation presence across the surfaces your buyers actually use.

The problem

Get cited when your buyers ask AI who to hire.

Ranking first no longer means being found. Buyers research inside ChatGPT, Perplexity and Google's AI Overviews, and those systems answer without sending a click. A firm can hold every position it has ever held and still watch its pipeline thin out, because the question that used to produce ten blue links now produces one paragraph naming three competitors.

What does ai visibility work actually involve?

Technical AI accessibility

Before anything can cite you, it has to be able to read you. Most AI crawlers do not execute JavaScript.

  • robots.txt audited and configured per bot, treating the search crawlers (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, PerplexityBot) separately from the training crawlers (GPTBot, ClaudeBot)
  • AI crawler accessibility testing, including checks for Cloudflare or security rules silently blocking them
  • Server-side rendering verification, so primary content exists in the raw HTML rather than appearing after hydration
  • llms.txt authored and maintained as a low-cost hedge, not sold to you as the mechanism

Structured data and entity establishment

AI cites sources it can identify with confidence. Ambiguity about who you are is a reason not to name you.

  • Organization, Person, Service and Article schema implemented and validated
  • Entity disambiguation, so your firm resolves as a known organisation rather than an unrecognised string
  • Knowledge graph alignment across your site, LinkedIn and every directory profile carrying your name
  • Breadcrumb and hierarchy markup so crawlers can map how your pages relate

On-page optimization for retrieval

Retrieval pulls passages, not pages. A section that only makes sense in context cannot be quoted.

  • Question-and-answer structuring built from the phrasing buyers actually use with AI
  • Self-contained sections that hold up when lifted out of the page
  • Comparison tables in real HTML, which are disproportionately favoured for extraction
  • Definitional and comparative content written to be the source a model reaches for
  • A scheduled freshness cadence, quarterly on money pages and monthly on anything containing a figure

Measurement

Citation presence is measurable. We start with the two official reports, which cost nothing.

  • Baseline audit of current citation presence across ChatGPT, Perplexity, Google AI Overviews, Gemini and Copilot
  • Google Search Console generative AI performance reporting, plus verification that the site is included in generative AI features
  • Bing Webmaster Tools AI Performance reporting for Copilot citations, cited URLs and grounding queries
  • Named competitor benchmarking and share of voice tracking

What is the difference between AEO, GEO and SEO?

The three overlap heavily in the work, and less in the outcome. SEO earns a position in a ranked list. AEO earns a citation inside a generated answer. GEO is used mostly interchangeably with AEO, though in practice it tends to emphasise content and formatting where AEO covers the technical and entity layers as well.

We do not price these separately, because splitting them would be a billing decision rather than a real distinction. One engagement covers all of it.

SEOAEO and GEO
GoalRank in a list of resultsBe the source quoted in an answer
Unit of competitionThe pageThe passage
Who decidesA ranking algorithmA retrieval and generation pipeline
Measured byPosition, impressions, clicksCitation presence and share of voice
Fails whenYou rank below the foldYou rank first and are never quoted

How do AI engines decide what to cite?

Differently from each other, which is the part most single-platform work gets wrong. One study of 34,234 AI responses found only around 11% of domains are cited by both ChatGPT and Perplexity, with brand citation rates of 0.59% on ChatGPT against 13.05% on Perplexity. Optimising for one does not carry over to the other.

Across all of them, a few things hold. Content has to be reachable and parseable, which is a crawler and rendering problem before it is a content problem. The source has to be identifiable, which is what structured data and consistent identity signals do. The passage has to answer the question on its own, because retrieval pulls passages rather than pages. And recency matters more than most people expect: pages updated within three months average around 6 citations against 3.6 for stale ones, and Perplexity weights content under 30 days old roughly 3.2 times higher.

Why do best-of lists keep showing up instead of firm websites?

Because for recommendation queries, that is what the models reach for. Ahrefs found across 750 top-of-funnel prompts that best-of lists were the most prominent page type in ChatGPT's sources, including for agency recommendations. When someone asks an AI to recommend a compliance consultancy, the answer is frequently assembled from third-party lists rather than from any consultancy's own site.

You cannot outrank those lists from your own domain. You get onto them. That is a separate workstream from on-site optimization and we treat it as one.

What we will not claim

No agency can guarantee you appear in AI answers. Google states plainly that indexing and serving are never guaranteed, and OpenAI promises eligibility rather than inclusion. Anyone selling certainty here is either not reading the documentation or counting on you not to.

Two specific claims circulate widely in this industry and neither survives checking. The first is that FAQ schema makes a page several times more likely to appear in AI Overviews; the figure traces back to a source page that no longer exists, and Google's own guidance states no special schema is required. The second is that llms.txt is essential; Google confirmed on 15 June 2026 that it does not affect Search, and crawler logs show it is rarely fetched at all.

We ship llms.txt anyway because it costs an hour and the agentic web may yet make it matter. We just will not tell you it is the mechanism.

Why this matters commercially

Our buyers are analytical professionals who check things for a living. Being the agency that says the content does the work rather than the markup is a real differentiator, and repeating a debunked statistic in front of a compliance consultant is a fast way to lose the engagement.

The process

How the engagement runs

01

Baseline

We test how your firm currently appears across ChatGPT, Perplexity, Google AI Overviews, Gemini and Copilot, against three competitors. Without a baseline there is no way to show the work moved anything, and most firms have never seen this data.

02

Access and rendering

We audit robots.txt per bot, confirm nothing at the CDN or security layer is quietly blocking the crawlers, and verify your primary content exists in the raw HTML. Nothing else matters if the systems cannot read the page.

03

Entity and structure

Schema implemented and validated, identity signals made consistent across your site and every profile carrying your name, and page structure reworked so a retrieval system can lift a clean, self-contained answer.

04

Measure and maintain

Monthly citation reporting against your named competitors, plus a scheduled freshness cadence. Citation presence decays, so this is maintenance rather than a project with an end date.

What you get each month

Your firm is the one named when a prospect asks an AI who to hire, on the surfaces where that research now happens, while your competitors are still measuring rankings.

Proof

83% more form submissions in 3 months, organic only

A risk management consultancy went from 2.4 website leads a month to 4.3, with 83% more conversions in 3 months of organic website optimization and no paid ads.

This engagement was SEO and conversion optimization.

Read the full case study

Is this right for your firm?

This tends to be worth doing when most of the following are true.

FAQ

Questions about this service

What is answer engine optimization (AEO)?

Answer engine optimization is the work of getting a business cited, referenced and recommended by AI answer engines. It spans four areas: technical accessibility for AI crawlers, structured data and entity establishment, on-page structuring so passages can be extracted cleanly, and measurement of citation presence across the surfaces buyers use.

What is the difference between AEO, GEO and SEO?

SEO optimizes for position in a ranked list. AEO optimizes to be the source an answer engine cites. GEO, generative engine optimization, is used mostly interchangeably with AEO, though it tends to emphasise the content and formatting side over the technical side. In practice the same engagement covers both, and we do not charge for them separately.

What is an llms.txt file and do I need one?

It is a markdown file at your site root summarising what your business does and linking to key pages. It is cheap to ship and we do ship one. It is not a citation lever: Google confirmed on 15 June 2026 that it does not affect Search, adoption sits near 10% of domains studied, and crawler logs show it is rarely fetched. Treat it as a hedge.

How do you get cited by ChatGPT and Perplexity?

The two behave differently, which is why single-platform work underdelivers. One study of 34,234 AI responses found only around 11% of domains are cited by both. Perplexity leans heavily on recent content and cites far more freely. ChatGPT cites a narrower set and draws more on third-party lists. Work has to target each surface on its own terms.

How is AI visibility measured?

We start with the two official reports, which are free. Google Search Console added a generative AI performance report in June 2026, and Bing Webmaster Tools reports Copilot citations, cited URLs and grounding queries. On top of that we run prompt testing across five surfaces and benchmark you against three named competitors each month.

How long does AEO take to work?

Technical fixes such as crawler access and rendering can change what is possible within weeks. Citation presence moves more slowly, usually across two to four months, because it depends on content being crawled, indexed and then selected. Nobody can promise inclusion. Google states plainly that indexing and serving are never guaranteed.

Start with a conversation

We will tell you honestly whether this is worth doing for your firm, and turn it down if it is not.