Generative Engine Optimisation (GEO) for SaaS
GEO is the work of getting your brand cited, named and accurately described in the answers AI assistants generate. There is no ranking position to win: a model retrieves a handful of sources and writes prose from them. We work on which sources those are, whether you are among them, and whether what they say about you is true.
What we can actually evidence.
How these are measured
- SaaS brands served:
- n = 30 SaaS brands. company inception to August 2026. Count of distinct SaaS companies under a signed advisory agreement. Excludes one-off audits, workshops and prospect calls.
- Years in-house experience:
- n = 1 founding team. 2015 to 2026. Cumulative years spent in salaried, in-house growth and marketing roles at SaaS companies before founding Surge45.
- AI assistants tracked:
- n = 6 assistants. current monitoring coverage. ChatGPT, Gemini, Claude, Perplexity, Microsoft Copilot and Grok, plus the AI Overviews and AI Mode surfaces inside Google Search. A coverage fact, not a performance claim.
- Growth pillars, one system:
- n = 4 pillars. current advisory model. Discover, Convert, Attribute and Grow. A structural fact about how the advisory is organised, not a performance claim.
The gaps quietly costing you pipeline.
An assistant names three competitors in your category and does not mention you at all
You are named, but the description is wrong, and you cannot work out where the wrong information came from
Your organic rankings are stable and clicks are falling, which reads as an SEO failure and is not one
You appear in the first answer of a conversation and disappear by the third, and nobody has noticed because nobody tests follow-up turns
Your robots.txt was written before any of this existed and nobody has looked at it since
You are being sold llms.txt, schema packages and a blended visibility score, and cannot tell which of them does anything
How we drive compound growth.
Diagnosis before deliverables
Four things cause low AI visibility and they have wildly different remedies. We establish which one you have before proposing anything.
The sources, not just the site
Ten domains carry over a third of B2B SaaS citations and eight are sites you do not own. On-site work alone cannot deliver the outcome, and we say so before you buy.
Measurement you can reproduce
A frozen prompt set, repeated runs, per-platform reporting, against definitions we have published so you can hold us to them.
A stated line on what we will not do
No astroturfing, no bought reviews, no guaranteed placements, no llms.txt on the invoice. The exposure would be yours, not ours.
Why GEO is a discipline and not a rebrand
SEO optimises for a ranking function. GEO optimises for a retrieval-and-synthesis pipeline, and almost every practical assumption changes as a result.
- The unit of selection is the passage, not the page, so a buried answer loses to a clearer one on a weaker site
- Question-answer fit routinely beats accumulated domain authority, which inverts a decade of instinct
- Most citations come from domains you do not own, so the majority of the work is not on your website
- Answers vary between runs and change with model releases, so a single measurement is a screenshot rather than a baseline
- The worst outcome is not absence but being present and described wrongly
- Google's AI surfaces are grounded in Google Search, so conventional SEO remains the foundation rather than the legacy
What the work consists of
AI crawler policy
Retrieval crawlers and training crawlers are different things. Blocking the wrong one removes you from answers entirely, and it is the most common self-inflicted failure in the category.
Source concentration analysis
Which domains actually carry citations in your category, measured rather than assumed. This usually redirects the whole programme.
Third-party record correction
Stale review listings propagate into answers on every platform. Unglamorous, fast, and routinely the largest single accuracy failure.
Answer-first restructuring
Moving the answer into the passage a retrieval pass would pull, and making every passage survive being read alone.
Entity clarity
One consistent description of what you are, across your site, your listings and the properties that describe you. Models that cannot classify you substitute a competitor who is easier to classify.
The citation-sources programme
Review platforms, communities, analyst and publisher placement, video. The largest component and the slowest.
How a GEO engagement runs
Baseline
Declare and freeze the prompt set and competitor set. Run every prompt repeatedly across every surface. Record answers, sources and model versions.
Diagnose
Separate the four causes: blocked retrieval, poor extractability, weak entity clarity, or absence from the sources being cited. They have different remedies and different timescales.
Fix what is cheap and unambiguous
Crawler directives, stale third-party listings, opening passages on pages that already rank. Days and weeks of work, often the largest single movement.
Build the off-site programme
Review platforms first because they are fastest, then community, then publisher and analyst placement. Quarters of work, sequenced by where your citations actually sit.
Re-measure and report
The same frozen prompt set, per platform, with citation rate and mention rate kept separate. Changes in method start a new series rather than restating history.
Decide what to stop
Every review includes what is not working and what we would drop. A programme that only ever adds is not being managed.
Why SaaS teams choose us.
We publish our definitions
Citation rate, share of voice and brand accuracy are defined publicly, with the ways each is usually inflated. Use them against us.
We lead with the off-site work
Most citations come from domains you do not own. Selling only the on-site half is cheaper for us and worse for you.
We do not sell llms.txt
97% of these files received zero requests in a 137,000-domain log analysis, and Google says it has no effect. We publish one and charge nothing for it.
Per-platform reporting, always
A blended visibility score hides which platform is failing, which is usually what you needed to know.
We will tell you not to buy
Roughly a third of our free audits end with us saying the problem is elsewhere, or small enough to fix in-house.
Your team keeps the method
You keep the prompt set and the data whether or not we continue. It is your category's baseline, not our lock-in.
Questions teams ask.
Is GEO different from AEO and LLM SEO?
Barely, and not in ways that matter commercially. Generative engine optimisation is the broadest label; answer engine optimisation emphasises the content layer; LLM SEO and AI SEO are used interchangeably with both. What matters is not the word but whether the scope includes off-site citation sources and real measurement, because those are the two parts most commonly missing.
Does this replace our SEO?
No, and funding it by cutting SEO is self-defeating. AI Overviews are generated from pages already ranking, Gemini grounds in Google Search, and Copilot runs on the Bing index. Weakening conventional search weakens the AI surfaces built on top of it.
How quickly does anything change?
Accuracy fixes can move within days, because a stale third-party listing propagates fast and corrects fast. Extractability changes show up over weeks. The off-site citation programme runs over quarters. We report the three separately rather than blending them into one timeline.
Can you guarantee we will be recommended?
No, and nobody can. Model outputs are non-deterministic and change between versions without notice. What we commit to is measuring against a fixed prompt set, working on the sources cited instead of you, and re-measuring. A guarantee in this category is the tell.
What if we already have an AI visibility tool?
Bring its numbers. The most useful thing an audit often does is show that the tool's citation rate blends unlinked brand mentions into it, or averages platforms into one score. Our definitions are published so you can check yours against them.
Do you work with our existing SEO agency?
Usually, and it is generally the cheaper arrangement. They keep the technical and content execution; we set direction and run measurement. The part neither of us should skip is the off-site citation-sources programme, which most SEO retainers are not staffed for.
Specialised services.
Citation sources programme
Review platforms, communities, analyst and publisher placement, and video. Where most citations actually come from.
Learn moreChatGPT
The largest assistant, and the one whose behaviour differs most between answering from memory and running a search.
Learn moreGoogle Gemini
The second-largest assistant, and a separate product from AI Overviews and AI Mode.
Learn moreGoogle AI Overviews
The generated summary above the results, built from pages already ranking.
Learn moreWhat GEO costs
Market ranges for tooling, freelance, agency, advisory and in-house, and what you should not be paying for.
Learn moreThe measurement standard
Our published definitions for every AI visibility metric, with the ways each one gets inflated.
Learn moreMore ways we help.
SEO
The foundation Google's AI surfaces are grounded in. Not a legacy concern.
Learn moreContent & community
Content and community presence that shapes the shortlist before sales is involved.
Learn moreWhat is generative engine optimisation?
The definition, what the work covers, and what does not move it.
Learn moreGEO agency vs advisory
Two different business models behind identical language, and how to tell them apart.
Learn moreA worked AI visibility audit
The prompts, the observed responses, the scoring and the roadmap, redacted.
Learn moreFree AI Visibility Audit
The same method on your category, against your competitor set. No cost, no obligation.
Learn more