SaaS AI Search Advisory
For SaaS companies with a capable marketing team and nobody who has run AI search before. We set the direction, agree what gets measured and how, sit in your reviews and brief your leads, and your team does the work. The expertise compounds with you rather than with us, which is the whole point of buying it this way.
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.
Your team is capable and executing well on things they understand, and nobody has done this specifically
You are being pitched GEO by three suppliers and cannot tell which answers are correct
A tool is producing a dashboard nobody acts on, because it reports what is happening rather than what to do
You need to justify a new budget line to a CFO and have no defensible numbers to do it with
You have execution capacity and no confidence about where to point it
Somebody needs to say when the programme should stop, and no supplier paid for output ever will
How we drive compound growth.
Judgement, not output
Which prompts matter, whether an answer is materially wrong, which of four causes is producing your symptom, and when to stop.
Your team executes
We sit in your reviews and brief your leads. The knowledge ends up with you, which is what you are paying for rather than a deliverable.
Measurement you can reproduce
Against definitions we publish, so your numbers survive a challenge from someone who did not build the report.
We will tell you to stop
An advisory is one of the few models that can. Roughly a third of our free audits end with us saying the problem is elsewhere.
Why judgement is the scarce part
The tactics in AI search are not secret. What is scarce is knowing which of them applies to your situation, and being willing to say when none of them does.
- Four distinct causes produce the same symptom of low visibility, and each has a different remedy and cost
- Model outputs are non-deterministic, so telling a real change from run-to-run noise takes method rather than a dashboard
- Most citations sit on domains you do not own, so the work often belongs somewhere a marketing team does not usually look
- Vendors report metrics that quietly blend brand mentions into citation rate, roughly doubling the headline
- The category vocabulary is unsettled, so two suppliers can quote the same words for very different scopes
- The right answer is sometimes to do nothing this quarter, which no supplier paid for output can say
What we actually do
Diagnosis before anything else
Establish which of the four causes you have, so the quarter is not spent on the wrong remedy.
Sequencing
What to do now, what to do next, and what to leave. Ordered by damage done rather than by effort or by what is easiest to invoice.
The measurement layer
A frozen prompt set, the competitor set, the cadence, and the definitions. Set up once, then run by your team.
In your reviews
We sit in your pipeline reviews and brief your leads before board meetings, rather than delivering a document and leaving.
Supplier and tool selection
The nine questions worth asking, applied to whoever you are considering, including us.
Saying what to stop
Every review names what is not working. A programme that only ever adds is not being managed.
How an advisory engagement runs
Baseline and diagnose
Before any plan. The four causes have wildly different costs, and choosing a remedy before the diagnosis is choosing a treatment before the test.
Agree what gets measured
Metric definitions, cadence and who owns the number, settled before work starts rather than in the first review.
Sequence the work
What your team does now, next and not at all, with the off-site programme sized honestly as slow.
Support the execution
Your team runs it. We are in the reviews, answering the judgement calls as they come up.
Re-measure and re-sequence
The same frozen set. What moved, what did not, and what we would now drop.
Hand over properly
The method documented and running inside your team, so the engagement can end without the capability leaving.
Why SaaS teams choose us.
Paid for judgement, not volume
Our revenue is not tied to producing output, which is what lets us tell you to stop.
Published definitions
Every metric we report is defined publicly, including how each one is usually inflated. Use them against us.
The knowledge stays with you
By design. You keep the prompt set, the data and the method whether or not we continue.
We disclose our own tool
WriteWorks is a Surge45 product and we say so wherever we mention it. Ask the same of anyone else.
We qualify people out
Roughly a third of our free audits end with us saying the problem is elsewhere, or small enough to fix in-house.
Senior people only
The calls in this category are judgement calls. There is no junior version of this engagement.
Questions teams ask.
How is this different from buying GEO as a programme?
Scope of who does the work. The GEO programme assumes we run a good deal of it. This is direction and measurement with your team executing. If you have capacity but no senior direction, this is the cheaper and more durable option, because the capability ends up in your team.
When should we not buy this?
When your team has no capacity. Strategy handed to a team with no room to act produces a document and a slow-building resentment, and you should buy execution instead. Also when the problem is genuinely simple: if your crawler directives are blocking a retrieval bot and your G2 listing is two years stale, you need an afternoon and a checklist, and we will say so on the call.
Can you work alongside 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.
What does it cost?
Priced on scope and stage rather than a rate card, because the scope varies enough that a published number would mislead more often than it helped. Our guide to what GEO costs sets out the market ranges for advisory against the alternatives, so you can sanity-check any quote including ours.
How do we judge whether it is working?
Against the frozen prompt set, on the metrics agreed before the work started, reported per platform. If we changed the method mid-engagement, that starts a new series rather than restating history, and we would tell you so.
More ways we help.
Generative engine optimisation
The full discipline, run as a programme rather than as direction.
Learn moreAnswer engine optimisation
The content layer on its own, for teams whose diagnosis is extractability.
Learn moreAdvisory, agency or in-house?
What each resourcing model actually buys, including the case against each.
Learn moreHow to brief and evaluate a partner
Nine questions, most of which have one correct answer.
Learn moreThe measurement standard
Every metric we report, defined publicly so you can hold us to it.
Learn moreThe wider advisory
The four-pillar model this sits inside: Discover, Convert, Attribute and Grow.
Learn more