Article
If I started a GEO business, I would not start with SaaS
A founder-focused take on why GEO is better approached as a vertical, productized service before building another AI visibility dashboard.
If I started a GEO business now, I would not begin by building a SaaS dashboard.
That is slightly inconvenient, because I like building software more than selling services. But the market shape points the other way.
AI referral traffic as a share of total website traffic in SE Ranking's 2026 dataset.
Growth from 2024 to 2026. Tiny base, fast slope. Both parts matter.
Longer average session duration for AI visitors compared with organic search visitors.
The obvious SaaS idea is already crowded: track brand mentions across ChatGPT, Perplexity, Gemini, Claude, Copilot, and Google. Show a visibility score. Compare competitors. Send alerts.
That is useful. It is also where the strongest players are already going.
Profound has raised serious money and reached a reported $1 billion valuation. Peec AI reportedly reached $4 million ARR within 10 months. Ahrefs and Semrush can add AI visibility to existing SEO workflows. Adobe has its own enterprise angle. If you are a small team, building “another monitoring dashboard” means you are walking into the most obvious part of the market.
I would rather start where the work is messier.
the buyer wants more than measurement
Most teams do not wake up wanting a GEO dashboard.
They want to know why competitors appear in AI answers and they do not. Then they want someone to fix it.
A dashboard can show the gap. It does not automatically know which third-party source matters, what comparison page to write, which docs are outdated, which category language is missing, or whether the product positioning is too vague.
That is why I would start as a productized service.
The offer:
We help B2B SaaS and developer tool companies understand and improve how they appear in AI answers for high intent buying prompts.
The delivery:
- visibility diagnostic
- competitor and source map
- content recommendations
- owned content production
- third-party mention strategy
- monthly measurement report
That is less scalable than SaaS. Good. At the start, scale is not the problem. Learning is the problem.
pick one vertical
I would not sell this to everyone.
The first beachhead I would choose is AI and developer infrastructure SaaS:
- LLM observability
- eval and tracing tools
- vector databases
- RAG infrastructure
- agent frameworks
- developer APIs
- data infrastructure around AI products
The reason is not that this is the largest possible market. It is that the buyer behavior fits.
These products are comparison-heavy. Buyers ask “best X,” “X vs Y,” “open source alternative,” “how to choose,” and “what integrates with my stack.” Documentation matters. Community mentions matter. GitHub, benchmarks, docs, tutorials, and technical blogs all shape the answer layer.
Also, I understand this market better than I understand healthcare, insurance, or consumer travel. That matters. Vertical GEO depends on knowing which sources and questions matter inside the category.
sell a concrete diagnostic first
The first offer should be small and specific.
Something like:
I ran 25 high intent AI search prompts for your category across ChatGPT, Perplexity, Gemini, Claude, and Google. You were recommended in 3 of 75 answer samples. Competitor A appeared in 29. The biggest source gap is that three answer engines cite
{domain}, where you are absent.
That is a much better opener than:
AI search is the future and you need a GEO strategy.
The first one gives the buyer something to react to. The second sounds like a webinar.
Category
Best {category} tools in 2026
Best {category} for {use case}
Comparison
{product} vs {competitor}
{product} alternatives
Decision
How do I choose a {category}?
{category} pricing comparison
Implementation
How to implement {use case} with {stack}?
Best {category} that integrates with {stack}
Reputation
Is {product} reliable?
{category} tools recommended on Reddit
For pricing, I would keep it simple:
- free or low-cost diagnostic to get feedback
- $1,500 to $5,000 per month for visibility operations
- higher project pricing for category takeover work once case studies exist
The exact price depends on client size and scope. The important point is that execution is included. If the offer is only monitoring, you are competing with tools. If the offer includes action, you can charge for outcomes and learning.
build the software inside the service
I would still build software. I just would not sell it first.
The internal system would handle:
- prompt panels
- repeated sampling
- competitor tracking
- cited domain extraction
- source maps
- content briefs
- monthly reports
- change logs
At first, this can be ugly. Scripts, spreadsheets, a small admin UI, whatever gets the work done.
After 10 to 20 clients in the same vertical, patterns should appear. The same prompts repeat. The same source types matter. The same content gaps show up. The same report sections get copied. That is when software becomes obvious.
You do not need to guess the product. The service tells you what to build.
the moat is not the dashboard
I do not think the long-term moat is a chart.
The moat is the dataset and the judgment around it:
- real prompt panels by vertical
- source maps for each category
- knowledge of which third-party pages move visibility
- content patterns that actually get cited
- case studies with before and after baselines
- a reputation for honest measurement
That last one matters more than it sounds.
This market has too much fake certainty. A company that says “we cannot guarantee inclusion, but we can measure the baseline, improve the source graph, and report the uncertainty honestly” may sound less exciting. I think it will sell better to serious buyers.
kill criteria
I would set failure signals early.
If three months in one vertical produces no repeat customers, I would change the vertical or the offer.
If every client needs completely custom work and nothing becomes reusable, I would stop calling it a productized service.
If clients only care about the report and never want execution, I would question whether the pain is strong enough.
If results depend on sketchy tactics like fake reviews, fake statistics, or spam listings, I would rather quit the idea. That kind of work may make money for a while, but I do not want to build around platform manipulation that could collapse or become embarrassing.
the actual path
The 24-month version looks like this:
First 3 months: choose one vertical, run manual diagnostics, get 3 to 5 paying customers, and prove someone renews.
Months 3 to 9: standardize the diagnostic, content briefs, source maps, and monthly reports. Hire or contract execution help only after the founder can do the work manually.
Months 9 to 18: turn the repeated pieces into software. Keep the service layer, but make reporting and monitoring cheaper to deliver.
Months 18 to 24: either expand to an adjacent vertical or package the measurement layer as a paid product for teams that already have content execution.
This is not the fastest path to a shiny landing page.
It is probably a better path to understanding the market.
GEO is real enough to study, but not mature enough to trust from a distance. I would rather earn the product by doing the messy work first.