How I got my new product cited in AI overviews
I launched RevPages about eight weeks ago.
It helps SaaS teams find SEO opportunities and prioritise them based on potential revenue.
Recently it’s been getting cited in Google AI Overview for “best keyword tool for SaaS”.
That’s a good search for the product to appear in.
Someone looking for a keyword tool specifically for SaaS is very likely to be interested in RevPages.
I’m not saying we’ve 100% cracked AI SEO or AEO, or whatever you want to call it - yet.
The site has had just 18 visitors from organic search so far, and only two have tried the product.
But getting mentioned in AI overviews for one of our main use-cases is a huge win.
Here’s exactly how I did it
1. I made the website effortless to understand and crawl
When you’re building a new product, it’s tempting to pack the site with clever "marketing-speak" to impress visitors.
But visitors don’t use those kinds of terms when they search, and neither does a search engine trying to work out what the site's actually about
You have to make it painfully obvious what your product does and who it’s for.
I started with a simple home page that gives an overview of what the product does and who it’s for.
Then I built a Features page which lists all of RevPages core functions, with a brief overview of each
Then I created a dedicated page for each feature, where I went into more detail about each one and its benefits.
I headed all pages with terms people actually search for, while keeping the pages focused on explaining the product clearly.
I also ensured the pages are all logically linked, and the site loads quickly and passes all the Google web vitals tests.
2. I published guides around the problems RevPages solves
I’ve written a handful of in-depth guides on the main topics that RevPages operates in:
- SaaS keyword research
- SEO revenue forecasting
- SEO opportunity scoring.
Most of these pages aren’t ranking yet, but they still serve a purpose
At this stage, they’re there to help search engines and LLMs understand what RevPages is about, and why we are experts in subjects we talk about.
These guides probably won’t ever bring in tons of traffic (especially now that AI overviews are taking away the majority of clicks going through to information content) but they build a solid SEO foundation to start building on - a step that can’t be skipped.
3. I gave the product a presence beyond its own website
I published straightforward launch posts on my personal website and agency site. They explain what RevPages is, who it’s for and where people can try it.
Those aren’t huge publications with high domain authority, but they allowed me to put an accurate description of the product somewhere other than its own site - which is very important to help Google and LLMs get some third-party evidence of the claims made on its own website
4. I pitched my founder story
I reached out to publications that feature founder stories and landed placements with Indie Hackers and Starter Story (thanks guys)
The links are very good for the site’s authority and relevance, but they also got the story in front of people who build, market and buy software - strong potential users of the tool.
A founder story gives me room to explain why I’m making the product and who I hope it will help - which again allows me to feed plenty of information about the product into the search ecosystem.
5. I asked to be included in relevant tool lists
I found sites that had published “best SEO tools” type lists and emailed them to ask if they would include RevPages.
This tactic is getting very difficult to do successfully because everyone is trying to do it right now, (mainly because it’s understood to be a major factor in getting cited in AI search) but it’s ten times easier if:
- You’ve built a genuinely good product that deserves to be in the lists
- You write personalized emails to the author and don’t spam everyone with automated outreach tools.
This still involved plenty of emails that went unanswered, as well as rejections, but so far, I’ve secured six list placements in relevant lists.
When users ask LLMs questions like “what is the best tool for saas keyword” research they pull their answers directly from lists like these - so earning mentions for your product is vital for AI search visibility.
What the numbers actually look like
Being cited is a decent milestone, but it hasn’t turned into meaningful traffic or product usage yet:
Here’s where RevPages is so far:
- 18 visitors from organic search
- Two people who tried the product
- One relevant AI Overview citation that I’ve spotted
For me, the encouraging part is where RevPages is appearing. “Best keyword tool for SaaS” is much closer to the product than a broad SEO search with lots of traffic but little buying intent.
The next challenge is turning that early visibility into visits, usage and feedback.
What I’m doing next
I’m planning to publish more detailed content that follows the work people do in RevPages, to build out topical authority and attract people who have a genuine need for the tool.
Another thing I’m going to really focus on is building more free tools for SEO teams and agencies.
Informational articles can still be useful, but AI answers mean fewer people need to click through for a simple explanation. A tool gives someone a reason to visit your site and get a job done - and I see them being a crucial part of SEO in the future.
Further down the line, I’d like to publish original data and studies. I used that approach a lot at my previous business, StandOut CV, to earn links. It takes work to get a study seen, especially when your site is new, so I’m waiting until our domain authority is a little higher.
And I’ll keep looking for relevant tool lists where RevPages would be a useful addition.
Eight weeks in, this is a small signal, not proof that the strategy works. But it’s nice to see the product start appearing for the searches I built it to serve.
The citation is real, but I would hold off on reading anything into the 18 visitors and 2 trials, because that pair of numbers cannot yet tell you whether you have a conversion problem or not.
Two out of eighteen reads as about 11%, which sounds like a number. Run it as a proportion and the 95% interval runs from roughly 3% to 33%. That range contains a product that converts terribly and a product that converts well, so at eighteen visitors the observation is consistent with almost any story you want to tell about the funnel. The citation signal and the trial signal are also being read together, and they are different instruments: the citation is about whether a model names you, which repeats across runs and has its own variance, while the trial rate is a proportion over people, which behaves like a coin and needs a much larger denominator than people expect. Eighteen is well inside the range where the second one is uninformative.
The arithmetic for making it informative is unfriendly but worth having. To pin the trial rate to within five percentage points you need roughly a hundred and fifty visitors, and to within three points, roughly four hundred. You are at eighteen, so the honest position is that the funnel measurement starts being readable somewhere around eight times your current traffic, and until then the citation count is the only thing moving on a timescale you can observe.
The reason that matters is that the two halves of your own post have different evidence standards. Your structural work — the crawlability, the feature pages, the third-party descriptions — is justified by reasoning rather than measurement, and it is cheap to keep doing. The conversion claim needs a denominator you do not have. Keeping those separate stops the second from being credited with the first.
Where I am coming from: I have a financial interest in the other side of this argument, since Piramyd (https://piramyd.cloud) exists to make growth metrics readable at sample sizes this small.
Given the current traffic, how many weeks of it do you need before you would trust a change in that trial rate at all?
Nice — getting picked up in AI overviews usually comes down to making your content easy to crawl, understand, and cite, while building enough external validation around it.
I’d focus on creating citation-friendly pages like methodology, real examples, use cases, and FAQs rather than relying mainly on the homepage. Clear headings, short answer sections, concrete numbers, and screenshots can make those pages much more useful as sources.
I’d also work on relevant third-party mentions and keep the product naming consistent across the web. Then track indexing, impressions, mentions, and which pages actually start getting visibility.
Since you launched only 8 weeks ago, I’d be especially interested in which type of page got cited first — homepage, methodology, or a specific example?
Great write-up, and refreshingly honest numbers. One thing I've learned building a tool that measures AI visibility: a single sighting can be misleading, because AI answers vary a lot from one run to the next. Asking the same question several times, on several engines (ChatGPT, Gemini, Perplexity, Claude), and tracking the share of answers that name you gives a much more reliable signal than one AI Overview screenshot. It's also worth checking other languages: a product can be cited in English and invisible in French or Arabic. Curious to see how your citation rate evolves as your list placements grow.
Your distinction between third-party evidence and the product’s own site is the part I’d keep testing. I’d log the exact question, engine, date, and whether the result names the product versus cites a specific page, then rerun the same prompt with a few natural phrasings before treating a citation as durable. One AI Overview mention is a useful signal, but the qualified visit and trial are the stronger checks.
Have you checked whether the citation persists across days and related searches?
getting cited in ai overviews is the new seo frontier - traditional rankings matter less every month. we're experimenting with the same for swapfile.live: clear factual pages that answer one question well. did you find any specific format (lists, tables, faqs) that gets cited more?
The 18 organic visitors and two trials are the perfect reality check: the citation is a signal, not a victory lap. I like the shift toward free tools—AI can answer the “what,” but a useful calculator still gets to keep the browser tab open.
With the citation already driving some organic traffic, what behavior would convince you AI visibility is creating qualified demand rather than just branded exposure—product trials, repeat usage, or paid conversions?
Love how honest the numbers are. Getting cited for a high intent SaaS keyword while traffic is still tiny is still a real signal. The part that stuck with me is making the site painfully obvious for both crawlers and LLMs. Guides for authority plus list placements for the answers those models actually pull from feels like the right AEO stack right now. Curious which of the six list mentions you think actually fed that Overview.