A SaaS founder emailed me last week. He'd been paying an SEO agency $8,500 a month for fourteen months, and he wanted to know why a competitor of his, a two-person startup with a fraction of his budget, kept showing up in ChatGPT and Perplexity recommendations while his own company was invisible.
I looked at the agency's most recent report. It had keyword rankings, backlink counts and domain authority trends, plus a nice pie chart showing "content published this quarter." What it didn't have, for $8,500 a month, was a single mention of AI search, or one data point on whether ChatGPT, Claude or Perplexity had ever cited his brand.
The report that tells you nothing
Most SEO agencies are still running the 2019 playbook. They track keyword positions, count backlinks and send you organic traffic graphs, and all of that still matters, kind of, since Google still drives the majority of organic traffic to most SaaS websites. The trouble is that the ground shifted under search and the agencies stayed where they were.
According to our citation tracking, 82.9% of ChatGPT citations come from third-party sources rather than directly from brand websites, so most of what ChatGPT cites about you was written by someone else. Backlinks, the metric agencies love most, explain roughly 2.8% of whether an AI model decides to cite you. The correlation between traditional SEO metrics and AI search visibility is so weak it might as well not exist.
So when the monthly report says "we earned 47 new backlinks this month," ask the obvious follow-up: do any of those backlinks influence whether Perplexity recommends your product when someone asks "what's the best project management tool for startups"? Your agency can't answer that, because it doesn't track it, and most agencies wouldn't know how to start.
What changed, and why agencies missed it
For twenty years, SEO was a Google game. You optimized for Google's algorithm, ranked on Google and got your traffic from Google, and agencies got good at it because the feedback loop was so clear: do a thing, check the ranking, adjust.
Then AI search arrived from several directions at once. ChatGPT launched search and Perplexity went mainstream. Claude started answering questions with web citations, and Gemini absorbed Google's index and began giving answers directly. Suddenly the question changed from "where do I rank on Google?" to "does AI even know my product exists?"
Most agencies treated this as a feature update. A "GEO" slide went into the quarterly deck, and maybe "AI search optimization" appeared on the homepage, but the core playbook stayed the same: rank higher on Google and everything else follows. That assumption is wrong now. We wrote about the tension between GEO and SEO tactics recently, and the data keeps confirming it.
Strong Google rankings do help with AI citations. The relationship only runs one way, though, so you can rank #1 for a keyword on Google and still be completely absent from every AI search engine's responses. What actually moves the needle is whether you exist in AI training data and whether trusted third-party sources mention you, with a page-one position as a bonus on top rather than a requirement.
26 content formats. Published on your domain. Matched to your brand.
Guides, comparisons, listicles, case studies, and more. RankControl generates content that gets cited by ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode.
The four things your agency can't do
The people at these agencies are mostly smart SEO professionals doing good work within the framework they know, so none of this is a comment on anyone's competence. The framework itself is what breaks, and it breaks in four specific places.
They can't measure what matters
Ask your agency how many times ChatGPT cited your brand last month. Then ask which specific questions trigger a mention of your competitor but not you, or whether your latest blog post is being pulled into Perplexity answers. You'll hear "we're looking into that," or you'll get pointed toward a tool they just started testing. You can't optimize what you can't measure, and an agency with no AI citation data is guessing.
Content built for crawlers instead of understanding
Traditional SEO content is built to satisfy crawlers through keyword density, header structure, meta tags and internal linking, and all of that is important. AI models just don't read content the way Googlebot does. They extract entities and evaluate factual claims against their training data, then judge whether a source has authority on the specific topic being asked about.
That's how an article can rank #1 on Google thanks to strong backlinks and perfect on-page SEO and still get zero AI citations: it reads like marketing copy with no substance. AI models want depth and real expertise. Given the choice, they'll cite the article that explains why your approach works over the one that says "our platform is the leading solution for X."
Rankings get reported, visibility doesn't
A ranking tells you where you show up on a Google results page, while visibility tells you whether you exist in the new search ecosystem at all, and those are different questions. A SaaS company that ranks #4 on Google for "best CRM for startups" might appear in zero AI search responses for that query. Meanwhile a smaller competitor sitting at #15 might get cited in ChatGPT, Perplexity and Claude consistently, because it published a detailed, opinionated comparison on its blog.
The agency reports the #4 ranking as a win. The founder is left wondering why customers keep saying "I heard about your competitor from ChatGPT."
Competitor tracking that stops at Google
Competitive intelligence in traditional SEO is mature, and half a dozen tools will show you a competitor's rankings, backlink profile and content strategy. Knowing whether AI search engines recommend that competitor is a different job. It means monitoring actual AI responses across multiple platforms and thousands of queries, continuously, and most agencies don't have the tooling for it. Some are trying to build it, but for now the founder who asks "where do my competitors show up in AI search?" gets a shrug.
This is the gap that worries me most. The monitoring blind spot costs you deals you don't even know you lost, because your competitor might be getting cited in ChatGPT for the exact queries your buyers use without any of it reaching your agency's monthly report.
Your competitors are getting cited by AI. You're not.
Every day without citation tracking is a day your competitors pull ahead in ChatGPT, Perplexity, and Claude.
The great unbundling of SEO
SEO used to be a single discipline. One agency handled technical audits, content strategy, link building and reporting, and that made sense when "everything" meant "Google." What I think founders in particular need to see is that search has since fragmented into at least four separate disciplines:
| Discipline | What it involves |
|---|---|
| Technical SEO and indexing | Google and AI crawlers can access your content, your structured data is clean, your site is fast. Table stakes: a one-time audit plus a developer who reads the docs |
| Content engineering | Content that earns both Google rankings and AI citations. It takes entity clarity, topical authority and an understanding of how AI models evaluate source credibility, and it's closer to product marketing than traditional "SEO content" |
| Brand distribution | Mentions on the third-party sources AI models actually trust: Reddit threads, industry reports, podcast transcripts, expert roundups. PR and community work, not link building |
| Citation monitoring | Tracking where your brand appears across ChatGPT, Perplexity, Claude and Gemini, which queries trigger competitor mentions, and whether your optimizations actually move the needle in AI responses |
A generic SEO agency handles the first of those and a watered-down version of the second. It ignores brand distribution because that's "not SEO," and it can't do citation monitoring because it doesn't have the tools.
The founders who are winning right now aren't spending $10K+ on agencies. They split search into these parts and handle each one deliberately, doing some of the work in-house and automating some of it, and very little of what they do looks like traditional SEO.
The DIY stack that actually works
I'm not saying every founder should fire their agency tomorrow. If yours tracks AI citations, optimizes content for both channels and reports on visibility across the full search ecosystem, keep it, because agencies like that are rare and valuable. If your agency's monthly report looks the same as it did in 2023, though, a modern stack has three layers.
The first is content creation, at 4-6 hours a week. You're writing content that answers real questions with genuine expertise, deep and opinionated and specific enough that an AI model would want to cite it because it actually says something useful. Keyword-stuffed blog posts and thin pages targeting long-tail queries don't count.
The principle behind that layer is that one great article that gets cited by AI beats ten mediocre articles ranking on page 2 of Google. It holds whether you write manually or use our content engine, which automates this part.
Citation tracking comes second, and it should be automated. You're monitoring where your brand appears in AI search responses, tracking competitor citations and measuring changes over time. It has to run every week, because AI models update their responses constantly and what gets cited today might not get cited next week. Done by hand, that's 8-12 hours a week per platform across four or five platforms, which isn't sustainable, so this is the layer you automate.
The third layer is the technical foundation: clean structured data, a fast site, proper robots.txt and llms.txt configuration, and verified AI crawler access. It's a one-time setup that takes a developer a day or two, plus quarterly check-ins to make sure nothing broke. Your agency charged you monthly for this, and it shouldn't be monthly.
All in, you're looking at $500-$2,000 a month in tools instead of $5,000-$15,000 for a generic agency, and 6-10 hours a week with the right automation instead of 20+ hours doing everything manually.
"But my agency says they do AI search now"
A lot of agencies added "AI Search Optimization" or "GEO Services" to their websites in the last six months. Some of them mean it, but most relabeled the services they already sold. Four questions will tell you which kind you're paying for:
- "Show me where our brand appeared in ChatGPT and Perplexity responses last month."
- "Which of our competitors gets cited more than us in AI search, and for which queries?"
- "What percentage of our content optimizations this quarter targeted AI citation improvement versus Google ranking improvement?"
- "How do you measure whether a content change improved our AI visibility?"
If they can't pull the data for the first one, they're not tracking AI visibility. Not knowing the answer to the second means they have no competitive intelligence for AI. A zero on the third makes their "GEO services" vaporware, and an agency that can't explain the measurement methodology behind the fourth is guessing.
The agencies that can answer all four questions are worth paying for. They exist, even if they're not the majority, and if your current agency fumbles on question one, you already know where you stand.
The part nobody talks about
This might be the most important point in the whole article. SEO agencies can be good at what they do and still be the wrong buy in 2026, because what they do only covers one channel, and for SaaS founders the missing channel is the one where buyers are increasingly making decisions.
We've seen it in our own data: AI search visitors convert at rates that blow traditional organic traffic out of the water. When someone asks ChatGPT "what's the best tool for X" and gets a recommendation, they arrive on your site with higher intent than someone who clicked a Google result, because a source they trust has already pre-sold them.
If your marketing strategy doesn't account for where your brand shows up in those AI responses, you're optimizing for traffic while your competitors optimize for trust. In an AI search world, trust is what drives the click that matters, the one from a buyer who already knows what they want and just needs to confirm your product is the right answer.
You can do all of this manually, tracking citations on every AI platform yourself and auditing your content for entity clarity. Or RankControl's agents can track your citations across six engines every week and publish the content, while you focus on building your product.
200+ SaaS teams already track their AI citations.
They know exactly when ChatGPT mentions their brand, and when it stops. Do you?
