People ask me some version of the same question on almost every first call. “So what do you actually do to get us into ChatGPT?” Usually with a bit of suspicion in it, which is fair. A lot of what’s sold as AI SEO right now is a PDF with “add schema” on page three and a monthly invoice.
So here’s the whole process, in the order we run it. No secret sauce, because there isn’t one. It’s six pieces of work, most of them unglamorous, and every one of them ran on our own sites before a client paid for it.
- We start with an audit of what ChatGPT, Perplexity and Google’s AI already say about you, run many times per prompt, because single answers are close to random.
- The first fixes are usually technical: crawler access, firewall rules and snippet settings that quietly keep you out of AI answers.
- Then we make your facts consistent and machine-readable everywhere, so an AI can verify you instead of guessing.
- Content gets rebuilt answer-first, with real prices, comparisons and honest trade-offs, the kind of pages AI answers actually quote.
- Off-site mentions matter as much as your own site, so we track who gets cited in your market every month and go after the gaps.
- We report AI impressions, manual citation logs and leads. We don’t promise rankings or citations, because no one honestly can.
Why “getting cited” is a different job from ranking
Classic SEO fought for a spot under the answer. Now the answer sits on top, and most people never scroll past it. Ahrefs re-ran its click study on December 2025 data and found that when an AI Overview appears, the top-ranking page gets about 58% fewer clicks. Position one still matters. It just isn’t the finish line anymore.
We see it on our own site too. One Search Console export we published showed 39,310 impressions against 123 clicks over 87 days. Seen a lot, clicked a little. That gap is the reason this service exists.
The good news, and it’s real good news, is that Google says plainly that optimizing for generative AI search is still SEO. Same raw material: crawlable pages, clear facts, real authority. What changes is how you shape it and where you measure it.
Step 1: The AI visibility audit
Before touching anything, we find out what the machines currently say about you. We write the prompts your buyers actually type, not your brand name, and run each one many times across ChatGPT, Perplexity, Google AI Mode and AI Overviews.
Many times matters. SparkToro’s research found AI tools give the same list of brands less than once in 100 runs. So we don’t report “you’re third in ChatGPT.” We report “you appear in 18% of answers for these prompts, your main competitor in 61%, and here are the three sources the AI keeps citing for them.”
The audit also catches the wrong stuff. Old prices, a product you discontinued, a competitor’s feature credited to you. That goes on the fix list too. If you want the short version of this step on your own, our piece on why AI search recommends your competitor walks through the 20-run test.
Step 2: Crawler and agent access
This is boring and it’s where we find the most embarrassing problems. A security plugin from 2023. A CDN rule that challenges every bot it doesn’t recognise. A nosnippet tag someone added to a template years ago.
OpenAI documents that sites blocking OAI-SearchBot will not be shown in ChatGPT search answers. Google needs a page to be indexed and eligible for a snippet before it can appear in AI features. So we check robots.txt, firewall rules and server logs, and we confirm in the logs that the search bots actually get through. We block the training crawlers you don’t want and let the search ones in. Our robots.txt rules are published, so you can see exactly what we’d set.
Step 3: Entity and schema work
An AI recommending you is, in a way, an AI vouching for you. It wants to confirm who you are before it does that. So we make your basic facts the same everywhere: business name, address, phone, what you sell, who it’s for, what it costs.
On your site that means structured data, an About page that states facts instead of slogans, and prices in real HTML, not inside an image or a “contact us” form. Off your site it means Google Business Profile, directories, review platforms, LinkedIn, and anywhere else your details live. Schema alone won’t get you cited, and Google says there’s no special AI markup. But contradictions will absolutely keep you out.
Step 4: Citation-grade content
AI answers quote pages that answer the question in the first two lines, with specifics. Not “we offer flexible solutions.” More like “from $49 a month, for clinics with 2 to 20 chairs, includes SMS reminders.”
So we rebuild your money pages answer-first. Real numbers. First-hand experience. Honest trade-offs, including who you’re not a good fit for, which oddly makes AI more willing to trust the rest. We also build the comparison pages, “you vs competitor” and “alternatives to competitor,” so when buyers ask those questions the AI grounds its answer on your version instead of someone else’s.
Volume isn’t the goal here. Often we publish less than the client expected and rewrite more.
Step 5: Competitor citation recon and mentions
Here’s the part most people underestimate. A big chunk of what AI says about your category comes from pages you don’t own. Ahrefs studied 10,000 brands and found Reddit mentions in ranking threads had the strongest link to AI visibility, around 0.75, with news, YouTube and general web mentions close behind.
So every month we log who gets cited for your market’s prompts and which sources the AI leans on. The “best of” lists, the review sites, the Reddit threads, the YouTube reviews. Then we work on getting you into those places properly: pitching list authors with real data, getting genuine customer reviews, answering in communities as yourself and not as a sock puppet. We also watch for vacancies, like when a competitor opts out of AI features or a cited page goes stale. Those slots get reassigned to someone. We try to make sure it’s you.
Step 6: Measurement that doesn’t lie
Reporting in this space is messy and I’d rather say that upfront. Search Console now has a generative AI report, but it shows impressions for AI Overviews and AI Mode only. No clicks, no queries. ChatGPT and Perplexity don’t give site owners a dashboard at all.
So we combine three things: Search Console AI impressions by page, a manual citation log for your money prompts (run monthly, many times each), and the number that actually matters, leads and sales from AI referrals in your analytics. You see the same dashboards we do.
How it fits together
| Step | What we do | Typical timing |
|---|---|---|
| 1. Visibility audit | Buyer prompts run many times per engine, gaps and errors logged | Weeks 1 to 2 |
| 2. Crawler access | robots.txt, firewall, snippet tags, log checks | Weeks 1 to 3 |
| 3. Entity and schema | Consistent facts on site and across profiles | Weeks 2 to 6 |
| 4. Content | Answer-first money pages, pricing, comparisons | Months 1 to 3, then ongoing |
| 5. Recon and mentions | Monthly citation tracking, lists, reviews, communities | Ongoing, results usually from month 2 to 4 |
| 6. Measurement | AI impressions, citation log, AI referral leads | Monthly |
What we won’t do
- Promise you’ll be “number one in ChatGPT.” Lists change on nearly every run. Anyone selling a position is selling a screenshot.
- Sell llms.txt as the fix. Fine to have, but Google says its search doesn’t use it.
- Pump out fifty AI-written posts. That’s how you get more pages nobody quotes.
- Fake reviews or fake Reddit accounts. They get found, and then the AI has a new story to tell about you, a worse one.
That’s it really. Six steps, no magic. The part that takes the longest is the part outside your website, because you’re changing what the rest of the internet says about you. Most clients see the technical and content fixes show up within weeks, and the mention work start to move the numbers somewhere between month two and month four.
FAQ
How does AxoVox get businesses cited by AI search?
Six steps: an AI visibility audit, crawler and firewall fixes, consistent entity data and schema, answer-first content with real prices and comparisons, monthly competitor citation recon with mention building, and honest measurement.
Can you guarantee my business will appear in ChatGPT?
No. AI tools return a different list almost every time and nobody outside OpenAI or Google controls the answer. We commit to tested methods, transparent dashboards and honest reporting of what moved and what didn’t.
How long does it take to get cited by AI?
Technical and on-site fixes often show within a few weeks. Earning mentions in lists, reviews and communities usually takes two to four months to show up in AI answers.
Which AI platforms do you optimize for?
ChatGPT, Perplexity, Google AI Overviews, Google AI Mode, Claude and Copilot. They use different indexes, but the same foundations of crawl access, clear facts and third-party mentions help on all of them.
Who does the work?
The founder, on every account. Trusted specialists for design or video join by name when needed. There’s no hand-off to a junior team.
Disclosure: this is our own service, explained by the founder. The outside studies we cite were checked on November 10, 2026. Nobody can guarantee what an AI answer will say, and that includes us.