Last week I typed our product name into ChatGPT, Perplexity, Gemini and Claude, and we got mentioned in exactly one of the four. That stung. It also gave me a project, which was to fix it in seven days, track everything, and see what actually moved the needle.
What follows is the sprint I ran, the actual day-by-day checklist with time estimates rather than theory or "best practices from our research," so you can steal it. Budget about 15-20 hours across the week, or 2-3 hours a day if you don't get distracted. I did get distracted on Day 4, when I went down a rabbit hole auditing our competitors' llms.txt files.
Day 1: Set your AI search baseline (2-3 hours)
You can't measure improvement until you know where you stand right now, so Day 1 is pure measurement, with no optimizations yet.
Ask ChatGPT, Perplexity, Gemini and Claude the same 10-15 queries your ideal customers would ask, then run the same set through Copilot as the fifth. Think "best [your category] tools 2026," "alternatives to [competitor]," "[specific use case] software," and "how much does [your product] cost." Screenshot every response and note which ones mention you, which mention competitors and which mention neither. That's your citation baseline.
Then check your AI referral traffic. In Google Analytics, or whatever you use, look for referrals from chat.openai.com, perplexity.ai, gemini.google.com and claude.ai. Most SaaS founders have never checked this, and you might be surprised, since industry data shows AI referral traffic converts at 4.4x the rate of organic search for SaaS products.
Last, run a crawlability audit to confirm AI crawlers can reach your content at all. Check your robots.txt for blocks on GPTBot, PerplexityBot, ClaudeBot and Google-Extended. Some WordPress security plugins block them by default, and if you've been blocking them, that's your first fix.
This Day 1 audit usually reveals that founders overestimate their AI visibility. Across our customer base, the average SaaS company appears in about 12% of relevant AI queries. Most think it's closer to 40%.
We'll show you exactly where your brand stands in AI search.
No commitment. $0 due today, cancel anytime. See how ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode talk about your brand today.
Day 2: Technical foundation (2-3 hours)
Today you make your site readable by AI crawlers, which comes down to two files and one schema update.
Create your llms.txt file
If you do one thing from this entire sprint, make it this. Put a plain text file at yoursite.com/llms.txt that tells AI agents what your product does. It needs a one-paragraph description of what you do and who it's for, plus links to your most important pages, such as pricing, docs, comparison pages and case studies. Add your key differentiators too, stated plainly rather than in marketing speak.
Ethan Mollick from Wharton noted recently that companies like HeyGen are already doing this. He also pointed out that most llms.txt files could do more to get AIs "excited" about using the product, by explaining things in plain English instead of just listing tech specs, and I think he's right. Write your llms.txt the way you'd explain your product to a smart friend. For the full technical checklist on llms.txt and AI crawler configuration, we wrote a complete guide on llms.txt, robots.txt, and AI crawlers.
Update your robots.txt
Make sure GPTBot (ChatGPT) and PerplexityBot are explicitly allowed, along with ClaudeBot / Claude-Web and Google-Extended (Gemini). If you had them blocked, the impact of unblocking typically shows up within 2-4 weeks as crawlers re-index your pages.
Add FAQ schema to your top pages
AI search engines love FAQ schema because it hands them pre-structured question-and-answer pairs they can cite directly, so add FAQPage JSON-LD schema to your top 5-10 content pages. On its own it won't get you cited, but it makes your content far easier for AI to parse and quote, which cuts the friction between your content and the AI's response.
Day 3: Content audit and rewrites (3-4 hours)
This is the longest day, since you're restructuring existing content so it's easier to cite. Start with your top 10 pages by organic traffic. They already have authority and just need restructuring for AI consumption. Run four checks on each one:
- Does the first paragraph directly answer the page's target query?
- Is there at least one comparison table?
- Is the H2/H3 hierarchy clear?
- Does the page have specific numbers?
On the first check, AI engines pull from the first 30% of page content about 44% of the time, so front-load your answers. AI also loves tables, and if a page compares features, pricing or tools in prose, convert at least one section into a markdown or HTML table.
Heading structure is how AI crawlers understand your content's organization, and random or missing heading levels make it harder to parse. Numbers help as well: pages with concrete data points like pricing, percentages and benchmarks get cited 2-3x more than pages with only qualitative descriptions.
For a deeper dive on restructuring content specifically for AI citations, check out our playbook on writing content AI agents actually cite. Don't try to rewrite everything today, though. Pick the three pages with the highest potential, meaning the most traffic and the worst AI structure, fix those, and work through the rest over the following weeks.
Day 4: Brand mention audit (2 hours)
This one is a slight detour from the technical work, but it matters. AI search engines synthesize information from across the entire internet as well as your website, so when your brand comes up on Reddit, in comparison articles, on G2 reviews and in industry discussions, those mentions feed into how the engines understand and recommend your product.
Start with Reddit. Search your product name and read what people say, whether they're recommending you, complaining or comparing you to competitors, because that's what the AI is reading too.
Then check whether you're in the major "best [category] tools" listicles on page 1 of Google for your category. These listicles are the #1 source AI engines cite when answering "what's the best X" queries, and we wrote about why listicle placements drive AI citations in depth.
After that, audit your competitors' mentions: run the queries where they get cited, and look at which content the engines cite and how it's structured. That tells you exactly what the AI considers citation-worthy in your category.
What surprised me in our own audit was a 6-person startup getting recommended over Salesforce-level incumbents in Claude's answers about CRM alternatives. Domain authority mattered less than content structure there: their blog posts were structured like API documentation, with clean hierarchy, direct answers in the first 100 words, and comparison tables that used actual feature names instead of marketing fluff.
Know exactly what AI says about your competitors.
RankControl's Recon Agent monitors competitor citations across ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode. See where they show up and you don't.
Day 5: Create missing content (3-4 hours)
By now you know where you're invisible, and Day 5 is about filling those gaps. Your Day 1 baseline and Day 4 audit probably turned up queries where your competitors get cited and you don't. You likely also found some where nobody gets cited well, which is a chance to be first, and some where you have relevant content that's buried or poorly structured.
Start with a comparison page. "Your Product vs [Top Competitor]" pages are citation magnets because AI engines pull from them every time someone asks a comparison question, so give yours a feature table, pricing side by side and a clear "pick this one if..." recommendation.
Next comes a FAQ or knowledge base page built from the 10 questions your sales team hears most often, answered in two to three sentences each, with FAQ schema added. That single page can become your biggest citation source, since AI engines treat FAQ pages like pre-packaged answers.
The page people skip is pricing. If yours is hidden behind "Contact Sales," AI engines can't cite you when someone asks "how much does [your product] cost," and they'll cite whoever has public numbers instead. Even ballpark tier ranges are better than nothing.
You don't need to create all of this from scratch, either. Most SaaS companies already have the raw material in their help docs, sales decks and internal wikis, and the work is restructuring it for the web with proper schema and heading hierarchy.
Day 6: AI search monitoring setup (1-2 hours)
You've done the work, and now you need to know whether it's working and when things change. AI search optimization breaks down the moment you stop paying attention. The engines update their models, re-crawl content and shift citation patterns without warning, and a page that gets cited today might get dropped next month when a competitor publishes something better structured.
Citation tracking comes first. At minimum, run your Day 1 queries again and compare, though doing this by hand every week is brutal: about 50-75 queries across five platforms, each one screenshotted and compared, which adds up to about 3-4 hours a week of tedious work. (If you want to skip the DIY approach, we compared the 9 best AEO tools for automating this.)
We built RankControl's AI visibility tracking specifically to automate this. Every week it checks your brand in ChatGPT, Perplexity, Claude and Gemini, and in Grok and Google AI Mode too. It tells you when you earn a new citation and shows what changed in the Monday brief. Whether you use a tool or do it by hand, though, the monitoring piece is what separates one-time optimization from sustainable visibility.
Set up AI referral traffic alerts as well, with a custom segment or alert in your analytics for traffic from AI search referrers, so you see week-over-week trends instead of a single snapshot. Then benchmark against competitors by tracking your own citations alongside your share of voice relative to competitors. If you're mentioned in 3 out of 10 relevant queries and your main competitor is mentioned in 7, that ratio is what you're trying to shift.
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.
Day 7: Measure, adjust, repeat
Run the same 10-15 queries from Day 1 and compare, but keep your expectations realistic, because you probably won't see dramatic citation changes in 7 days. AI models re-crawl and update on their own schedules. Some changes, like unblocking crawlers and adding llms.txt, take 2-4 weeks to show up, and content restructuring tends to show results faster in Perplexity, which crawls more frequently, than in ChatGPT.
By Day 7 you should be able to confirm this:
| What you should see | How to confirm it |
|---|---|
| Your llms.txt is being crawled | Server logs show GPTBot, PerplexityBot and ClaudeBot hitting /llms.txt |
| FAQ schema is validated | Run your pages through Google's Rich Results Test |
| Your restructured pages have cleaner heading hierarchies | Check with any accessibility audit tool |
| You have a baseline to measure against | This is the real win. Without it, you're optimizing blind |
The sprint doesn't end on Day 7, since what you've built is a foundation. The companies that win at AI search visibility treat it the way they treat traditional SEO and GEO, as an ongoing discipline instead of a one-time project.
The time math
Here's what the sprint costs in hours:
| Day | Task | Hours |
|---|---|---|
| 1 | Baseline audit | 2-3 |
| 2 | Technical setup (llms.txt, robots.txt, schema) | 2-3 |
| 3 | Content audit and rewrites | 3-4 |
| 4 | Brand mention audit | 2 |
| 5 | Create missing content | 3-4 |
| 6 | Monitoring setup | 1-2 |
| 7 | Measure and adjust | 1-2 |
| Total | First sprint | 14-20 |
Then add 5-10 hours per month to maintain it: re-running audits, updating content, creating new pages and monitoring citations. You can do all of this manually. Or RankControl's agents can run the baseline checks, track the citations, generate the content and show what changed, every week, while you focus on building your product. One plan, $400 a month.
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.
