When a founder, a marketing lead or a developer needs a new tool, they don’t always start on Google any more. More and more often they open ChatGPT, Perplexity or Google’s AI Overviews and ask something like “What’s the best onboarding tool for a small B2B SaaS team?” The answer is a short list of names. If your product isn’t on it, you never reach the shortlist, and you won’t even know that you lost the deal.
That’s what Generative Engine Optimization (GEO) deals with. GEO means making sure AI assistants understand your product, trust it and name it when buyers ask questions you should be winning. This guide gives you a practical 30-day plan. You can run it yourself, and it’s the same structure we use when we advise startups on AI visibility.
Why AI answers are the new page one
For twenty years, “page one” meant ten blue links. You fought for a position and then fought again for the click. AI answers change both parts of that:
- Fewer slots. An assistant usually names three to seven options, not ten results plus ads. Being “number eleven” in an AI answer means you don’t exist.
- Answers come pre-filtered. The buyer gets an opinion instead of a list of pages: “Tool A is best for startups, Tool B for enterprise.” That framing shapes the shortlist before anyone visits your site.
- The sources are different. Assistants build answers from many places: your website, comparison articles, review platforms, documentation, forums and community discussions. Ranking your homepage is no longer enough. The wider web has to tell a consistent story about you.
ChatGPT, Perplexity and Google AI Overviews each behave a little differently. Some lean on live web search and show citations, and others rely more on what the model has already absorbed. The principle is the same for all of them, though. AI recommends what the web clearly and repeatedly says about you. GEO is the work of making that signal clear, accurate and hard to ignore.
Week 1: Audit where you show up (and who shows up instead)
You can’t improve what you haven’t measured. In the first week you build a simple baseline.
Step 1: Write 15–20 buyer-intent prompts
Write questions the way a real buyer would ask them, not the way you’d describe yourself. Mix these types:
- Category prompts: “What are the best [category] tools for startups?”
- Use-case prompts: “How can a 10-person SaaS team automate [job to be done]?”
- Comparison prompts: “[Competitor] vs alternatives for [use case]”
- Constraint prompts: “Affordable [category] tool with [integration] and SOC 2”
- Problem prompts: “Why is our [metric] dropping and what tools help fix it?”
Skip prompts that already contain your brand name. You want to know whether AI brings you up when nobody mentions you.
For example, a startup selling a customer onboarding tool might start with prompts like these:
- “What are the best user onboarding tools for early-stage SaaS?”
- “How do I reduce churn in the first week after signup?”
- “Alternatives to [Competitor] for a small product team”
- “Onboarding software that integrates with HubSpot and Segment”
- “Which product tour tool is easiest to set up without developers?”
Step 2: Run them across assistants
Run every prompt in ChatGPT, Perplexity and Google (check whether an AI Overview appears). Use a fresh or logged-out session where you can, so your own history doesn’t skew the results. Record the following in a spreadsheet:
- Were you mentioned? (yes/no)
- Your position in the list, if you were named
- How you were described, and whether it was accurate
- Which competitors were named
- Which sources were cited, where the assistant shows them
Step 3: Find the patterns
By the end of the week you’ll usually see three things: prompts where you’re missing completely, prompts where you’re named but described wrongly or vaguely, and a handful of competitors who keep appearing. Pay extra attention to the cited sources. They show you which pages and platforms the assistants trust in your category, and that becomes your target list for Weeks 2 and 3.
Week 1 checklist:
- 15–20 buyer-intent prompts written
- Prompts run in ChatGPT, Perplexity and Google
- Baseline spreadsheet filled in
- Top competitors and cited sources listed
Week 2: Fix the sources AI reads
Now you improve the material assistants draw from, starting with your own site and moving outward.
Make your positioning impossible to misread
Assistants have trouble recommending a product they can’t summarise. Look at your homepage and main product pages and ask whether a stranger could answer these questions in one sentence each: What is it? Who is it for? What problem does it solve? How is it different? Replace clever taglines with plain statements. “The onboarding platform for B2B SaaS teams that want users activated in their first session” gives an AI far more to work with than “Delight, reimagined.”
Use the same wording everywhere: website, docs, social profiles, directory listings. Consistency is a signal in its own right.
Publish honest comparison and alternatives pages
Buyers ask comparison questions all the time, and assistants look for pages that answer them. Create pages such as “[Your product] vs [Competitor]” and “Best [Competitor] alternatives”. Keep them fair. Say where a competitor is a better fit and where you’re stronger. Balanced pages tend to be more credible and more useful as a source than one-sided sales copy.
Add structured data
Schema markup helps machines read your pages without guessing. For most SaaS sites the useful types are Organization, SoftwareApplication (or Product), FAQPage and Article. Make sure your name, category, pricing model and key details are marked up correctly and match what’s in the visible text.
Strengthen third-party listings
Your own website is only one voice. Assistants also lean heavily on what others say about you:
- Review platforms such as G2 and Capterra: complete your profile, choose accurate categories and ask happy customers for honest reviews.
- Review roundups and “best of” articles: go back to the sources you noted in Week 1 and contact the authors with accurate, useful information about your product. Don’t send a generic pitch.
- Directories and marketplaces relevant to your niche, including integration marketplaces for the tools you connect with.
The goal isn’t to be everywhere. You want to be present, accurate and consistent in the places that already shape answers in your category.
Week 2 checklist:
- Homepage and product pages rewritten in plain language
- At least one fair comparison or alternatives page live
- Schema markup added and validated
- Review profiles completed and outreach to roundup authors started
Week 3: Reddit, the source AI keeps quoting
Reddit gets cited in AI answers again and again, especially for “what should I use?” questions. That makes sense. Reddit is full of real people comparing tools, sharing frustrations and recommending what worked for them, which is exactly the kind of content assistants find valuable.
For a SaaS startup, this is both an opportunity and a trap. Communities can spot marketing from a mile away and punish it quickly. The approach that works is slow, helpful and human.
Find the right conversations
- Search Reddit for your category keywords, competitor names and the problems your product solves.
- List the subreddits where your buyers actually talk: founder, marketing, developer and industry-specific communities.
- Note which threads show up in AI answers or Google results for your Week 1 prompts. Those are high-value conversations.
Show up as a useful expert
- Read each community’s rules first and follow them. Every subreddit has its own norms on self-promotion.
- Lead with help. Answer the question properly even if your product never comes up. Share frameworks, lessons and honest trade-offs.
- Mention your product only when it really fits and is relevant to what the person asked, and be fair about alternatives.
- Be consistent over time. A history of thoughtful contributions carries far more weight than a burst of activity.
- Listen. What people complain about in your category is free positioning research. Feed it back into the pages you fixed in Week 2.
Reddit isn’t a channel you “hack” in a week. Week 3 is where you build the habit and set up a sustainable rhythm that compounds over the following months.
Week 3 checklist:
- Relevant subreddits and community rules reviewed
- High-value threads from your Week 1 prompts identified
- A weekly time slot set aside for helpful replies
- Recurring buyer pain points noted for your website copy
Week 4: Monitor, measure and iterate
GEO isn’t a one-off project. Models update, competitors publish, and new threads and articles appear every day. Week 4 turns your work into a repeatable system.
Track share of mentions
Re-run your prompt set from Week 1 under the same conditions and compare. The core metric is share of mentions: across all prompts and assistants, how often are you named compared with your main competitors? Track these alongside it:
- Accuracy of how you’re described
- Average position when you are named
- New sources being cited in your category
- Prompts you’ve gained or lost
Set a monthly rhythm
Repeat the audit once a month. Early shifts can be small, and some changes take time to show up in AI answers, so look for trends rather than reacting to every single result. Each month, pick the two or three biggest gaps and decide what to do about them: a new comparison page, a refreshed listing, outreach to a roundup author, or more effort in one community.
Expand your prompt set
As you learn how buyers phrase things, from sales calls, support tickets and Reddit threads, add new prompts. Over time your prompt set becomes a living map of how your market thinks.
Week 4 checklist:
- Prompt set re-run and compared with the baseline
- Share of mentions calculated against your main competitors
- Top two or three gaps chosen for next month
- Monthly audit date added to the calendar
Frequently asked questions
What is the difference between GEO and SEO?
SEO focuses on ranking your pages in traditional search results. GEO focuses on getting your brand named and described accurately in AI-generated answers. They overlap a lot, since clear content, structured data and authority help both, but GEO also covers off-site sources like review platforms, roundups and community discussions that AI assistants draw from.
How long does it take to see results?
It varies by category, competition and how much groundwork already exists. Some fixes, such as clearer positioning or better listings, can show up relatively quickly in assistants that use live search. Others, like building a presence on Reddit, compound over months. That’s why we recommend monthly tracking rather than judging the work after a week.
Can I pay to be recommended by ChatGPT?
Not in the way people usually mean. Organic recommendations come from what the model and its sources say about you. The dependable path is to earn those mentions with clear positioning, useful content and a real presence where your buyers talk.
Do I need to be on Reddit to get recommended?
It’s not mandatory, but Reddit discussions are a frequently cited source for product recommendations, so ignoring them leaves a gap. If Reddit isn’t a fit for your audience, focus on the communities and platforms that are. The same principle of helpful, authentic participation applies.
Can a small startup compete with big brands in AI answers?
Yes, often more easily than in traditional search. Assistants try to match the specific question, so a startup that is clearly positioned for a particular use case or audience can be named next to, or instead of, larger generalist tools.
Want a second pair of eyes on your AI visibility?
You can run this 30-day plan on your own, and we hope you do. If you’d rather move faster, or you want an experienced view of where you’re losing recommendations and why, we can help. At Overhorizon Media we advise SaaS founders on GEO, Reddit and startup growth. Together we’ll audit your AI visibility, prioritise the fixes that matter and build a plan your team can carry out. If you need hands-on execution as well, we can arrange it on request through our partner agencies.
Consulting sessions start from $249/hour. Book a consulting session or explore our services to see how we work.