How to build the business case for GEO

You build a business case for GEO the same way you build one for any marketing spend: state what it costs, state what you can measure, and show how many extra customers it would take to pay for itself. What you can't do is borrow an industry benchmark, because no reliable one exists for how much traffic or revenue AI answers produce.

GEO (generative engine optimization) is the work of getting your business mentioned and cited in AI-generated answers. Its returns show up in four places you can track with your own data: visits referred by AI assistants, conversions those visits contribute to, growth in searches for your brand name, and how often you appear in answers to the questions your buyers ask.

A credible case uses your own baseline, a break-even calculation, and a range of scenarios. It also says plainly what can't be measured. That honesty is what makes a finance team trust the rest of the numbers.

Why a GEO case is harder to make than a paid ads case

With paid search you pay per click and see each conversion. With GEO, three things get in the way.

First, AI exposure can produce no click. Someone asks an assistant for the best accounting firm for restaurants, sees your name, and searches for you the next day. Analytics records a branded search or a direct visit, not an AI referral.

Second, the platforms report different things. Google reports impressions. Microsoft reports citations. Neither reports what was said about you, and as of October 2026 there is no equivalent site-owner report from OpenAI, Perplexity or Anthropic that we could confirm.

Third, the work overlaps with SEO. Google says optimizing for its generative AI features is "still SEO," so a clearer service page may lift regular rankings and AI visibility at once. Separating the two is often impossible, and you don't need to. Make the case for the combined result.

The four value signals you can measure

AI referral traffic

Google Analytics 4's default channel group includes a channel named AI Assistant. Google's channel definitions describe it as the channel "by which users arrive at your site from sources like ChatGPT, Gemini, Deepseek, Copilot, or Grok. It excludes Google's AI Overviews and AI Mode." Classification depends on the visit passing a referrer that matches Google's list, so treat the number as a floor. Our guide to tracking AI traffic in GA4 walks through the setup.

Assisted conversions

Count the leads and sales where an AI-referred session appears anywhere in the path, not only as the last click. Then add a low-cost check that no analytics tool can replace: a "How did you hear about us?" field on your forms and a note in your sales process. When buyers start answering "ChatGPT," you have evidence that analytics would have filed under something else.

Branded search lift

In Search Console's Performance report, the queries data shows what people typed. Watch the searches that contain your business name, month over month. A rise is consistent with growing awareness, including awareness from AI answers, but it doesn't prove the source. Present it as supporting evidence.

Share of answers

Pick a fixed set of questions your buyers really ask. As a suggestion, not a standard, 20 to 50 is enough to show a pattern and still small enough to check by hand. Run them through the assistants your customers use, and record whether your business appears, whether it is described accurately, and who else is named. Repeat on a schedule with the same prompts. Answers vary from run to run, so trends matter more than any single result. Tracking tools automate this, and GEO tools vs. hiring an agency explains what they do and don't cover.

For Google and Microsoft surfaces, use the free first-party reports. Google's Generative AI performance reports show impressions in AI Overviews and AI Mode, and Google says that data is also included in the overall performance report, so don't add the two together. Microsoft's AI Performance report in Bing Webmaster Tools shows citation counts and grounding queries, the phrases the AI used when retrieving content. Microsoft says the grounding query data is a sample.

Set a baseline before you spend

1. Record the last three to six months of AI Assistant sessions and conversions from Google Analytics 4.
2. Export impressions from the Generative AI performance report in Search Console, if the report appears for your site.
3. Note total citations from Bing Webmaster Tools' AI Performance report.
4. Record searches containing your brand name from Search Console.
5. Run your fixed prompt list once and save the answers with the date.
6. Write down your current lead-to-customer rate and average customer value from your CRM or books.

Without this, any later improvement is an anecdote. With it, you have a before and an after.

Start with break-even, not a forecast

A forecast claims to know what GEO will produce. Nobody does. A break-even calculation asks a question you can answer: how many extra customers per month would this spend need to bring in to pay for itself?

The formula is simple. Divide the monthly cost by the gross profit from one new customer. The result is the number of additional customers you need each month. Then ask whether that number is plausible given your baseline.

A worked example (hypothetical)

The figures below are invented round numbers for illustration. They are not benchmarks, market data or results from any real business.

In this hypothetical case, the business needs about 10 extra qualified leads a month from the combined effect of the work. Now compare that with the baseline. If the site currently gets 5 leads a month in total, 10 extra is a stretch and the spend is hard to justify yet. If it gets 150, a lift of 10 is a modest ask.

Then show three scenarios instead of one prediction.

Use your real margin, real customer value and a real quote. For what goes into a quote, see how much GEO and AEO services cost.

What to tell a CFO

Finance leaders don't need to be sold on AI. They need to know the downside, the test, and the exit. Cover these points on one page.

• What we are buying. Hours of specific work: fixing how the site explains what we do, filling content gaps, and correcting inaccurate information about us across the web.
• What it costs. The monthly fee, any tool subscriptions, and internal time for approvals.
• What we will measure. The four signals above, against the baseline, reported on a fixed schedule.
• What we can't measure. Exposure in answers that doesn't produce a click, and the split between SEO and GEO effects.
• What is not guaranteed. Placement in any AI answer. Google states that "indexing and serving aren't guaranteed," and the same uncertainty applies to other engines.
• The review point. A date when we look at the trend and decide to continue, change scope or stop, and the contract terms that let us do that.
• What we keep if we stop. Better pages, cleaner business data and a measurement setup, all of which also serve regular search.

The last point matters. Much of the work improves assets you own, and those remain after the spend stops.

Mistakes that weaken the case

• Quoting a statistic you can't source. One unsupported "X% of searches" claim makes a careful reader doubt everything else.
• Crediting GEO with all organic growth. Say what is shared with SEO.
• Using last-click only. It can hide much of the effect.
• Changing the prompt list. If you swap prompts each month, the trend means nothing.
• Promising a date. Some changes show quickly and others take a long time. How long GEO takes to work explains which is which.
• Reporting visibility as revenue. Impressions and citations are leading indicators. Label them that way.

How the case changes by business type

A local service business with a short sales cycle can connect an AI-referred visit to a booked job fairly directly. A company with a six-month sales cycle and several decision makers can't, and should lean on pipeline influence and self-reported attribution instead. We cover that situation in GEO for B2B companies.

Whatever the model, agree on the KPIs before the work starts. Our overview of KPIs for GEO campaigns lists the options.

Conclusion

A good GEO business case is modest. It sets a baseline from your own data, shows the break-even point, gives a range instead of a promise, and names a review date. That is easier to approve than a confident forecast, and easier to defend later. If you need a view of where you stand before you build the numbers, eSEOspace offers a free GEO audit of your website structure and visibility. More guides are in our GEO and AEO Guide.

Frequently asked questions