Forecasting ROI for AI Search Work When There's No Spend-to-Click Line
A paid search forecast is straightforward: spend more, get roughly proportionally more clicks, until you hit a ceiling. AI search and generative engine visibility work doesn't follow that curve, since the input is mostly hours and the output is citation, not a click you paid for directly, which means a forecast built on the paid model will either wildly overstate or understate what this work will return.
Here's a simpler model built around three inputs, and how to check it against a channel you can already benchmark.
Vendors Covered in this Article
Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.
Why This Doesn't Forecast Like a Paid Channel
Paid spend and paid output move together in a fairly predictable ratio, at least within a normal range. AEO/GEO effort doesn't: a well-targeted page can keep generating citations for months without more hours going into it, while a poorly chosen topic won't improve no matter how much time you add. That makes a straight-line "spend more hours, get proportionally more result" forecast actively misleading rather than just imprecise.
What is a simple model for forecasting AI search ROI?
Instead, forecast off three things: the number of pages you're actively targeting for citation, your best estimate of what share of those pages are currently getting cited based on your prompt audits, and the conversion rate you're seeing on the pages that do get cited. Multiply through and you get a rough range for expected outcomes, which is honest about being an estimate rather than pretending to precision the underlying data doesn't support.
Build the forecast from three inputs:
- Count the pages you are actively targeting for citation, rather than every page on the site.
- Estimate what share of those pages currently get cited, based on your prompt audits.
- Note the conversion rate you are seeing on the pages that do get cited by assistants.
- Multiply the three together to get a rough range, and label it clearly as an estimate.
- Replace the estimated share with your measured share after six months and redo the math.
How do you check the forecast against a paid baseline?
To sanity-check the range, compare it against what the same result would cost through paid search. If the average Google Ads cost per click runs about $5.26 across industries1 and your landing pages convert at a median of 6.6 percent2, you can back into roughly what a similar volume of conversions would cost through paid, and use that as a rough dollar value for what your AEO/GEO forecast is worth if it holds up.
Where a Visibility Tool Fills the Data Gap
The weakest input in this model is usually your estimate of what share of targeted pages are actually getting cited, since that requires ongoing prompt tracking most teams don't have the bandwidth to run consistently by hand. A tool such as Semrush can help fill that gap with more systematic visibility tracking, which makes the forecast's middle input less of a guess.
Revising the Forecast With Real Data
Treat the first version of this forecast as a starting hypothesis, not a commitment. After six months, replace the estimated share of pages getting cited with your actual measured share, and redo the math. Most teams find the real number is lower than their first guess and the conversion rate on cited pages is higher, which shifts the useful lever from "target more pages" to "get more of the pages you already have actually cited."
Comparing Apples to a Genuinely Different Kind of Apple
A common forecasting mistake is comparing this channel's expected return directly against paid search's historical return, as if they're competing for the same dollar in the same way. They're not quite: paid search buys attention on demand, while AEO/GEO content, once it earns a citation, can keep generating that outcome for months without additional spend, which is a different risk and return profile entirely, closer to an asset than a recurring cost.
A fairer comparison treats the AEO/GEO forecast as closer to an investment in an appreciating (or occasionally depreciating, if the page goes stale) asset, and asks not just "what's the return this quarter" but "how long does a successful page keep paying off before it needs a refresh." That's a harder question to answer precisely, but pretending the two channels are directly comparable dollar for dollar understates what a durable citation is actually worth over time.
Share both the forecast and its underlying assumptions when you present it, not just the final range. A stakeholder who can see that the forecast rests on an estimated citation share of one in ten pages is better equipped to judge how much confidence to place in the number than one who's only shown a dollar figure with no visible reasoning behind it.
What Good Looks Like
Good forecasting here means using a three-input model built on targeted pages, estimated citation share, and downstream conversion rate rather than a spend-scaling model, checking the result against a paid-channel baseline, and replacing early estimates with measured data after a few months.
Building The Capability (5-Stage Skill Ladder)
How to Get Started
Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.
Frequently Asked Questions
How far out should we forecast for this channel?
Six months is a reasonable first horizon for this channel. It's long enough to see whether recently published or updated pages start getting picked up, but short enough that you're not committing to a full year of assumptions that could be badly wrong. Revisit and extend the forecast after that first check.
What if we can't estimate the share of pages getting cited at all yet?
Use a conservative placeholder, such as one in ten, and label it clearly as a guess in whatever forecast document you share. A labeled guess that gets replaced with real data in six months is more useful than skipping the forecast until you have perfect information, which for this channel may never fully arrive.
Should this forecast include the cost of the hours spent, or just the expected return?
Include both, since a return number without its cost isn't a real ROI figure. Pair this forecast with the hours-based cost estimate from a budget worksheet so you're comparing a real cost to a real, if rough, expected return.
Sources
Where we quote a benchmark, we show its source. Other figures in this guide are estimates or general guidance, so check them against your own numbers.
- Google Ads Average CPC (All Industries). WordStream 2025 Google Ads Benchmarks, 2024-2025.
- Landing Page Conversion Rate (Median, All Industries). Unbounce Conversion Benchmark Report.
Related Guides
Ranking for High-Intent Queries When an AI Overview Sits Above You
How to structure pages for high-intent, commercial search terms so both traditional rankings and AI answer overviews can actually extract and cite them.
Why AI Citations Show Up as 'Direct Traffic' and How to Fix It
Most AI assistants strip referrer data, so traffic from a citation shows up as direct in your analytics. Here are three practical ways to recover the signal.
Why AI-Referred Visitors Convert Differently (and What to Track)
AI-referred visitors arrive mid-decision, not mid-search. Here's how to compare their conversion rate fairly against organic and paid traffic.
How to Build a Free Calculator People Actually Finish
A step-by-step approach to building an interactive ROI calculator that gets used, captures leads honestly, and doesn't become a maintenance burden.
How to Tell if a Podcast Sponsorship Is Actually Working
A comparison of host-read and produced podcast ad formats, how to pick shows worth sponsoring, and how to measure results honestly.
Splitting Budget Between Paid Search and AI Search Visibility
AI search visibility work costs mostly time, not media spend. Here's a simple worksheet for splitting budget between it and paid search fairly.