ExplainerStrategy

Budgeting for AI Visibility: Where the Money Actually Goes

by 3 min read

Short answerbudgeting for ai visibility

The first question we get from a finance lead is "what does this cost," and the honest answer is "it depends on what is broken." But that is not a budget. So here is how the money actually breaks down and what you can leave out at each level.

We are not going to give dollar figures for a retainer, because they vary by market and scope. What we can do is show the categories, the rough share each takes, and what cutting each one costs you.

The five cost categories

1. Measurement. Building a prompt set that reflects what your buyers actually ask, running it across the major engines on a schedule, and recording who is named and what is cited. This is a tool subscription plus a few hours a week, or part of an agency fee. It is the smallest line and the one most often cut first, which is backwards. Without it you cannot tell whether anything else worked.

2. Technical fixes. Crawler access, JavaScript rendering, schema, llms.txt, and entity consistency. Mostly one-time engineering time. On a simple site this is days. On a large site with a JavaScript framework and no server-side rendering it can be a real project.

3. Content. Pages written to be quoted. This is the largest ongoing line for most companies. The cost is writer time plus subject-expert review time, and the expert time is the part companies underestimate.

4. Off-site presence. Getting into the sources engines cite: Reddit, G2 and other review sites, YouTube, and roundup lists on publisher sites. This is PR and community work. The cost is people time, not media spend. If someone proposes media spend here, they are proposing paid placements, and we recommend against that.

5. Reporting and coordination. Someone turning the measurement into a weekly readout and deciding what to do next. Small, but it is the line that keeps the other four moving.

A rough split

For a company starting from nothing, we tend to see the first six months split roughly like this: technical 30 percent, content 35 percent, off-site 20 percent, measurement 10 percent, coordination 5 percent. After the technical fixes land, that line drops to near zero and content and off-site take most of the budget.

Treat those numbers as a shape, not a rule. A site with a rendering problem that hides every product page from crawlers has a 60 percent technical problem until it is fixed.

What to do at each budget level

Almost nothing. Run the free checks: crawler access, JavaScript visibility, answer readiness, schema, llms.txt. Fix whatever they flag yourself. Write one quotable paragraph at the top of your ten most important pages. Pick 20 prompts and run them by hand once a month in a spreadsheet. This costs a few days of one person's time and finds the biggest problems.

A small monthly budget. Add a proper audit, a tool or service for weekly measurement, and a content plan of two to four quotable pages a month, written by your own expert and edited for structure. Skip off-site for now unless you already have community presence.

A real program. Everything above plus off-site work: a source map of where engines cite in your category, a plan to be present in each, and someone whose job is to execute it. This is the level where share of answer tends to move visibly within a quarter.

What is a waste of money

Paid placements in AI answers, where a vendor claims to insert you into engine results. Bought reviews. Bulk AI-generated content that answers nothing specific. Thick monthly decks. Any tool that promises a "ranking" in a system that does not have rankings.

Also a waste: content before technical. If crawlers cannot read your site, new pages will not be read either.

How to justify it

Tie the budget to share of answer on prompts that lead to revenue. Pick the 20 prompts closest to purchase. Record the starting number. Report it weekly. If it moves, the budget is justified by the number. If it does not move in a quarter, something in the plan is wrong and the measurement will tell you which part.

Our bias, stated

At AEO GEO Labs we sell the audit, the measurement, the content, and the off-site work, so we have an interest here. We also decline paid placements and fake reviews, which costs us some deals. Start with the free tools from AEO GEO Labs and you will know whether your problem is technical or editorial before you spend anything.

Keep reading

More on strategy

All posts

Rather not do it yourself? We'll get you named in AI answers.

Reading is the slow way. In 20 minutes we'll show you who ChatGPT recommends instead of you, and the three fixes we'd make first.

Book a call

Or get the free AI visibility report by email.

no strings, really