How to Make the Business Case for AI Readiness

You've read enough to be convinced that AI search matters for your business. Your boss, your client or your finance lead hasn't. They've heard "AI" attached to a dozen budget requests this year, most of them vague, and they want to know what they'd be paying for and how they'd know it worked.

Most advice on getting buy-in tells you to pitch a full AI search programme: test what ChatGPT says about you, run a 90-day pilot, track citations, report back. That's reasonable advice, but it asks for a bet. You're asking someone to fund work whose result is decided by OpenAI and Google, not by you, and bets are the hardest thing to get approved.

Hidden inside that bet is a much smaller and much easier request: fixing your own website so AI systems can reach it, read it and tell who runs it. That part is in your control, it has a finish line, and you can prove it got done. This post shows you how to separate the two, build the case for the part you can prove, and leave the bet for later, when you have something to show.

Two Different Things You Might Be Asking For

When people say "we should invest in AI search," they usually mean two different things at once. Pulling them apart is the single most useful thing you can do before you write a word of your pitch.

The fix (AI readiness). Work on your own website: letting AI crawlers in, making pages readable, stating clearly and consistently who you are and what you do, and answering the basic questions customers ask. You decide whether it happens. Each item is either done or not done. You can check it the day after.

The bet (AI visibility). Whether ChatGPT, Gemini, Perplexity or Google's AI answers actually name or cite you. You can influence it, but you can't control it. It moves around from week to week, and no one can honestly promise a result.

	The fix (readiness)	The bet (visibility)
What it is	Changes to your own site	How AI tools treat you
Who decides the outcome	You	The AI providers
Does it end?	Mostly a one-off project, plus upkeep	Ongoing
Can you prove it's done?	Yes, item by item	Only as a trend over time
Can anyone guarantee it?	Yes, the work itself	No
Easy to approve?	Usually	Harder

Our post on readiness versus visibility explains the difference in full. For budget approval, what matters is that pitching both as one request tends to go wrong in one of two ways: the person signing off judges the fix by the bet's uncertainty ("you can't promise we'll show up in ChatGPT, so why spend anything?"), or you overpromise to get the bet through and lose credibility three months later. Pitch the fix on its own terms first.

Step 1: Find What's Actually Broken on Your Own Site

The strongest evidence in any business case is a specific problem on your own website. Industry statistics are easy to wave away ("that's not our market"). A screenshot of your own site telling AI crawlers to go away is not.

Start with an AI readiness audit of your site. You're looking for findings that a non-technical reader will understand as a defect, not a nice-to-have. The kinds of findings that land best:

• AI crawlers are blocked. A rule in your robots.txt file (the file that tells automated visitors which parts of a site they may visit) turns away OpenAI's or Perplexity's crawlers, often added years ago by someone who has since left.
• A security service is turning them away. Your firewall or bot protection blocks AI crawlers even though robots.txt says they're welcome. Nothing on the site itself shows this.
• Your business details are missing or inconsistent. Your name, address, phone number or what you sell differs between pages, or isn't stated anywhere a machine can pick it up reliably.
• Key pages can't be read without running code. The content only appears after scripts run in a browser, so some crawlers see an empty page.
• Basic questions go unanswered. Your site never clearly states your prices, your service area, your opening hours or who the company is.

Then add one piece of outside evidence: ask ChatGPT or Gemini a few plain questions about your business and save the answers. If one gets your opening hours wrong, names a competitor when asked for your service, or confuses you with a similarly named company, that's a concrete problem your boss can see for themselves. Keep it to three or four questions, and ask each one a few times, because AI answers change from one ask to the next and a single answer proves little. Use a fresh, temporary chat that isn't shaped by your own history; our guide to checking what ChatGPT says covers how to get a clean result. A wrong answer that comes back on most tries is far more persuasive than a spreadsheet of 50 prompts.

Write each finding as a plain sentence with a consequence. "Our robots.txt blocks GPTBot" means nothing to most people who approve budgets. "Our website tells ChatGPT's crawler not to read any of our pages, so it can't learn anything about us from our own site" does.

Step 2: Use Outside Numbers Carefully

Once you have your own findings, a small amount of outside context helps explain why this matters now. Use it sparingly, and only from sources you've checked yourself.

A few numbers that hold up because they come from the original research:

• People click less when Google shows an AI summary. Pew Research Center tracked the browsing of 900 US adults in March 2025. When a Google results page showed an AI summary, people clicked a regular result link in 8% of visits, against 15% when there was no summary. They clicked a link inside the summary itself in just 1% of visits, and they were more likely to stop browsing altogether (26% versus 16%), according to Pew's analysis.
• The top result loses clicks when an AI Overview appears. Ahrefs compared 300,000 keywords and found that an AI Overview being present was linked to a 58% lower average click rate for the top-ranking page, using December 2025 data (Ahrefs' study). Ahrefs is careful to call this a correlation, not proof of cause.
• AI visitors can be few but valuable. In one 30-day period, AI search sent about 0.5% of Ahrefs' own website traffic but 12.1% of its signups, and those visitors converted at 23 times the rate of regular search visitors (Ahrefs' own data). That's one company, a software company with a very technical audience, and they say so themselves. Present it as "here's what one company saw," never as "here's what we'll see."

What these numbers support is modest but useful: more of the first contact between customers and businesses now happens inside an AI answer, before anyone reaches a website. They don't support a forecast of your traffic or revenue, and you shouldn't use them as one.

Check every number at its source before it goes on a slide. Statistics about AI search get passed from article to article and quietly change along the way. While researching this post, we found a vendor guide repeating Ahrefs' conversion figure as "23% higher than organic search." Ahrefs' own post says 23 times. That's the difference between a modest bump and a claim your finance lead would rightly question. If you can't find a number's original source, leave it out. One unverified figure that gets challenged in the meeting can undo the rest of your case.

Leave out predictions dressed as facts. "Search traffic will fall by X% next year" is a forecast, and forecasts about AI search have a poor track record. Your case is stronger built on things that are already true about your own site.

Step 3: Work Out What the Fix Actually Costs

People who approve budgets want to know three things about cost: how much, for how long, and whose time. Vague answers here kill more requests than anything else.

List the fixes, then cost each one. Take the findings from Step 1 and turn each into a task with an owner. A typical list might look like this:

Finding	Task	Likely owner
AI crawlers blocked in robots.txt	Edit the file to allow the crawlers you want	Developer or whoever runs the site
Firewall turns away AI crawlers	Change the bot rules, then confirm with the crawler's published addresses	IT or security
Business details inconsistent	Agree one correct version, update the pages	Marketing, signed off by the owner
No structured business information	Add structured data (a standard, machine-readable description of who you are)	Developer or a tool
Missing answers to common questions	Write or expand an FAQ or service pages	Marketing or content

Ask each owner for a rough estimate in hours. Don't invent the numbers yourself; an estimate from the person who'll do the work is far more believable, and it gets them on board early.

Separate one-off work from upkeep. Most readiness work happens once: you fix the robots.txt file, correct your business details, add structured data. A smaller amount continues: new pages need the same treatment, and site updates sometimes undo earlier fixes (a redesign that drops your structured data, or a new security setting that blocks crawlers again). Show both, so nobody is surprised by the second bill.

Include the tool cost if you're using one. Some of this work can be done by hand by a developer. Some of it, especially structured data across many pages and checking that nothing has broken since, is easier with a tool. As one example, our own pricing starts with a free plan that includes the full readiness report and fixes for up to 25 pages, and paid plans start at $19 a month for larger sites and weekly re-checks. Whatever you use, put the real monthly figure in the case, along with how to cancel.

Flag the hidden approvals. Two items often need sign-off from someone outside marketing, and finding that out late can stall the project for weeks:

• Firewall changes need IT or security. If your security service is blocking AI crawlers, someone with access to it has to change the rule, and they'll want to know exactly which bots and why. Our firewall guide covers the questions they'll ask.
• Business details need an owner's sign-off. Which phone number is the main one? Which address? Is the founder's name spelled the same everywhere? These sound trivial and regularly take longer to agree than the technical work.

Step 4: Say What You Will and Won't Promise

This is where most AI search pitches go wrong, and where yours can stand out by being honest.

What you can promise is the work itself, with a date:

• Every finding on the list fixed by a set date.
• AI crawlers confirmed able to reach your key pages.
• One correct, consistent version of your business details across the site.
• A readiness score (or a checklist) that's measurably better than before, re-checked to show it stays that way.

What you can't promise is what AI tools do with your site afterwards. No one can promise that ChatGPT will start recommending you, that Google's AI Overviews will cite you, or that a specific amount of traffic will arrive. Anyone who does promise that is guessing, and the person approving your budget will probably know it.

Being upfront about this makes your case more believable, not less. It shows you understand the difference between what you control and what you don't, which is exactly the judgment someone wants from the person spending their money.

The fix is low-regret. Here's the argument that tends to settle it: the fixes are mostly the same things that make a good website anyway. Google itself says there are "no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary" (Google Search Central). In other words, the fix is ordinary good-website work that AI tools happen to depend on. If AI search turned out to matter less than everyone expects, you'd still have a better website. That's a much easier thing to approve than a bet.

Step 5: Name the Cost of Doing Nothing, Without Scare Tactics

Every business case needs an answer to "what happens if we don't?" For the fix, the honest answer is concrete:

• If crawlers are blocked, AI tools can't learn about you from your own site. They'll describe you from whatever else they find: directory listings, old reviews, a competitor's comparison page.
• If your details are inconsistent, AI tools may repeat the wrong ones to customers who never visit your site to check.
• If basic questions go unanswered on your site, AI tools answer them from somewhere else, or not at all.

These are costs that exist today, on your current site. You can point at each one.

What you should avoid are the vague warnings that fill a lot of AI search pitches: "competitors will lock up AI answers forever," "this is the last chance to get in early," "by next year it'll be too late." There's no good evidence for any of them, and the person reading your case has heard the same lines about every new channel for 20 years. A specific defect on your own site is far more persuasive than a general fear.

Step 6: Plan How You'll Report Back

Before the work starts, agree what you'll show at the end and when. This turns a request into a commitment with a review date, which is far easier to say yes to.

A simple before-and-after report:

• The fix list. Each finding, its status, the date it was done.
• The readiness check, re-run. The same audit as Step 1, showing what changed.
• AI referral traffic. Visits that arrive from ChatGPT, Perplexity, Gemini and similar tools. Most sites can see this in their existing analytics; our guide to tracking AI traffic in Google Analytics shows how. Expect this number to be small, and don't present it as proof either way in the first month or two.
• The same three or four questions from Step 1, asked again the same way and the same number of times, with the answers saved. If the wrong opening hours now come back right most of the time, say so. If nothing has changed yet, say that too.

Set a decision point. At the review, the question isn't "did it work?" in some general sense. It's: "The fix is done and verified. Do we now want to fund the bet, an ongoing effort to track and improve how AI tools mention us?" That's a decision your boss can make with real information, rather than one you asked them to make on faith at the start.

Tailor the Case to Who Signs Off

The same facts need a different emphasis depending on who's reading. A few common readers and what they care about:

A small business owner usually cares about time and money, in that order. Lead with the one or two findings they can see for themselves (the wrong answer from ChatGPT, the blocked crawler), the total cost, and how long it'll take. Keep it to a page.

A marketing lead cares about how customers find the business and how that compares with competitors. Lead with what AI tools say when asked for your service, who they name instead of you, and how the fix makes your own site a clearer source.

A finance lead cares about whether the cost is fixed or open-ended, and whether it can be stopped. Lead with the cost table from Step 3, the split between one-off and ongoing costs, and the review date. Avoid revenue forecasts entirely unless you can source every number in them.

IT or security cares about what changes and what the risk is. Lead with exactly which crawlers you want to allow, what they're allowed to read, and how you'll confirm they're the real crawlers and not something pretending to be them. Make it clear that you're not asking to switch off protection, only to adjust which automated visitors get through.

A client, if you're an agency, cares about what they're getting and whether you're overpromising. The fix-versus-bet split works especially well here: it lets you sell a defined project with a clear end, and offer ongoing tracking as a separate, optional service.

A One-Page Template You Can Copy

Here's a structure that fits on a single page. Fill in each line with your own findings.

Request: Approve [number] hours of work and [tool cost, if any] to fix how our website presents itself to AI search tools, finished by [date].

Why now: [One or two sentences on how customers increasingly meet businesses inside AI answers, with one sourced number.]

What we found on our own site:

1. [Finding, in plain words, with its consequence]
2. [Finding, in plain words, with its consequence]
3. [Finding, in plain words, with its consequence]

What we'll do: [The task list from Step 3, with owners.]

Cost: [One-off hours and cost] plus [ongoing monthly cost], which can be stopped at any time.

Needs sign-off from: [IT for firewall changes, owner for business details, etc.]

What we will deliver: Every finding above fixed by [date], confirmed by re-running the same check.

What we can't promise: How often AI tools mention or cite us. That's decided by the AI providers, not us.

How we'll report back: A before-and-after report on [date], covering the fix list, the re-run check, AI referral traffic and the same test questions asked again.

Decision at review: Whether to fund ongoing tracking and improvement of how AI tools mention us.

If your case runs past one page, it's usually because the findings section has turned into a technical report. Keep the detail in an attachment and the page itself in plain words.

Objections You'll Hear, and Honest Answers

"Isn't this just SEO with a new name?" Partly, and that's good news. A lot of the fix overlaps with work that helps regular search, which is why it's low-regret. The parts that are specific to AI tend to be the ones nobody had a reason to check before: whether AI crawlers in particular are allowed in, and whether your site states who you are clearly enough for a machine to be sure.

"AI traffic is tiny for us." Probably true. Check your analytics and say so honestly. The case for the fix doesn't rest on traffic. It rests on specific defects on your site and on what AI tools tell people who never click through at all.

"Can you guarantee we'll show up in ChatGPT?" No, and you should say so plainly. You can guarantee that your site no longer stands in the way. What the AI tools do after that is up to them.

"Let's wait a year and see." That's a fair answer for the bet. It's a weaker one for the fix: a blocked crawler or a wrong phone number costs you today, and waiting doesn't make the repair cheaper. If your boss wants to wait, offer to do the fix now and hold the bet for a later decision.

"We blocked AI bots on purpose." That can be a deliberate and reasonable choice, for example to stop your content being used to train AI models. It's worth knowing that many AI companies run separate crawlers for training and for answering live questions, so it's often possible to block one and allow the other. The point of the business case isn't to overturn that decision. It's to make sure it was a decision and not an accident.

Common Mistakes

Leading with the trend instead of your own site. Opening with "AI is changing everything" invites the reaction you're trying to avoid. Open with something specific and true about your website.

Promising citations or traffic. It might get the first approval. It almost always costs you the second one.

Asking for the bet before the fix. An open-ended visibility programme is a hard first ask. A defined fix with a finish line is an easy one, and it gives you results to point at when you ask for more.

Repeating statistics you haven't checked. Numbers about AI search change shape as they're passed around. Go to the original study, quote what it actually says, and note who it covers.

Leaving out the people who have to approve parts of it. Discovering in week three that only IT can change the firewall, or that nobody agrees on the company's main phone number, makes the whole project look unplanned.

No date and no owner. "We'll improve our AI readiness" isn't something anyone can approve. "Developer fixes items one to three by the 30th, marketing fixes four and five by the 15th of next month" is.

Leaving re-checks out of the budget. If the plan has no line for checking again later, the first redesign or security change that undoes a fix goes unnoticed. Put a re-check date in the case itself.

FAQ

What is a business case for AI readiness?

It's a short document that asks for time or money to fix how your website presents itself to AI search tools. A good one states what's wrong on your site, what it costs to fix, what will be delivered and when, and what can't be promised. It's usually easier to approve than a broader "AI search strategy" because the work is defined and can be checked when it's done.

How much does AI readiness cost?

It depends on what's broken. Many fixes are small: editing a robots.txt file, correcting business details, adding structured data. The biggest costs are usually people's time and, if you use one, a tool to handle structured data and re-checks across many pages. Get an hours estimate from whoever will do each task, and show one-off and ongoing costs separately.

How long before we see results?

Most of the fix shows results quickly: once a change is live and the site has been checked again, you can see each item resolved. A few items take longer to show their full effect, such as getting outside profiles and directories to match your corrected details. Changes in what AI tools say about you take longer and can't be scheduled, because they depend on when each AI tool next reads your site and how it decides what to use. Plan a review a few weeks after the fix and treat any change in AI answers as a bonus at that stage, not the test.

How do I measure the return on AI readiness?

Measure the fix and the bet separately. For the fix, the return is the list of defects removed, confirmed by a re-check. For the bet, look at visits from AI tools in your analytics, whether the test questions you asked at the start now get correct answers, and, over a longer period, whether new customers mention finding you through an AI tool. Don't promise a revenue figure at the start.

Do we need an agency to do this?

Not necessarily. Much of the fix is within reach of whoever already runs your website, especially with a tool that handles structured data and re-checks. An agency can make sense if your site is large, if several teams need coordinating, or if you also want ongoing visibility work. Either way, the business case looks the same: findings, tasks, cost, what's promised and what isn't.

Start With Your Own Evidence

The fastest way to build this case is to see what's actually wrong on your site. Run our free AI readiness checker on your homepage: it checks whether AI crawlers can reach you, whether your pages can be read and whether your business details are clear, and gives a first look at how well your site answers customer questions. Each finding comes in plain words, ready to drop into Step 1 of your business case. A free account adds the full report.