The business case

AEO plays on the word-of-machine effect. People trust AI more when they’re buying for practical, functional reasons, and trust people more when they’re buying for taste or experience.

Even though AI search naturally does grow as the LLM user base grows, you can do work that influences the outcome. Without it, buyers might:

Most CEOs know how important AI visibility today, but the business case will be about how to spend on it.

Start with who uses answer engines while buying. Surveys disagree on the size, which is itself worth showing a finance lead.

Then make it about your category. Run ten buying prompts through the engines your buyers use. Record which brands get named in each answer. A table with your name missing from prompts your competitors win is the strongest slide you will show.

Companies receive customers through AI answers without doing any AEO work. That baseline belongs in your case. It proves the channel exists for you. It also stops anyone crediting a program for results that were arriving anyway.

If you are already getting AI referral traffic and customers saying they heard about you through AI, it shows that there's something worth investing in.

Clicks and signups can part ways. One Flying Cat client saw traffic rise 32% while signups rose 75% over six months. Another saw traffic fall 40% in a quarter while demo requests rose 70%. Build the case on signups and demos, with clicks as supporting context.

Accuracy is another area of AI visibility worth investing in. I have a client that is a huge name brand, so they're mentioned 100% of the time, almost. But they've expanded/change categories recently, and the internet mostly still knows them for what they used to be.

I was able to build an entity analysis worksheet and fact checker and found that in their market, only 33% of buyers were told what they currently do. This helped build the case for investing in AEO even if they were already highly mentioned.

Every solid business case compares options, including the option to keep things as they are. There are many options out there (and the next chapter is about how to use a vendor!)

Doing this well, in-house, requires some level of senior expertise, alongside some execution. The role that decides whether AEO works is the lead, someone tapped into the brand and the market who oversees on-site work, outreach and brand activity and sets priorities across teams. Specialists can carry much of the execution, but they need that lead above them.

In October 2026 we searched LinkedIn for people at SaaS companies whose titles or role descriptions include AEO. We took the three companies below as examples of a built up AEO department. Every title below comes from the employees' own profiles. The reporting lines are our inference from those titles and descriptions, not actual reporting lines. Salaries are from Glassdoor range data.

There may be a strong temptation to let Claude do all of your AEO: strategy, updates, outreach. While Claude is an extremely powerful tool that can do a lot of work for you, it can make some judgement calls that, if you don't know any better, sound great but are ineffective at best, harmful at worse.

Claude can help significantly speed up:

• Mass data analysis
• Operations
• Outreach research and verification against a specific, given set of guidelines
• Contextual internal linking (which used to be extremely time consuming but is pretty high impact)
• The work of a junior SEO specialist
• Mass updates across the website of a specific thing chosen by an expert

But the best way to use Claude is with someone who knows what they are doing, who has a deep knowledge of your brand and a pulse on the market and AEO/SEO practices. Trust me, I've tried to use Claude to automate any aspect of my work. And while it can really help get insights from data, it doesn't have enough experience to make good judgement calls. Remember - it takes from the most average probability.

Anybody who doesn't know anything about a topic will think whatever Claude creates about it is great, and anyone who works deeply in it will know that most of its strategic and creative output is bang-average. It works really well for operations.

Writing is the worst use case for AI. We use AI every day for research, briefs, competitor tracking, accuracy checks and updating pages at scale. All people find it painful AND painfully obvious to read AI-generated content, even if you edit it. Agents speed up execution, but a team that a $20 subscription could replace was weak to begin with.

AEO happens across your whole site and the wider web. Teams who expect it to run from the blog, the way SEO often did, end up with too little to show. A program needs four kinds of work, and the chapters in the Do section cover each one in depth.

This is why an AEO retainer looks different from an SEO retainer built around a monthly quota of blog posts. There are also some things that an AEO agency just can't do alone (or the cost of that investment would be so high, and span so many departments, it's a bit of a hard sell).

AEO's popularity has drawn some new risky band-wagon-jumpers. Some SEOs claim AEO expertise without changing how they work, doing traditional SEO under the AEO name. Some newcomers jumped on the trend with little SEO knowledge, and their advice can damage a site that took years to build. Google's own guidance from May 2026 treats AEO as part of SEO and names several popular AEO tactics it calls unnecessary.

Good AEO takes an expert who understands SEO, brand and marketing, and who knows what to prioritize. Chapter 10 covers how to vet a partner.

List every cost first. Then, work out what the program has to return to pay for itself.

Every dollar benchmark we found comes from a company that sells the service, so treat these as market prices.

We like to split an AEO budget evenly between content on your own site and outreach for third-party mentions.

AEO works more like brand building than a funnel, so we can't work our way backward from a revenue goal like we used to with SEO. Les Binet and Peter Field studied decades of advertising effectiveness data and found that brands with a share of voice higher than their market share tend to grow, and brands below it tend to shrink. They coined this gap excess share of voice. In their B2B research with LinkedIn's B2B Institute, 10 points of excess share of voice went with about 0.7 points of market share growth a year.

AI answers are a new place to have a voice, so the same logic applies there. We call it excess share of AI voice.

Branded search stands in for market share, since Binet's later research showed that share of search tracks share of market and often moves first. To calculate it, follow these steps.

Walk through an example. Say you track 50 unbranded prompts about your category across the main AI assistants. Your brand appears in 12% of the brand mentions those answers contain, and your three main competitors split the rest. In a keyword tool, your brand name gets 20% of the branded searches across the same four brands. Your excess share of AI voice is 12% minus 20%, or −8 points. AI answers give you less voice than your brand demand would predict, so buyers who start their research in AI hear about your competitors more often than they would in Google.

A competitor with 30% of AI mentions and 18% of branded search sits at +12 points. Binet and Field's research suggests that brand is the one positioned to gain ground over the next few years.

How to use it in a business case. Treat the gap as a direction and a trend. A negative gap is the cost of doing nothing, stated in brand terms, because it shows where competitors are building demand you are missing. A gap that moves toward zero, then above it, over several quarters is the sign the program is working, well before closed revenue shows it. Changing share of voice is a lagging indicator that takes a long time investing to change, though.

How it compares with Binet and Field. Their rule rests on hundreds of campaigns over decades, with paid share of voice measured against audited market share. Excess share of AI voice borrows their logic for a newer channel and uses branded search as a stand-in for market share. Use the direction as your signal, pair it with pipeline and conversions, and leave the 0.7 figure out of any revenue forecast until there are years of AI answer data to test it against.

Check your data before you trust the gap. Branded search breaks down when existing users search a brand to log in. A learning platform used by students, for example, sees its branded searches rise and fall with the school year. Brand names that are also common words cause the same problem. As a quick check, the brands in the top half of branded search should also sit in the top half of AI answers. When they do, the gap is worth reporting.

Leadership approves reversible bets more easily than open-ended ones. Jeff Bezos called these two-way doors, decisions you can walk back through if they go wrong. Discovery-driven planning works the same way and releases money as milestones are met. We recommend an intensive three-month sprint followed by a lighter maintenance phase, with a review at each checkpoint.

Test the assumptions that decide the case. Douglas Hubbard found that in a typical business case, only a handful of uncertain variables are worth measuring. For AEO they tend to be how many of your buyers ask AI, how often you show up today, and how those buyers convert. Measure those first, using the Recommendable Framework from chapter 8.

Timing helps. Ask during the fourth-quarter planning cycle, since money requested mid-year has to come out of a budget that was already approved for something else. Keep the case to one page, or a short 5-deck presentation, built from the parts above.

These came up again and again across client calls, LinkedIn threads and the webinars we reviewed.