More AI visibility is not the return.

These can all be positive AEO outcomes: more brand mentions; stronger recommendation position; more citations; better answer accuracy; greater AI share of voice; more ChatGPT referral traffic. None of them is ROI. They are indicators that the system may be moving in the right direction.

ROI begins when those changes create measurable commercial value. That distinction matters because it is very easy to build an AEO report that looks impressive without answering whether the work deserves another rand of investment.

Return needs a time dimension. A simple CAC payback calculation is:

That is commercially very different from the same CAC with a 24-month payback. Marketing reporting often celebrates the revenue number. Finance cares how long the cash takes to come back. Both perspectives matter.

The relevant question is not "Is our AI visibility increasing?" It is "Given another R100,000, where should we put it?" Potential options may include: AEO/GEO; Google Ads; Meta; LinkedIn; SEO; CRO; content; partnerships; sales enablement; product improvements.

If another channel can create higher-quality customers faster and at better economics, it may deserve the investment instead. That does not make AEO unsuccessful. It makes capital allocation rational.

Attribution asks: Which observable marketing interaction gets credit? Incrementality asks: Did the marketing create an outcome that would otherwise not have happened? That second question is harder.

Suppose a customer clicked through from ChatGPT. They already knew your company. Would they have bought regardless? Maybe. Or suppose a customer ultimately arrived through branded Google search after first discovering you through an AI recommendation. Did AI create incremental demand? Possibly.

This is why attribution and incrementality should not be treated as the same thing. For large enough programmes, incrementality can eventually be investigated through techniques such as: controlled market tests; holdouts; staggered implementation; exposed vs unexposed cohorts; time-series analysis.

For most businesses beginning AEO, that level of experimentation will not be practical immediately. Start with good attribution. Do not pretend it proves causality.

Imagine AI referrals increase 50%. Leads increase 30%. Revenue increases 20%. Excellent. But during the same period: paid search spend doubled; the website was redesigned; a new PR campaign launched; the sales team grew; pricing changed.

What exactly caused the result? This is why the baseline needs more than a starting AI visibility number. Track the broader commercial environment too. Otherwise every positive business movement becomes conveniently attributable to the latest marketing initiative. That is not a business case.

For a recurring-revenue business, a customer may continue generating value for years. If you calculate ROI using lifetime value, be careful. Forecast LTV is not realised cash. Use assumptions you can defend. Useful versions include:

This matters particularly early in a programme. Suppose: AI visibility is improving; citations are strengthening; new commercial pages are beginning to rank; referral traffic is increasing; qualified leads have started appearing; no deals have closed yet. ROI is still negative.

That does not automatically mean the programme has failed. It means it has not yet paid back. The question then becomes whether the leading indicators and sales cycle justify continued investment. A six-month enterprise sales cycle cannot reasonably be evaluated as if it sells shoes online. Measurement needs to reflect how the business actually makes money.

AI search is new. The economics are not. The business still invests money. Customers still buy things. Products still have margins. Sales cycles still take time. Cash still matters.

So I would evaluate AEO and GEO the same way I evaluate every other acquisition investment: What did it cost? What behaviour changed? What commercial value followed? How confident are we in the connection? How quickly did the investment come back? Would the next rand be better spent here or somewhere else? That is the standard.

QFD works across both sides. On one side: Can AI systems correctly understand, cite and recommend the brand? On the other: Can the business measure what happens when they do?

That means the work can extend from prompt visibility, entity architecture, citation strategy, content, technical implementation through to analytics, lead-source capture, CRM attribution, opportunity tracking, revenue, unit economics.

Because the interesting result is not "We got your brand mentioned in ChatGPT." It is "We can see what changed, what demand followed, what became revenue and whether the economics justify further investment."

That is the bridge between AEO & GEO Consulting and Revenue Measurement Architecture. If you want to see where the connection breaks in an existing setup, start with an AEO/GEO Visibility Audit, and if you want the measurement discipline behind it, read How to Measure AEO Performance or the story of The Attribution Detective.