Your Marketing Team Is Optimizing for the Wrong Search Engine

Ask your CMO what percent of new customers are finding you through AI. Most cannot answer.

That is not a marketing problem. That is a finance problem, because the dollars are still flowing as if Google were the only search engine, and a measurable channel has opened up that your budget does not reflect.

What Has Actually Changed About Search in the Last 12 Months?

The short version: people are routing questions through AI assistants before they ever open Google. "What wineries should I visit in Santa Ynez?" "Which fractional CFOs work with DTC brands?" "What is the best payroll software for a 15-person agency?" These used to be searches. Now they are prompts.

ChatGPT crossed 500 million weekly users in late 2025. Perplexity publishes citations directly next to answers, which means its users often arrive at your site already pre-qualified by the AI response. Claude and Gemini are doing the same thing inside the products people already use at work. The share of research and shopping decisions that starts in an AI chat instead of a search bar has been climbing steadily for 18 months.

Attribution data from DTC analytics platforms tells the same story. Traffic from AI sources is a smaller share of the top of the funnel, but it converts dramatically better than cold social or generic display. When customers arrive from an AI answer, they arrive with intent.

Why Does the CFO Need to Care, Not Just the Marketing Team?

Three reasons, all of them about dollars.

Capital allocation. Your marketing line items still assume Google is the search engine. Your SEO retainer, your content agency, your technical SEO audits are all pointed at ranking in the SERP. None of that is wasted, but if a meaningful share of high-intent discovery is happening in AI chat, you are under-invested in a channel that is measurable and fixable.

Measurement. You cannot make capital allocation decisions on a channel you cannot measure. Most finance teams have never seen a number for AI visibility. That is not because the number is hard to produce. It is because no one has asked for it.

Payback math. If AI-sourced traffic converts at multiples of traditional paid social, the CAC payback window for dollars spent on AI visibility work is much shorter than the dollars spent on topping up Meta. That changes where the next $10K of marketing budget should go, and it changes the ROI story you tell the board.

What Is the Difference Between SEO, GEO, and AEO?

Three acronyms, three different optimization targets. Your marketing team probably only uses the first one.

SEO (Search Engine Optimization) is the classic. Ranking in Google. Measured in clicks, impressions, and average position. Your existing SEO spend is here. It still matters.

GEO (Generative Engine Optimization) is whether AI crawlers can reach your site at all and parse what is there. Unlike Google, which forgives a lot, AI crawlers like GPTBot and ClaudeBot are strict: they do not solve CAPTCHAs, they do not retry, they see empty pages if your content is JavaScript-rendered. The most common GEO failure is a Cloudflare bot challenge blocking the crawlers without anyone realizing.

AEO (Answer Engine Optimization) is whether your content is structurally quotable when an AI does successfully read it. Short, snippet-shaped paragraphs. Question-shaped H2s. FAQ and HowTo schema. The difference between a paragraph an AI paraphrases from memory and a paragraph it quotes directly with a citation.

SEO is necessary. GEO is the minimum bar for being visible in AI answers at all. AEO is what gets you actually quoted. The three are stacked, not substitutes.

How Do You Know if Your Site Is AI-Visible?

Two numbers worth tracking.

Brand mention rate: the share of AI responses that name your company when asked an industry-relevant question. Zero percent means the AI does not know you exist. Ten to twenty-five percent means the AI knows you but does not lead with you. Twenty-five to fifty percent is the "frequently recommended" tier. Above fifty percent means you are a default answer for the category.

URL citation rate: the share of AI responses that link to your actual domain. This is almost always lower than the mention rate, because AI models paraphrase from memory more often than they cite. Watching citation rate climb over time is the single best signal that your content is being quoted, not just remembered.

Both numbers are produced the same way: take a list of 10 to 15 questions a customer might plausibly ask, run them through ChatGPT, Claude, and Perplexity, and score the responses. It takes about an hour. We do it as a quarterly measurement for our own site and for our clients, and the number moves enough quarter over quarter to drive real decisions.

Where Should the Next $10K of Marketing Budget Go?

Not a blanket recommendation. A framework.

Measure first. A baseline AI visibility test tells you which problem you actually have. Without it you are guessing. Running the baseline is the cheapest thing on this list and it almost always changes the prioritization.

If you are at zero percent, the fix is mechanical, not editorial. Allowlist reputable AI crawler user-agents in your bot-protection dashboard. Publish Organization schema with a populated sameAs array pointing to your LinkedIn, Instagram, Crunchbase, and any press. Add an llms.txt at your site root. These are engineering tasks, not ad buys, and they often move the needle more in 60 days than a full quarter of paid social.

If you are in the 10 to 25 percent range, you are visible but not leading. Now the work is editorial: question-shaped pages that directly answer the queries you want to win, an entity graph dense enough that the AI is confident about which domain belongs to your brand, and earned mentions from authoritative third parties that link to your site.

Only after that structural work should you think about paid. Retargeting visitors who arrived from AI citations. Sponsored content on the third-party sites the AIs cite most frequently. Paid layers only compound on top of a structurally visible site.

What Should a CFO Ask in the Next Marketing Review?

Three questions.

What is our current AI visibility number, and when did we last measure it? If the answer is "we haven't," that is the whole answer. Book the measurement before you approve any SEO renewal.

Of the current marketing budget, how much is pointed at Google ranking versus AI visibility? If the answer is one hundred percent Google, the portfolio is not balanced for where the market actually is.

What would it take to move our AI visibility number from X to Y over the next two quarters, and what is the CAC payback on that work? This is the real question. It turns AI search from a vague conversation into a measurable line item with an expected return.

If you cannot answer those three questions by the end of Q2, you are making marketing allocation decisions without the most important signal that has emerged in the last 18 months. That is fixable. It just has to start at the finance seat, not the marketing one.