AEO Budget Is Expansion, Not Subtraction: The CFO Math

The dominant story about AEO budgets is a subtraction story: SEO is dying, so move the money. That story is wrong, and if you carry it into a CFO meeting it will cost you the budget. The smart money treats GEO as expansion — a parallel investment topped up onto search, not a knife taken to it — and the finance math backs that framing hard.

• The reallocation is expansion, not a knife. PMG recommends piloting GEO at 1.5–2x the current search budget. “If a brand was spending $10,000 a month on SEO, they would now expect at least half of that budget to also cover GEO” (Noise Media Group). One pet-food brand kept SEO and added 30–35% experimental GEO spend on a new line (Pawco, via Digiday). Nobody winning is gutting search.
• The market has already moved. 12% of enterprise digital budgets went to AEO/GEO in 2025; 94% of CMOs plan to increase AEO investment in 2026; 55% of marketers already have GEO dollars allocated (Conductor; Scribewise/Digiday). You’re not asking to bet on something unproven — you’re asking to keep pace with your peers.
• The CFO clincher is the conversion gap. AI-referred traffic converts at 14.2% vs 2.8% for Google organic — 4–5x (Opollo, 312 technology-service firms). Semrush independently measured the average AI visit as 4.4x as valuable. This makes GEO a revenue argument, not a cost line.
• The real cost is measuring the wrong thing. Every analyst-hour spent producing rank-tracking decks measures a channel that, for high-intent buyers, converts at a fraction of the AI channel — while the metric that predicts next-quarter pipeline goes untracked.
• Speak the CFO’s two languages. Share of Voice is demand creation (how much demand exists for you); ROAS is demand harvesting (how efficiently you capture demand that already exists). They are different jobs. And Binet & Field’s IPA databank proves 10 points of excess share of voice predicts ~0.5–0.7% annual market-share growth — a forty-year-old law, now on a new surface.
• Connect citation share to the three things boards already fund: pipeline, competitive position, and total addressable audience. A budget ask framed as “improve our SoV” dies; framed as “we’re recommended 38% vs the leader’s 55%, AI buyers convert at ~5x, here’s the pipeline math” — it survives.

Who should read this: the in-house marketer (Avatar 6) who needs to win a GEO budget from a skeptical boss or CFO, and the B2B operator-founder (Avatar 5) who is that skeptic and wants the honest finance case before they sign. This is the post you forward to win the room.

1. The Framing Mistake That Loses the Budget

Most marketers walk into the budget conversation having already lost it, because they’ve accepted the wrong frame. The headline narrative of 2026 — “SEO is collapsing, AI is the future, move the money” — sounds bold in a LinkedIn post and lands like a threat in a finance meeting. Because to a CFO, “move the money from SEO to AEO” parses as: the thing we’ve funded for years was a mistake, please now fund a new thing I can’t yet prove. That’s a confession dressed as a strategy. It invites exactly the response that kills the ask: “Let’s wait and see.”

The operators who are actually winning AI visibility didn’t make that mistake. They reframed the entire conversation from subtraction to expansion — and the data shows that’s not spin, it’s literally what the smart money is doing. From Digiday’s reporting on how marketers are actually allocating:

If a brand was spending $10,000 a month on SEO, they would now expect at least half of that budget to also cover GEO. — Joe Levi, Co-founder/CEO, Noise Media Group

Read the word also. Not instead. The $10k SEO budget doesn’t shrink to $5k SEO + $5k GEO. The expectation is that at least $5k of additional GEO scope sits on top of the existing search investment — because the two do different jobs and feed each other. PMG’s recommendation is even more explicitly expansionary: pilot GEO at 1.5 to 2x the current search budget (Matt Allfrey, PMG Head of SEO EMEA, via Digiday). VML’s global chief innovation officer describes “increased testing and experimentation within search budgets — specifically allocation to GEO” (Brian Yamada, via Digiday). The pattern across every source is the same: topping up, not cutting over.

This is the single most important move in the whole post, and it’s a framing move before it’s a math move: you are not asking the CFO to abandon a sunk investment. You are asking to fund a parallel, higher-converting channel at the same moment every competitor is doing the same. That’s a fundamentally different — and fundable — request. The rest of this post is the evidence that makes it stick.

2. The Market Already Voted: You’re Keeping Pace, Not Gambling

A CFO’s first instinct on any new-channel ask is “is this a fad I’ll regret funding?” The adoption data answers that directly, and it’s the cheapest credibility you can put on the table.

• 12% of enterprise digital budgets were already allocated to AEO/GEO in 2025 (Conductor, 250+ enterprise respondents).
• 94% of CMOs plan to increase AEO investment in 2026 (Conductor). Not “consider” — plan to increase.
• 56% made “significant or high” AEO investments in 2025 (Conductor).
• 55% of marketers say they already have dollars allocated to GEO, and 70% of those say GEO accounts for 11–20% of their budget (Scribewise, 205 executive leaders, via Digiday).

(All figures mid-2026, perishable — date-stamp them; adoption percentages move fast.)

Here’s how to use this in the room. The fad objection assumes you’re early and exposed. The data says the opposite: the budget shift is already mainstream, and the risk has flipped. The risk is no longer “we funded a fad.” The risk is “every competitor funded the channel where half our buyers now start, and we sat it out.” When 94% of your peers’ marketing leaders are increasing this exact line item, the unfunded position isn’t prudent — it’s conspicuous. You’re not asking the CFO to gamble on the frontier; you’re asking them not to be the one company on the sideline.

And the concrete case beats the percentage. Pawco, a pet-food brand, increased its budget 10% in Q1 2026 for LLM-discovery testing and allocated 30–35% of its new brand’s budget as experimental GEO spend (Ryan Bouton, VP of Growth, via Digiday). The telling narrative detail: budget that previously funded SEO content production was “expanded to fuel AEO expert partnerships and platforms.” Expanded. Not redirected. That’s the word that wins the meeting.

3. The Clincher: AI Traffic Converts 4–5x. This Is a Revenue Argument.

Adoption data answers “is it a fad.” This section answers “will it make money” — and it’s the part you lead with if your CFO is numbers-first.

The single most powerful fact in the budget case is the conversion gap. From Opollo’s 2026 benchmark across 312 technology-service firms:

AI-referred traffic converts at 14.2% versus 2.8% for Google organic.

That’s a 4–5x conversion gap. Semrush measured the same phenomenon independently and from a different angle: the average AI-search visit is 4.4x as valuable as a traditional organic visit, by conversion rate (Semrush, 500+ high-value topics). Seer Interactive’s platform-by-platform data (June 2025) shows the same shape — ChatGPT visitors converting at 15.9%, Perplexity at 10.5%, against Google organic at 1.76%. And the mechanism, documented across the dark-funnel research, explains why: an AI engine delivers a brand to the buyer as a recommendation — synthesized, contextual, framed as an answer — so the buyer arrives already persuaded, not at the top of the funnel but near the bottom of it. (Ahrefs found this on its own site: 0.5% of its visitors, the ones from AI, drove 12.1% of signups — a 23x concentration.)

(Conversion figures are mid-2026 and perishable — cite with date; the durable point is the gap and its mechanism, not the exact percentages.)

Now translate that into the CFO’s language. Two channels:

A CFO does not fund “AI visibility.” A CFO funds the channel that converts at five times the rate of the one we already pay for. Same product, same sales team, dramatically higher conversion — because the buyer was pre-sold by the recommendation. That’s not a marketing pitch; that’s unit economics. When you frame GEO spend against that table, the question stops being “can we afford this?” and becomes “why are we under-funding the highest-converting channel we have?”

One honest caveat, because the operator-founder reading this will think it and you should say it first: the conversion premium concentrates in high-consideration / research-heavy verticals and weakens for impulse e-commerce (Semrush is explicit about this). That’s not a weakness for Biostack’s buyer — it’s a bullseye. The complex, high-ACV B2B operator is exactly the profile where the AI conversion premium is largest. Concede the limit, then point out you’re standing in the center of where the limit doesn’t apply.

4. The Cost of Measuring the Wrong Thing

Here’s the reframe that turns the budget conversation inside out — and it’s the line Patel gestures at without finishing. The real cost in front of the CFO isn’t the GEO spend. It’s the measurement you’re already paying for.

Every quarter, your team spends real analyst-hours producing a rank-tracking board deck — pulling keyword positions, charting organic-traffic lines, building cost-per-lead tables. All of that labor measures a channel that, for your high-intent buyers, now converts at 2.8% while the unmeasured AI channel converts at 14.2%. You are paying people to carefully measure your lower-converting channel, while the metric that actually predicts next-quarter pipeline — your presence in the AI shortlist — goes untracked entirely.

That’s the opportunity cost. Restate it for the CFO in their own framing: “We’re spending measurement budget characterizing a channel that converts at a fifth the rate of the one we’re not measuring. The cost isn’t the GEO program — it’s that we’re flying blind on the channel that’s actually feeding the pipeline.” Misinformation in AI answers compounds this; one CMO-reporting framework calls brand-accuracy errors in AI “revenue leakage, not a content error” (akii.com) — reframing the fix as P&L protection rather than a marketing nicety.

This is also where the DORA J-Curve becomes the honest expectation-setter. Adopting a new capability (here, AI-visibility measurement and the work to move it) produces a temporary dip before the gain — the J-Curve. The teams that win are the ones who fund through the trough instead of cutting at the bottom of it. (The full two-groups argument lives in the DORA J-Curve post; the budget-relevant point is: tell the CFO the curve dips before it lifts, so the month-three “is it working yet?” question is answered before it’s asked.) Cutting the GEO budget at the bottom of the J-Curve is the most expensive thing a team can do — it pays all the setup cost and captures none of the compounding return.

5. Demand Creation vs Demand Harvesting: Speak the CFO’s Two Languages

The deepest reason budget asks fail is a category error: marketers pitch Share of Voice as if it were a performance metric, and CFOs evaluate it as one — and it loses, because as a short-term ROAS proxy it is weak. The fix is to name the two different jobs explicitly. From Orr Consulting:

ROAS is a measurement of how efficiently you are harvesting demand that already exists, while share of voice determines how much demand exists for you in the first place.

That distinction is the whole game in a finance conversation. Demand harvesting (paid search, retargeting, ROAS-measured channels) captures buyers who are already in-market. Demand creation (Share of Voice, brand presence, AI citation share) determines how many buyers think of you — are pre-disposed to you — before they’re in-market. They’re not competing line items; they’re sequential. Harvesting with no creation is squeezing a shrinking pool harder. Creation is what refills the pool.

A CFO already funds both halves elsewhere; they fund demand harvesting (performance marketing) and demand creation (brand) as separate, legitimate categories. Your move is to locate AI Share of Voice correctly: it’s a demand-creation / leading-indicator metric, and it belongs in the same mental bucket as brand investment — judged on whether it grows future market position, not on this quarter’s ROAS. Pitch it against ROAS and you’ll lose. Pitch it as the AI-era surface of demand creation, paired with your ROAS-measured harvesting, and it slots into a category the CFO already believes in.

And there’s a forty-year-old evidence base that makes “demand creation drives market share” a fact rather than a hope — which is the subject of the next section, and the single most powerful card in the deck.

6. The Forty-Year Law: Excess Share of Voice → Market Share

If you put one piece of evidence in front of a skeptical operator-founder, make it this one — because it predates AI entirely, which is exactly why it’s credible.

Binet & Field’s analysis of the IPA databank established that 10 points of excess share of voice (ESOV) above your current market share predicts roughly 0.5–0.7% of annual market-share growth, varying by category. This isn’t a single study — it’s cross-corroborated: Nielsen found “roughly the same number” across 123 brands; the B2B Institute measured ~0.6% per 10 points of ESOV; and brands with negative ESOV (spending less voice than their market share) lost share in roughly 80% of cases (all via Orr Consulting).

Here’s why this wins the room. The operator-founder’s deepest objection to any AI-marketing pitch is “this is hype, you’re selling me the shiny new thing.” The ESOV law lets you agree with their skepticism about hype and close the sale: “You’re right to distrust the AI hype. So don’t trust it. Trust the forty-year-old marketing law that share of voice predicts market share — the one Nielsen and the IPA databank confirmed across hundreds of brands. AI Share of Voice is simply that same proven law, on the surface where half your buyers now start their research. I’m not asking you to believe a fad. I’m asking you to apply a principle you already trust to the new front door.”

That reframe does something no amount of conversion data can do: it moves AI Share of Voice from the “unproven new thing” mental category into the “established law, new application” category. The number is novel; the law is old. Fund the law.

7. The Reporting Discipline That Makes the Budget Survive Next Quarter

Winning the budget once is half the job. The other half is making sure the program survives the next board review — because a budget granted on hype gets cut the first quarter the number doesn’t obviously move. The discipline that protects it, from Security Boulevard:

The reporting framework that survives actual board reviews connects citation share to three things boards already care about: pipeline, competitive positioning, and total addressable audience.

Connect the citation number to those three, every quarter:

1. Pipeline. Tie AI Share of Voice to pipeline as a leading indicator — “if we’re not in the AI shortlist today, we’re not in the pipeline next quarter.” Pair the SoV trend with AI-referred traffic and its conversion rate (the 4.4–5x multiplier) and, ideally, pipeline sourced or influenced by AI discovery. SoV says whether you’re in the answer; the conversion data shows the revenue that produces.
2. Competitive positioning. Report as a gap to a named competitor, per platform: “we’re recommended 38% in ChatGPT vs the leader’s 55%.” A gap implies a plan; an absolute number implies nothing.
3. Total addressable audience. Frame citation share as your share of the buyers who now start in AI — a growing slice of your TAM.

A budget ask framed as “give us money to improve our Share of Voice” dies, because it connects to nothing the board funds. A budget ask framed as “we’re recommended 38% vs the leader’s 55%, AI buyers convert at ~5x, and here’s the pipeline math if we close that gap” survives, because every clause connects to something the board already cares about. (The literal four-row board slide that operationalizes this is the subject of the scorecard post; the budget-relevant point is: report against pipeline / competition / TAM or watch the budget evaporate at the next review.)

And the one rule that governs all of it — because getting this wrong undoes everything above: when you report Share of Voice, never report a single blended number. The same brand sits at 40% in ChatGPT and 12% in Perplexity for identical queries (Orr Consulting). A blended figure is the new vanity metric — it hides the only fact that implies an action. Split by platform, every time. A CFO who catches you presenting a tidy aggregate that turns out to be meaningless will defund the program and doubt the next ask. The platform split isn’t pedantry; it’s the credibility that keeps the budget alive. (The full method for producing that per-platform number lives in the 50-Prompt Trap post.)

8. The Honest Caveats (Say Them First, or the Operator-Founder Won’t Trust You)

The B2B operator-founder is burn-scarred. They’ve been sold marketing magic before and watched it underdeliver. The fastest way to lose them is to over-claim; the fastest way to win them is to voice the caveats before they have to — concede the true part, and the rest of your case gains credibility by association. Here are the limits, stated plainly:

1. The volume is still small today. AI search is roughly 1% of total website traffic across most B2B sites (multiple sources). The bet is on trajectory, not current volume — which is precisely why AI Share of Voice is a leading indicator. Say this. An operator who hears you concede “the volume is tiny right now” trusts the rest of your numbers more, not less.
2. It’s a leading indicator, not revenue on its own. Share of Voice is “a very effective mirror for your gestalt digital presence” that “often closely resembles your market share” (Orr Consulting) — but it must be paired with pipeline and conversion data to mean anything financially. Don’t pitch it as a revenue metric; pitch it as the metric that predicts the revenue metric.
3. You can’t sprint it. Roughly 250 substantial documents are needed to meaningfully move how an LLM perceives a brand, and earned media drives the majority of citations. It compounds slowly, like brand equity. The corollary for the budget: this is a recurring investment, not a one-time project — which is honest, and which happens to align with how the work actually gets done.
4. A single aggregate number lies. Covered in Section 7, but it belongs on the caveat list too: report per platform or the number misleads.

Notice that none of these caveats weakens the budget case — they strengthen it. The small volume makes it a leading indicator (fund it before the volume arrives). The slow compounding makes it recurring (the retainer is the right structure). The leading-indicator nature is exactly why it predicts pipeline. An operator-founder who watches you disclose the limits up front concludes you’re the rare vendor selling the real thing — and that trust is worth more than any single statistic.

9. The 5 Counter-Intuitive Findings

1. The reallocation is expansion, not subtraction. Counter to the “SEO is dead, move the money” narrative, the smart money tops up — PMG pilots GEO at 1.5–2x the existing search budget; Pawco kept SEO and added AEO. Treat GEO as a parallel investment, not a replacement line. (Digiday)
2. The real cost is measuring the wrong thing. The expense in the room isn’t GEO spend — it’s the analyst-hours spent measuring a 2.8%-converting channel while the 14.2% channel goes untracked. (Opollo; akii.com)
3. Conceding the limits wins the skeptic. The operator-founder trusts the vendor who says “AI is 1% of traffic today” more, not less — because it proves you’re not selling hype. The honest caveat is a closing tool. (Orr Consulting)
4. Share of Voice isn’t a ROAS metric — and pitching it as one loses. It’s demand creation, not demand harvesting. Locate it in the brand-investment bucket the CFO already funds, paired with ROAS, and it slots into a believed category. (Orr Consulting)
5. The newest metric rests on the oldest law. AI Share of Voice is novel; the ESOV→market-share relationship behind it is forty years old and Nielsen-confirmed across hundreds of brands. Fund the law, not the fad. (Binet & Field / IPA)

10. FAQ

Should I cut my SEO budget to fund AEO/GEO?

No — and pitching it that way will likely lose you the budget. The dominant pattern among teams winning AI visibility is expansion, not subtraction: PMG pilots GEO at 1.5–2x the current search budget, and one widely-cited figure is that a $10k/month SEO budget should expect at least half that amount to also cover GEO — on top, not instead. SEO and GEO do different, complementary jobs and feed each other. Frame the ask as funding a parallel higher-converting channel, not abandoning a sunk investment.

What’s the single strongest number to put in front of a CFO?

The conversion gap: AI-referred traffic converts at 14.2% versus 2.8% for Google organic — a 4–5x gap (Opollo, 312 firms), corroborated by Semrush’s finding that the average AI visit is 4.4x as valuable. This reframes GEO from a cost line into a revenue argument: you’re under-funding the channel that converts at five times the rate of the one you already pay for. The premium concentrates in high-consideration B2B — exactly the profile most likely to be reading this.

How much budget should I actually ask for?

The benchmarks point to piloting GEO at 1.5–2x your current search budget (PMG), or adding GEO scope equal to at least half your existing SEO spend (Noise Media Group). For context, 12% of enterprise digital budgets already went to AEO/GEO in 2025, and 94% of CMOs plan to increase that in 2026. Start with a pilot sized to your category’s complexity, framed explicitly as expansion, and tie the number to a pipeline outcome rather than to “improving Share of Voice.”

How do I answer “isn’t AI search too small to matter?”

Concede it, then reframe. AI search is roughly 1% of total website traffic today — so the bet is on trajectory, not current volume, which is exactly why it’s a leading indicator you fund before the volume arrives. Pair that with the conversion gap (14.2% vs 2.8%) and the buyer-behavior shift (half of B2B buyers now start research in an AI chatbot). The honest “it’s small now” actually strengthens the case: you’re funding presence in the channel before it’s expensive to enter, not after.

What’s the difference between Share of Voice and ROAS, and why does it matter for budget?

ROAS measures how efficiently you harvest demand that already exists; Share of Voice measures how much demand exists for you in the first place (Orr Consulting). They’re sequential, not competing — demand creation refills the pool that demand harvesting drains. This matters because pitching Share of Voice as a ROAS-style performance metric loses (it’s a weak short-term ROAS proxy). Locate it instead in the demand-creation / brand-investment category your CFO already funds, paired with your ROAS-measured channels.

How do I keep the budget from getting cut next quarter?

Connect citation share to the three things boards already fund: pipeline (as a leading indicator, paired with AI-referred conversion data), competitive positioning (a gap to a named competitor, per platform), and total addressable audience. Report the trend, not the absolute level, and set expectations with the DORA J-Curve — the number dips before it lifts. And never report a single blended Share-of-Voice figure; split it by platform, because a tidy aggregate that turns out to be meaningless will cost you both the program and the next ask.

Is this just rebranded “brand marketing” that I already do?

Partly — and that’s the strongest part of the case. The law underneath AI Share of Voice is Binet & Field’s finding that excess share of voice predicts market-share growth (~0.5–0.7% per 10 points), confirmed by Nielsen across 123 brands. That’s a forty-year-old, peer-respected marketing principle. AI Share of Voice is that same proven law applied to the surface where buyers now start their research. You’re not funding a fad; you’re applying a principle the CFO already trusts to a new front door — which is precisely why it survives finance scrutiny.