You don't have to choose between AI, SEO and paid advertising: you have to figure out how much weight to give each one right now — and rebalance every quarter as the market shifts. Those who cut a channel to chase the new thing usually end up chasing two problems instead of one. There's a way to keep everything in balance without wasting budget, and the difference comes down to having a precise rule for where to move the money and when.
It’s not AI vs SEO vs Ads. It’s AI and SEO and Ads, with budget that shifts in real time. Anyone who doesn’t think in terms of mix loses efficiency on the whole 100%.
I’ll say it right away because in my articles I often see this fixation: “Roberto, do I need to pull budget out of Google Ads to move it to AI visibility?”. Blunt answer: it depends on your conversion data, and it’s almost never a total shift. The right question isn’t which channel to kill, but how much budget to rebalance each quarter based on the cost per lead of each one.
In this series I’ve already talked to you about how to measure incoming AI traffic and how to calculate the cost per AI mention. Now let’s close the loop: how that data feeds into the overall budget, alongside SEO and Ads.
The reframe that changes the question
For years in Italy, anyone managing marketing has thought in terms of siloed channels: so much to Google Ads, so much to Meta, so much to SEO. AI visibility arrives and the first reaction is identical: “how much do I allocate to the AI channel?”.
Wrong framing. Visibility in AI answers is not a standalone channel that lives on its own. It’s the result of brand authority, content structure, distribution across sources cited by the models. Many of these signals are the same ones that feed SEO. And part of the Ads budget serves to build the brand recognition that AI models then use as a trust signal.
So it’s not a pie to be split into three independent slices. It’s an ecosystem where moving 10% in one direction also shifts signals in the others. That’s why calculating the mix can’t be done by “feeling”; it has to be anchored to per-channel conversion data.
The principle: every euro goes where it converts best
In the world of classic performance marketing the principle has been documented for decades: budget is allocated where the cost per acquisition is lowest, at the same lead volume and quality. It works for Google Ads vs Meta Ads, it works for SEO vs Ads, and today it works for AI visibility vs the other two.
From this it follows that, for a company that wants to show up in AI answers without throwing money away, the optimal mix is not a fixed percentage decided in advance. It’s the result of three numbers:
- Cost per lead from organic SEO (content + technical budget / organic leads)
- Cost per lead from Google Ads (Ads spend / leads from Ads)
- Cost per lead from AI referral (AI-friendly content budget + monitoring tools / leads from AI)
In practical terms: you measure, you compare, you shift 10-20% each quarter toward the channel that converts best, without zeroing out the others. Because if you zero out SEO, AI visibility collapses too (the models pull from well-ranked web sources), and if you zero out Ads you lose coverage on high-intent commercial queries.
Anyone who measures after 30 days and says “it’s not working” is judging too soon.
The Stintino case: how you actually rebalance
Let me tell you about a real, anonymized case. A boutique hotel in Stintino (province of Sassari, northwest coast of Sardinia), 18 rooms, mid-to-high-end target. Marketing mix at the start of 2025:
- Google Ads: 4,500 euro/month (60% of the budget)
- SEO: 2,000 euro/month (27% — agency + blog content)
- AI visibility: 1,000 euro/month (13% — content designed to be cited by Perplexity/ChatGPT on queries like “boutique hotel northwest Sardinia”)
In September 2025 we measured the cost per qualified booking (direct inquiry + booking, excluding OTAs):
- Google Ads: 78 euro per qualified booking
- Organic SEO: 41 euro
- AI referral (traffic from link.perplexity, chat.openai, trackable Google AI Overviews): 29 euro
The AI figure was surprising because it was on small volumes (12 bookings in the month versus 58 from Ads), yet the unit cost was clearly lower. Honest limitation of the test: three months of tracking, a sample of a single hotel, impossible to rule out seasonality. It’s not a study, it’s an operational observation.
What we did: shifted 800 euro/month from Google Ads (still leaving 3,700, because they cover high-intent last-minute queries) toward AI visibility (raised to 1,800 euro/month). SEO unchanged. The following quarter: leads from AI rose to 21/month, cost per AI booking dropped to 24 euro. Ads bookings fell from 58 to 49 but the unit cost stayed stable.
Quarter conclusion: overall mix efficiency improved by about 18%, calculated as the weighted average cost per qualified booking. Without having touched SEO, without having “killed” Ads.
In practical terms: you measure, you compare, you shift 10-20% each quarter toward the channel that converts best, without zeroing out the others.
The test you can run in 60 minutes
You don’t need an enterprise-consulting attribution model. You need a spreadsheet and discipline. Here are the three steps.
First: open Google Search Console and isolate the converting organic traffic of the last quarter. Divide the quarterly SEO cost (agency + content + tools) by the number of leads generated. You’ve got your cost per SEO lead.
Second: inside Google Ads or Meta Ads, take the cost per conversion you already know. If you don’t have the conversion pixel set up properly, start from the form CPL (cost per lead). A number that’s immediately available.
Third: inside Google Analytics 4 segment the referral traffic by filtering on source: `chat.openai.com`, `perplexity.ai`, `gemini.google.com`, `copilot.microsoft.com`. Count the leads generated and divide by the budget you’ve invested in content designed for AI answers (see how to track AI referral traffic). You’ve got your cost per AI lead.
Three numbers, a column next to them with the cost per lead. The one with the lowest number at the same useful volume deserves 10-20% more budget next quarter. The one with the highest number loses 10%. Never zero out anything. Never move more than 20% per quarter, because the data is noisy.
Honest entry-level check: this is a first step. Real attribution across channels with multiple touchpoints requires professional marketing mix modeling tools. But for an SME this three-column calculation does 70% of the work.
The mistakes I see most often
Four patterns I run into systematically when I help a company rebalance its mix.
Confusing volume and unit cost. “But Ads bring in 60 leads, AI only 15, AI sucks”. No: 15 leads at 25 euro each are worth more than 60 leads at 80 euro, if the quality is comparable. Quantity without unit cost tells you nothing.
Cutting the channel with the highest CPL all the way to zero. Terrible idea. If Google Ads has the highest CPL but covers immediate commercial queries, zeroing it out means losing that high-intent audience. You rebalance, you don’t amputate.
Not factoring in time. SEO and AI visibility don’t respond in 7 days the way Ads do. If you shift budget toward AI-friendly content today, you’ll see the impact in 60-90 days. Anyone who measures after 30 days and says “it’s not working” is judging too soon.
Treating AI visibility like Ads. They’re different things: AI visibility is built with brand authority, structured data, an author the models recognize, E-E-A-T. You don’t “buy” it the way you buy a click. The AI budget is content production + monitoring, not direct media spending.
What to do in the next 30 days
- Open a spreadsheet with three rows (SEO, Ads, AI) and three columns (monthly budget, leads generated, cost per lead)
- Fill the first two columns with solid data from the last 90 days
- Calculate the third column (simple division)
- Identify the channel with the lowest CPL and the one with the highest
- Next quarter: shift 10-20% from the highest to the lowest
- Redo the calculation at the end of the quarter. If the trend holds, shift another 10%. If it reverses, go back
- Never emotional moves: everything is driven by quarterly numerical evidence
Then compare with the 3-5 direct competitors in your sector: whoever gets cited by AI on your main queries is investing differently than you. It’s no accident.
Why this matters for AI visibility
Everything I’ve told you in this series about how to measure visibility in AI answers (share of voice, accuracy rate, maturity model) only makes sense if it then feeds into budget decisions. Measuring for the sake of measuring doesn’t move revenue. Measuring to rebalance the mix does.
The company that shows up in AI answers more than its competitors isn’t the one that “found a trick”. It’s the one that understood before the others that AI, SEO and Ads are a single system, and started shifting 10% of budget per quarter toward the channel that converts best. Three years of discipline on this principle make the difference between a brand that gets cited and one that’s invisible.
In the next articles in the series I’ll show you how AI visibility becomes a real competitive moat and how to connect AI visibility and lead attribution to close the loop between measurement and revenue.
