AI Search Budgets: PR and SEO Merge, the CMO Roadmap

Why is the line between PR and SEO disappearing?

I look at sites every week. And for several months now, I’ve seen the same pattern come back.

On one side, a PR team landing great press mentions, interviews, citations in industry articles. On the other, an SEO team optimizing pages, tags, internal linking. Two teams. Two budgets. Two dashboards. And almost no conversation between them.

That used to work. It no longer works as well.

According to Neil Patel, PR and SEO have long evolved in separate corners of the marketing organization. AI Search is starting to change that relationship. As AI systems become the way people discover companies, products and services, the sources those systems rely on gain importance. And among those sources, third-party coverage is gaining ground.

Read that sentence again, slowly.

Third-party coverage is the historical territory of PR. The sources engines consult are the historical territory of SEO. The two territories now overlap.

Your teams are working on the same surface without knowing it.

In practice, when an AI engine answers a buyer’s question, it doesn’t just read your site. It cross-checks what you say about yourself against what others say about you. An article in a recognized media outlet. A comparison by a third party. A quote from an executive. Each of these pieces becomes an element of the answer.

In my view, this is the most underestimated structural change of 2026 on the marketing leadership side. Not because it’s spectacular. Because it touches the budget line, therefore the org chart, therefore habits.

The question I ask every CMO on the first call: who, in your company, is responsible for what AI engines say about your brand? The silence that follows is an answer.

What do AI engines actually read when they talk about your brand?

A classic search engine returns a list of pages. An AI engine writes an answer. That difference changes everything.

When it writes, it has to choose what to assert. And to assert with confidence, it looks for agreement. If your site says you’re a specialist on a topic, that’s a claim. If three independent publications say it too, that’s a fact. The engine needs facts.

That’s the mechanism. It’s not an opinion, it’s what I observe when I test AI assistant answers on my clients’ brands: the brands that appear with precise descriptions are the ones whose description shows up, almost word for word, in several places.

I’m not giving you a percentage here. I don’t have a reliable one, and nobody really does at this point. What I have is a repeated observation across semantic cocoon deployments: consistency across sources weighs more than the volume of sources.

Look at the difference between the old and the new reflex.

The table looks simple. The real question: who owns the narrative?

Take a common case. A software vendor describes itself on its site as a management platform for SMEs. In the press, it’s presented as an accounting tool. In directories, it’s listed as a CRM. Three descriptions, three teams that wrote independently.

The engine has to summarize. It hesitates. It generalizes. Sometimes it gets it wrong.

The real problem: consistency.

According to Neil Patel, the sources these systems depend on are becoming more important. My field observation: what really matters is what they say, and how they say it.

How do you reallocate budgets without breaking what works?

When a CMO hears « PR and SEO merger, » they often hear « budget cuts. » That’s not the point.

The point is to know where each euro works toward two goals at once. A budget that funds a press hit without thinking about its readability by engines wastes half of its effect. A budget that funds SEO content without thinking about its pickup by third parties wastes the other half.

Here’s how I approach reallocation with my clients. Three pockets. No imposed percentage: each mix depends on your market, your maturity and what your teams already know how to do.

• The foundation pocket. Your site, your architecture, your reference pages. This is the primary source. If it’s blurry, everything that comes from outside stays blurry too.
• The proof pocket. Mentions, citations, studies, speaking appearances and content published by third parties. This is the former territory of PR, now steerable with an SEO reading grid.
• The consistency pocket. The work of reviewing, aligning descriptions, and correcting gaps between sources. It’s rarely budgeted. Yet it makes the difference.

The third pocket always surprises. Nobody owns it in the org chart. It falls between teams. So it doesn’t exist.

My advice, drawn from what I observe with my clients: before moving a single line, take inventory of what each expense actually produces. Was a press hit reviewed for its brand description? Is a product page cited by a third party? Does a content partnership point to the right page?

These questions shift the trade-offs without increasing the envelope.

The best AI Search budget isn’t an additional budget. It’s a better-coordinated budget.

There’s also a question of timing. PR work produces its effects over months. So does site structure work. If you wait for AI answers to degrade before acting, you react a step behind. If you anticipate at the next allocation, you start a quarter ahead.

Should you merge the PR and SEO teams?

No. Not necessarily.

I defend a counter-intuitive approach. You often hear that convergence requires a big reorganization, a new role, a new department. I’ve seen that scénario. It consumes six months of internal discussions and leaves the same problems in place.

What works is smaller and more concrete: a shared map.

A map is a single document that both teams consult and contribute to. It contains four elements:

• The reference description of your brand, in one sentence and in one paragraph.
• The five to ten topics on which you want to be recognized.
• The pages on your site that are authoritative on each topic.
• The third-party sources targeted or already secured for each topic.

Nothing sophisticated. It’s what I build in a semantic cocoon, extended beyond the site. I’ve delivered 1 300+ cocoons since 2016. The logic stays the same: every page has a role, every link has a reason. You just apply it to external mentions.

With this map, the PR team knows which angle to serve, the SEO team knows which page should get the attention, and you know who does what.

According to Neil Patel, it’s the relationship between the two disciplines that is changing. Not their existence. I share that reading. The professions stay distinct. The objectives come together.

Three rituals are enough to maintain convergence:

1. A joint monthly meeting of forty-five minutes, with the map on screen.
2. A PR brief reviewed by SEO before every major launch.
3. A shared report that brings together mentions secured and pages strengthened.

It looks small. It’s what builds the habit.

And if your organization is highly siloed, start with a single topic. One product. One range. One market. Prove that the method produces more accurate answers. Then expand.

How do you measure the effect when the tools are still imperfect?

Let’s be honest. Measuring visibility in AI answers is not mature yet.

The tools exist, some are useful, none gives a complete picture. Answers vary by wording, timing, engine. If someone promises you a perfect dashboard, be skeptical.

That doesn’t mean we move forward blind. I work in four steps.

Step 1: a question panel. Write ten to twenty questions your buyers actually ask. Not keywords. Full sentences, as you would ask an assistant.

Step 2: a regular snapshot. Ask those questions to the main AI assistants, every two weeks for example. Note three things: is your brand cited, how is it described, which sources appear.

Step 3: a reading of the sources. This is the most valuable part. The cited sources tell you exactly where the engine goes to look. They are your shared PR and SEO targets.

Step 4: a gap log. Each time an AI’s description differs from your reference description, note the gap and trace its source.

This setup costs almost nothing in tooling. It requires discipline.

I own it: it’s a barometer, not a scientific measurement. It tells you whether the direction is right, not how many points you’ve gained. And it keeps you out of the opposite trap, the one where you measure nothing because nothing is perfect.

An imperfect barometer checked every two weeks is worth more than a perfect dashboard nobody opens.

One more point. Keep two things clearly separate: visibility in AI answers and commercial performance. The first is an early signal. The second is the result. Don’t ask your team to prove a direct link too quickly. First ask them to prove that descriptions are becoming more accurate and that cited sources are becoming more relevant.

The rest follows. Not always at the pace you hope for, but in the right direction.

On the sources of these measurements, let me be clear: my observations come from snapshots taken on my clients’ brands. They illustrate trends. This is not a quantified study – I don’t present it as one.

What roadmap should you follow before the next budget allocation?

You have the logic. Let’s move to action.

Here is the roadmap I propose to CMOs, in four stages. Each stage fits in one meeting and one deliverable.

Stage 1: the baseline. Query five AI assistants about your brand, your category and your competitors. Copy the answers. List the cited sources. In half a day, you get a snapshot your team has probably never seen.

Stage 2: the shared map. Bring PR and SEO together. Build the reference description, the priority topics and the authority pages together. One document. One designated owner.

Stage 3: the trade-off. Go back to your budget lines. For each one, ask: which topic on the map does this expense serve? Expenses with no answer become candidates for reallocation. Not for elimination. For reallocation.

Stage 4: the pilot. Choose one topic. Mobilize both teams on it for a quarter. Capture AI answers before, during and after. Present the result to your executive committee with the real excerpts.

That last point matters. An excerpt of an AI answer that finally describes your offer correctly convinces faster than a twenty-slide presentation.

The starting point of this reflection, as Neil Patel reminds us, remains simple: the sources AI systems use matter more and more, and third-party coverage is part of them. The rest is organization and discipline.

I’m not selling you the method. I’m showing you the pages. That’s why my first exchanges with a client take the form of a live audit: we open the AI answers, read the sources, and spot together the gaps between what you say and what’s said about you.

So, for your brand: if I asked an AI assistant ten buyer questions today, who in your organization could tell me where the answers come from?

Stéphane Jambu

SEO & AI Engineer

I build growth systems / AI / Neuroscience | 650+ clients · 80 LinkedIn testimonials · 30 years of expertise · 15 years of systems running without me.

Follow on LinkedIn