Here's something that should make every marketing team slightly uncomfortable: the model family deciding whether your organic content deserves to show up in AI Overviews is the same model family shaping how your paid ads get written, targeted, and placed.

Gemini now sits across both sides of what the industry still treats like a hard border — powering the creative generation, query understanding, and intent-matching in paid, while simultaneously running the organic AI experiences those ads appear next to. Two surfaces, one family of models forming its understanding of your brand and applying that understanding everywhere.

And most organisations are still running paid and organic out of separate rooms, with separate budgets, separate reports, and — let's be honest — separate Slack channels where each team politely questions the other's existence.

(If this describes your org chart, don't worry. It describes almost everyone's.)

Let's trace what's actually happening inside the machine, because this is where the comfortable fiction falls apart.

On the paid side, Google's newer campaign types have quietly stopped being campaigns in any traditional sense. Performance Max runs across search, YouTube, Gmail, Maps, Display, and Shopping simultaneously through a single automated system, deciding placement by placement where to show your ad and to whom. AI Max does the same thing specifically for search, reading your pages to decide which ads to write and which queries to target. In both cases, the advertiser isn't building the machine anymore. The advertiser is feeding it.

Now look at the organic side. AI Overviews. AI Mode. The conversational search experience that's eating traditional results from the top down. These are also Gemini, reading the same query, from the same user, in the same moment, forming the same judgement about what that person needs.

Put those two facts next to each other and the implication is quietly devastating. The same model family that shapes your paid creative and targeting is the one evaluating your organic content. Gemini is forming its understanding of your brand and then expressing that understanding on both sides of a line that your org chart still treats as load-bearing.

The line is not load-bearing. It's decorative. It's a feature wall in a building that's been structurally redesigned around it.

Here's where things get personally uncomfortable for anyone who built their career on the beautiful complexity of paid search management.

The old paid game was intricate. Keywords, match types, bid adjustments, audience segments, the elaborate campaign structures that separated a skilled account manager from someone clicking buttons. Complexity was the craft. The craft was the moat. You could spend years mastering it and feel justifiably proud of the expertise.

On Gemini-driven surfaces, most of those levers are gone. Not deprecated. Not simplified. Gone, in the way that a switchboard operator's skillset is gone — the infrastructure moved on, and the expertise that used to run on top of it went with it.

The model decides where an ad appears, what it says, who sees it, when, and at what price. The advertiser sets parameters and supplies information, but the decisions that actually determine outcomes are made by the system. You can't instruct Gemini to show your ad to a particular person in a particular way. You can only shape what Gemini understands about you before it decides.

That's not a tweak. That's the optimisation surface moving from settings to substance. What the model reads about your brand — how clearly it can establish who you are, what you offer, who you serve, and whether any of that is consistent and trustworthy — now matters more than any configuration inside the platform.

The industry calls this signal "the entity": the model's structured understanding of a brand as a known thing in the world, rather than a collection of keywords wearing a trench coat pretending to be a strategy.

(I realise "entity" sounds like something from a horror film. In fairness, watching your elaborate campaign architecture become irrelevant does have a certain body-horror quality.)

This is the part worth thinking about carefully, because it's where the convergence stops being theoretical and starts touching the budget.

A paid ad served inside an AI-driven experience has to be consistent with the organic answer the user is already looking at. If the sponsored result contradicts the surrounding context — different claims, a different picture of the brand, a mismatch in tone or fact — the experience breaks. The click dies. For the ad to stay consistent with the organic context, it has to be grounded in the same underlying material: the same search index, the same knowledge graph, the same model reading the same signals.

If that holds — and I want to flag this as a reasoned argument rather than documented gospel, because intellectual honesty matters more than sounding certain — then a specific and slightly counterintuitive consequence follows. The work you do to make the model trust you organically isn't separate from your paid performance. It's an input to your paid performance.

Teaching Gemini that you are who you say you are, that your facts are consistent, that you're the credible answer for a particular need — that lowers the cost of being treated as credible when you pay.

One training investment. Two outputs. The organic discipline stops being the budget line everyone cuts first and becomes the thing that makes paid cheaper.

(I know. The organic team just felt a vindication so intense it may require medical attention. Give them a moment.)

Follow the coupling into the money and it stops being abstract.

When the model's confidence in a brand is low, the cost shows up everywhere the model touches. Organically, it hedges. It softens claims it's unsure about, prefers a competitor it understands better in a direct comparison, and in the worst case simply doesn't surface you at all. None of that surprises anyone who's watched their visibility quietly erode inside AI answers.

What's new is that the same low confidence crosses into paid. If the model is unsure about you, it rewrites your ad creative to soften the same claims it hedges organically. It passes your ad over for a competitor it trusts more, even when your bid is competitive. And if it won't surface you organically, it won't run your ad either.

There is no bid high enough to buy your way into a consideration set the model hasn't put you in.

Read that again if you need to, because it applies whether you're a SaaS company bidding on competitor terms, a law firm buying "personal injury lawyer near me," or a local plumber trying to outspend the guy with 400 five-star reviews. The dynamic is universal.

That sentence deserves to sit on its own for a second, because it quietly invalidates a strategy that's kept a lot of brands afloat for years: using paid to compensate for a weak organic position. On these surfaces, that move is dying. You cannot outbid a model that doesn't trust you. You can only become a brand it trusts, at which point the bid matters less than it ever has.

The mirror image is the part worth dwelling on. When confidence is high, a well-defined placement does two jobs at once. It buys an efficient result today and teaches the model who the brand serves tomorrow. Trust earned on one side lowers cost on the other. The effect compounds. It's the closest thing digital marketing has to a flywheel that actually deserves the word.

None of this means abandoning paid. It means stopping running it as a sealed compartment.

The practical shape is a single loop rather than two parallel tracks. Paid is still the fastest way to discover which combinations actually convert — which audience, which intent, which margin band produce a profitable result — because in paid you declare all three of those things explicitly. Organic can't declare them directly. The model has to infer who a page serves from behaviour: who landed, who stayed, who converted, who never came back to search for the same thing again.

So use each side for what it's actually good at. Let paid tell you precisely which cohort-and-intent combinations are worth owning. Then build the organic pages around exactly those combinations, designed so specifically for the intended person that the way they behave on the page sends the model the same signal the paid campaign declared explicitly — this time for free.

A page built for a person rather than a query converts better because it genuinely serves that person. Better behaviour sends a cleaner signal. Cleaner signal strengthens organic visibility for that cohort. And stronger organic signal — on the grounding argument above — reduces what you pay in paid.

The same piece of work pays out in three places. The loop feeds itself.

This isn't a theoretical nicety. It's the difference between a brand that spends increasingly more to stand still and a brand that spends the same while reach compounds. The maths only works if you stop treating the two channels as separate cost centres with separate KPIs that happen to share a Google login.

The two teams who've been politely ignoring each other across the office are, whether they like it or not, building the same thing.

One model family now shapes both sides of what most brands still organise as two entirely separate functions. The two budgets are already being spent into the same system, evaluated against the same understanding of your brand, grounded — probably — in the same underlying index. The dashboards are separate. The teams are separate. The quarterly reports are separate. The machine reading all of it observes none of those boundaries.

The brands that figure this out first won't just save money. They'll compound trust in a system that rewards consistency and punishes fragmentation. The brands that keep running paid and organic as parallel tracks will keep wondering why their paid costs are climbing while their organic visibility shrinks — never connecting the two, because the org chart told them not to.

The real question isn't whether the two budgets eventually converge. It's what you're actually buying when you spend on paid at all — because the one thing you can no longer purchase is a way past a model that hasn't decided to trust you yet.

(And if that thought keeps you up at night, welcome to the club. We meet at 2 AM. Bring snacks.)

Jo Lambadjieva is an entrepreneur and AI expert in the e-commerce industry. She is the founder and CEO of Amazing Wave, an agency specializing in AI-driven solutions for e-commerce businesses. With over 13 years of experience in digital marketing, agency work, and e-commerce, Joanna has established herself as a thought leader in integrating AI technologies for business growth.

😍It was great, all I needed

🙂Good, not great

🙄It sucked