5 bets I’m making on 2027 in SEO and AI search
Introduction
Five years out, nobody knows. That is why I don’t make five-year forecasts — I make bets on an 18-month horizon and revisit them every six months.
It is a different genre. A forecast says «this is how it will be». A bet says «this is where I put my money, my time and my reputation, and here is why». A forecast demands being right. A bet only demands stating honestly where I might be wrong.
Here are the five bets I am making on 2027. Each one comes with the reasoning, the numbers, and an honest read of how confident I actually am.
In short
Five bets for 2026–2027: (1) AI Overviews will appear in more than 50% of search results by the end of 2026, (2) ChatGPT and Perplexity will take a significant share of B2B search away from Google, (3) the SEO market will split into two categories — strategic and commodity, (4) Ukrainian business in Poland will build its own infrastructure of digital services, (5) classic SEO won’t disappear, but the volume of manual work will roughly halve. Four of the five are observations in motion; one is an assumption. All of them can turn out wrong.
Bet 1 — AI Overviews will cover more than 50% of search results by the end of 2026
Confidence level: high, around 80%.
The reasoning. In January 2025, AI Overviews showed up in 6.49% of queries (Semrush). By March 2025 — 13.14%. In Q1 2026 — 25.11% (Conductor, based on an analysis of 21.9 million searches). The doubling period is roughly six months.
If that line holds, we end 2026 somewhere between 45% and 55%. This is not extrapolation to infinity — at some point growth slows down, because not every query suits an AI answer (transactional and navigational ones don’t). But I see enough room for one more doubling.
What it means in practice. Half of all search queries resolved without a classic click to a website, or with a click only to the cited sources. «Be in the Google top three» stops covering half the demand. Alongside SEO you need a search strategy that accounts for AI answers, not just for blue links.
Where I could be wrong. If Google runs into a legal or regulatory limit in the EU — and there have already been signals of that around the AI Act — the pace may slow. Or Google itself trims AI Overviews in response to pressure from publishers.
Bet 2 — ChatGPT and Perplexity will take a significant share of B2B search
Confidence level: medium, around 60%.
The reasoning. ChatGPT reported 900 million weekly active users in February 2026 (OpenAI). Perplexity — 22 million active users and 780 million search queries a month (May 2025). The number that matters most: visitors arriving from AI search convert at roughly 23 times the rate of ordinary organic traffic (Ahrefs, June 2025).
In B2B, search behaves differently than in retail. A CMO or a CTO does not type «best SEO tool» into Google. They write to ChatGPT: «I run marketing at a 50-person B2B SaaS company, what would you recommend for SEO». That is not a query — that is a brief. Google doesn’t answer briefs. ChatGPT does.
What it means in practice. If your ICP is a B2B decision-maker, your content has to be citable by AI systems, not only rankable in Google. That is a different kind of optimisation: original data, frameworks, named expert authorship.
Where I could be wrong. If Google moves fast and makes AI Mode the default interface — and the signals point that way — part of this shift returns to Google. My direction would still be right, but my call on «who wins» would be wrong.
Bet 3 — the SEO market will split into two categories
Confidence level: medium, around 65%.
What I have observed over the past 18 months. Client requests come in two shapes. The first: «do our SEO, cheaper». The second: «make us look like an authority — to people and to AI systems alike». These are different jobs that used to sit in the same basket labelled «SEO».
My bet. By the end of 2027 the market splits:
- The commodity segment — fast basic services, largely automated, with prices falling to 200–500 EUR a month. Only those who learn to sell at volume will survive there.
- The strategic segment — AI visibility, content strategy built on original data, brand authority. Prices of 5 000–15 000 EUR a month and up. Fewer clients, deeper work.
What it means in practice. If you run an agency, you are already standing on one of those two tracks, even if you haven’t named it yet. Being «a mid-sized agency that does everything» becomes economically unviable.
Where I could be wrong. A third segment may emerge — vertical-specific shops serving only clinics, or only law firms. I see early signs of it, but I am not convinced it becomes a mass trend.
Bet 4 — Ukrainian business in Poland will build its own digital services infrastructure
Confidence level: high, around 75% — and I am personally invested in this one.
The reasoning. Poland now hosts 29 044 Ukrainian-owned Sp. z o.o. companies and more than 90 000 sole proprietors (Opendatabot × KRS, July 2025), growing at about 30% a year since early 2022. It is the largest foreign entrepreneurial community in the country.
Polish agencies mostly don’t speak Ukrainian. Ukrainian agencies don’t have a Polish location or Polish network. A gap opens up in between.
My bet. By 2027 there will be 10 to 20 «Ukrainian-Polish» agencies and fractional specialists closing that gap. Be Visible is one of them. This is not one agency taking the market — it is several, each with its own specialisation.
What it means in practice. Ukrainian business in Poland gradually moves away from mid-sized Polish agencies, where the cultural connection is missing, and from remote Ukrainian ones, where the local understanding is missing. The segment forms between those two poles.
Where I could be wrong. If geopolitical circumstances push a large share of Ukrainian businesses to move back to Ukraine, the market contracts. That is a scenario I cannot forecast.
Bet 5 — classic SEO won’t disappear, but its volume will halve
Confidence level: medium, around 60%.
The reasoning. Classic SEO is a lot of manual work: technical audits, keyword research, on-page optimisation, link building. Between 2024 and 2026 a large part of that got automated — AI-generated meta descriptions, automated technical audits, programmatic SEO for large sites.
My bet. By the end of 2027 the amount of manual work in classic SEO drops by roughly half. It doesn’t vanish, because strategy, interpretation of results and coordination stay human. But «doing SEO by hand» for 40 hours a week is a dying profession.
What replaces it. Hybrid roles: SEO strategist plus AI-visibility specialist, SEO analyst plus data engineer, SEO copywriter plus prompt engineer. Each new role demands more expertise and less manual digging.
Where I could be wrong. It may go the other way — the workload grows because SEO itself gets more complex: more channels, more platforms, AI visibility as an extra layer. Then the volume doesn’t shrink, it redistributes.
What I don’t know — and why that’s fine
Honestly: there are things I cannot predict at all.
- What happens to AI regulation, especially in the EU. The AI Act can change the rules of the game.
- Whether an alternative to Google appears — not ChatGPT, not Perplexity, but something radically different. Technological history suggests such jumps happen once every 10–15 years.
- How user behaviour shifts — whether people keep typing queries, move to voice, or move to multimodal interfaces.
- What happens to the cost of AI compute — that determines how long Google can keep AI Overviews free.
Not knowing is fine. The honest part of my job is to describe the edge of my own map, rather than claim the whole territory. If someone in this field offers you 100% certainty, that is your cue to ask which data that certainty is built on.
How I’ll check these bets
Every six months: December 2026, June 2027. Each bet gets a specific metric that tells me whether the direction was right.
- AI Overviews above 50% → I read the Conductor and Semrush reports on AI Overviews presence.
- ChatGPT and Perplexity in B2B → I measure brand share of voice inside AI systems for my B2B clients through Profound.
- The market split → I watch how prices move across the market, using agencies’ public rate cards, and how many new premium agencies appear.
- Ukrainian business in Poland → I track Opendatabot data on Sp. z o.o. registrations, and the number of inquiries my agency receives from that segment.
- Classic SEO shrinking → I ask freelance colleagues about their workload and the nature of the tasks they get.
If any bet turns out clearly wrong, I will write about it publicly. Without excuses.
FAQ
Five questions I get most often about this format — what a bet is, why the horizon is 18 months, and what to do when a bet your strategy depends on turns out wrong.
What’s next
I will come back to this article in December 2026 and publish an update. For each bet: worked / didn’t work / still in motion. That is part of the method, not a footnote to it.
If you are making your own bets for 2027, share them — I’d like to compare. Not to debate, but to see other corners of the map.
Making bets in public is a discipline tool for thinking. It forces you to be specific where it would be more comfortable to stay general. It is not a rule, it is my own practice. It works, so I keep doing it.
If you want to talk through what these shifts mean for your own site, write to us — describe the business and send the link.
— Iryna Lehusha, Be Visible
External sources
Three resources behind the numbers these bets rest on. If you want to check my logic or build forecasts of your own, start here.
- Conductor — Research & SEO insights
Conductor’s research hub is the source of the analysis of 21.9 million searches in Q1 2026, which Bet 1 is built on — AI Overviews passing 50% by the end of 2026. It also publishes regular updates on SERP changes. - Statista — Digital markets data
A cross-source aggregator. The user numbers for ChatGPT (900 million weekly actives), Perplexity (22 million) and Google support Bet 2 on the redistribution of B2B search. Useful if you want to see the trend over time. - Gartner — Search engine volume will drop 25% by 2026
Gartner’s forecast of a 25% shift of search volume towards AI chatbots and virtual agents by the end of 2026 is the industry context for Bet 3 and Bet 5.
