A 2026 prediction list I'd bet money on.
Five hills I'm willing to die on. Probably. Written in December 2025, checked against reality in June 2026, with notes on what held up and what I got wrong.
Published1 June 2026
ByThomas Cox
Read time7 minutes
Filed underOpinion · Strategy · GEO
Prediction posts are a hostage to fortune. Most practitioners write them vague enough to claim credit either way. I find that useless, so I'm doing the opposite: five specific, falsifiable positions I held going into 2026, with an honest scorecard attached now that we're six months in.
The point isn't to look clever. It's to show how I think about where search is going, and to hold myself accountable to that thinking publicly.
Prediction one: citation share becomes a boardroom metric.
What I said: By mid-2026, at least one major B2B SaaS category leader will report LLM citation share in an earnings call or investor deck alongside traditional organic traffic.
June 2026 verdict: directionally correct, not quite there. We have not yet seen citation share in a public earnings call. What we have seen is a wave of B2B marketing teams adding AI visibility metrics to their quarterly OKRs. I've seen this in three separate engagements this year. The language is internal, not public, but the behaviour is there. I give this one eighteen more months before it surfaces in investor materials.
Why does this matter to your strategy now? Because the teams already tracking it are building institutional knowledge about what moves the number. The teams waiting for a public signal will be a year behind when that signal arrives. Citation share as a metric is not hypothetical. It is measurable today with nothing more than a prompt set and a spreadsheet.
Prediction two: traditional keyword rankings stop being the primary SEO deliverable.
What I said: By Q3 2026, the majority of serious B2B SEO teams will have dropped or significantly demoted weekly keyword ranking reports from their core deliverables, replacing or supplementing them with AI visibility metrics.
June 2026 verdict: happening faster than I expected. In-house SEO teams I speak to (particularly at Series B and above) are already having this conversation internally. Agencies are slower, for obvious commercial reasons (ranking reports are easy to produce and easy to show clients). The shift is being driven from the in-house side, not the agency side, and it is accelerating. The question is no longer whether to replace keyword rankings as the primary signal, but what to replace them with. I wrote about the practical measurement alternatives in how I measure AI search ROI when there's no click to track.
Prediction three: entity work overtakes link building as the dominant off-site SEO activity.
What I said: Link building budget at the median B2B SaaS company will decline by 20–30% in absolute spend in 2026, with the reallocation going into entity and PR work.
June 2026 verdict: too early to call definitively, but the directional pressure is real. Link building budgets have not collapsed. What has happened is that the marginal link (the kind you buy from an outreach agency at £150 a placement) has visibly lost value as a GEO signal. The question clients are asking has shifted from "can we get more links?" to "which links actually move AI citations?" That's a different brief, and it naturally selects for editorial and reference-source placements over volume. The practical framework for building source-worthiness rather than just link volume is what I keep coming back to here.
Prediction four: Google AI Overviews reach 60%+ of B2B SaaS queries.
What I said: AI Overviews will appear on more than 60% of commercial B2B SaaS queries by end of Q2 2026.
June 2026 verdict: correct. The data from my own prompt tracking shows AI Overviews now appearing consistently across the DevTools, MarTech, and security categories I monitor. The holdout categories (highly regulated verticals like fintech and legal tech) are close behind. The implication is that if you're doing B2B SaaS SEO and your content strategy doesn't account for how AI Overviews select and attribute sources, you're optimising for a SERP that fewer and fewer of your buyers are actually seeing.
Prediction five: the first GEO-native content agencies appear.
What I said: By mid-2026 there will be agencies that position exclusively around AI search visibility: not traditional SEO with an AI bolt-on, but GEO-first practices with their own methodologies, pricing, and case studies.
June 2026 verdict: already happened. They exist. Some are credible: practitioners who genuinely understand entity coherence, RAG retrieval, and the mechanics of how models cite sources. Many are not. These are SEO agencies that swapped "link building" for "GEO optimisation" in their deck without changing what they actually do. The market for GEO-native capability is real; the ability to distinguish signal from noise requires knowing what the work actually looks like at the execution level. Which is most of what I write about here.
/ Scorecard
Two correct, one directionally correct, one too early, one already happened. I'll take that. The broader point, that 2026 is the year AI search stops being a "watch this space" topic and starts being a "fix this now" problem, I'd stand by without hesitation.
For 2027, I'll post a fresh list in December. The predictions I'm currently sitting with: structured authorship becomes a ranking signal Google explicitly acknowledges; Perplexity overtakes Bing in B2B research tool usage; and the first major B2B SaaS brand publicly attributes a lost deal to poor AI visibility. I'll be more specific when I write it up.