Check your analytics against this time last year. Rankings holding steady, maybe even improving. Organic traffic down anyway. Leads from search quietly thinner than they used to be, and nobody in the marketing meeting can quite explain why the machine that worked for a decade has stopped paying out.
Here is why. Your buyers are still searching. They are just not clicking, because Google now answers their questions for them, and increasingly so do ChatGPT, Perplexity and the rest. The blog-post-to-traffic pipeline that B2B marketing was built on is being dismantled in real time.
The short version of what to do about it: shift from chasing clicks to earning presence, and put video at the centre of that shift, because video is proving to be the format that AI-driven search disrupts least and rewards most.
What actually changed
The numbers describe a structural break rather than a dip. Analysis of search behaviour by Semrush and others puts the share of Google searches ending without a single click at around 60%, and higher still on mobile. Gartner predicted back in 2024 that traditional search engine volume would fall 25% by 2026 as AI assistants absorbed query share, a forecast that looked bold at the time and looks conservative now.
The mechanism is AI Overviews: the generated answers Google places above the results. When one appears, most people read it and leave. Pew Research Center found that only 1% of users clicked a source link inside an AI Overview. Seer Interactive’s tracking shows click-through rates on affected queries falling by more than half since mid-2024.
And B2B is in the blast radius more than almost any other sector. BrightEdge’s sixteen-month tracking study found that 82% of B2B technology queries now trigger an AI Overview. The informational, how-does-this-work, what-should-I-consider queries that B2B content marketing spent fifteen years targeting are precisely the queries AI answers best. Your top-of-funnel blog traffic did not get outcompeted. It got summarised.
The direction of travel is set, too. Gartner now forecasts that by 2028, 90% of B2B buying will be intermediated by AI agents: software researching, comparing and shortlisting suppliers on a buyer’s behalf. Whether that number lands exactly is beside the point. Some meaningful share of your future buyers will first encounter you through a machine’s summary of you. Plan for it.
The game is now citation, not clicks
Before anyone drafts the funeral notice for search marketing, the same research contains the opportunity.
AI answers cite sources, and being one of those sources matters enormously. Seer Interactive found that brands cited within an AI Overview see around 35% more clicks than uncited brands on the same results page. Semrush’s data goes further: visitors who arrive via AI-driven referrals convert at more than four times the rate of traditional organic visitors, because by the time they click through, the AI has already done the qualifying.
So the strategy shifts. Fewer visitors, better visitors, and a new job for your content: convince the machines you are worth citing, so the humans hear your name inside the answer. Some call it generative engine optimisation. The label matters less than the behaviour: structure content to answer questions directly, demonstrate genuine authority, and be present on the platforms AI systems draw from.
Which is where video enters the story.
Why video survives the AI answer engine
Video is holding its value through this shift for four reasons that compound.
YouTube is not being cannibalised the way websites are. It is the world’s second-largest search engine, and when someone searches for how something works, the video results still get watched rather than summarised. A text answer about how your product works competes with an AI summary. A video showing it does not, because watching the thing is the point.
AI systems also lean on video for their own answers. EMARKETER’s generative engine optimisation research notes that AI models increasingly prioritise video transcripts and metadata as citation sources. A properly published video, with a full transcript, descriptive metadata and video schema markup, is legible to every AI system deciding what to cite, while doing its normal persuasion job for humans.
Then there is conversion. The traffic that still reaches your website matters more, and video converts it. Those AI-referred visitors converting at four times the organic rate land on your pages closer to a decision. An explainer at that moment is the difference between a bounce and a demo booking. Research cited by HubSpot has long put the conversion uplift of video on landing pages in double digits, and it compounds when the visitors arriving are pre-qualified.
Finally, video works in the channels the click apocalypse does not touch. LinkedIn video watch time grew 36% year on year by the platform’s own figures. Buyers researching in feeds, communities and inboxes never touched Google in the first place. Native video meets them there.
The dark funnel just got darker
There is a wider truth underneath the traffic charts. B2B buyers were already doing most of their research invisibly, across communities, peer recommendations, social feeds and content consumed without a form fill. AI answers extend that invisible phase: now even the visible bit, the website visit, often does not happen until the shortlist is drawn.
That changes what your content is for. If a buyer’s first meaningful contact with your brand is an AI-generated answer, a YouTube result or a clip shared in a LinkedIn feed, then brand-level clarity does the selling long before your website gets a chance to. The question stops being “how do we rank” and becomes “when the machine composes an answer about our category, are we in it, and when the buyer finally arrives, do they already understand us”.
A distinctive, well-distributed explainer is unreasonably good at both halves of that question. It seeds the visual identity and the core message everywhere buyers actually spend the invisible phase, and it closes the understanding gap in ninety seconds when they surface.
A practical playbook
What this looks like as a to-do list rather than a thesis.
Publish video where the machines can read it. Every explainer goes on your site with a full transcript on the page and VideoObject schema markup, and natively to YouTube with a descriptive title, chaptered timestamps and a proper description. The transcript is not an accessibility afterthought. It is the text AI systems index and cite.
Restructure key pages to answer first. Open with a 40 to 60 word direct answer to the question the page targets, then let the video and the detail expand it. Pages built this way get cited by AI Overviews. Pages that make readers scroll for the point get summarised by them instead, with no credit.
Aim content at the questions AI answers poorly. Generic how-to queries are gone; the AI answers them adequately and nobody clicks. Comparisons with real numbers, genuine case studies, original data and demonstrations of your actual product are much harder to synthesise, and they sit closer to the purchase decision anyway.
Distribute natively, everywhere. The explainer is the master asset. The cutdowns live natively on LinkedIn and YouTube, building the brand recognition that pays off when a buyer asks an AI for recommendations and your name is one it has seen everywhere.
Change what you measure. Track branded search volume, AI citation appearances, and conversion rate per visit alongside traffic. If traffic falls 20% while branded searches and demo bookings rise, you are winning the new game while grieving the old one.
And check what the machines currently say about you. Ask ChatGPT, Perplexity and Google’s AI Mode the questions your buyers ask, including “who are the best providers of” your category. If you are absent or misdescribed, that is your content roadmap written for you, one gap at a time.
What about the blog you already built?
If you have years of blog content that used to pull traffic, do not torch it. Triage it.
Some of it still earns clicks, usually the comparison pieces, the pricing explainers and anything with original numbers in it. Those pages deserve investment: retrofit a direct answer block at the top, add an explainer or a short video walkthrough, mark it all up with schema, and they become stronger citation candidates than anything you could write from scratch.
Some of it is generic how-to content that AI now answers without you. Those pages have a different job now: they are raw material. A post explaining a process your product handles is a video script waiting to happen, and the video version can live on YouTube and LinkedIn where the summarisers cannot reach it. The research is done, the structure exists, the argument is proven. Repurposing it into video costs a fraction of the original writing and reaches buyers the original no longer can.
And some of it should be quietly retired. Thin posts chasing keywords nobody clicks any more are dead weight. Fewer, stronger, video-backed pages beat a large graveyard of orphaned articles in an era where authority signals matter more than page count.
The paid search wrinkle
One more budget line affected by all this: your Google Ads. AI Overviews compress paid results too, and Seer Interactive’s data shows paid click-through rates on affected queries falling even harder than organic. Costs per click keep rising while the clicks themselves get scarcer.
The response is not necessarily to spend less, but to waste less of what each click costs. Sending expensive traffic to a generic homepage was always leaky. In 2026 it is indefensible. Every paid pound should land on a dedicated page whose message matches the ad, and a ninety-second explainer on that page is the highest-impact conversion element you can add to it. When clicks are this expensive, the video that converts an extra one in ten pays for itself faster than any bidding tweak.
The uncomfortable but useful conclusion
The era of publishing text and harvesting clicks is not coming back. Every B2B marketer is deciding, this year, whether to keep optimising for a machine that is being switched off or start building presence in the one replacing it.
Video is the through-line. It is the format buyers prefer, the format feeds reward, the format AI systems cite, and the format that converts the smaller, sharper stream of visitors who still arrive. The businesses treating video as the centre of their visibility strategy are the ones whose pipelines will make sense in two years.
We build explainers designed to work this way: transcript-ready, cutdown-friendly, and distinctive enough to be remembered from a feed. You can see the standard in our work, and our pricing is published openly.
If your traffic graph has been telling you something is wrong, get in touch and we will write your first script free. The fix is not more blog posts.
