A new report from Search Engine Journal, covered by Matt G. Southern, found that TikTok surfaces roughly three times as much AI-generated slop as YouTube. That's a big number for anyone running B2B content, paid social, or account-based marketing. The immediate temptation is to shrug it off as a consumer problem. It isn't. The same recommendation systems that flood a scroll feed with low-effort AI clips also decide which of your buyer's employees see your brand between the memes.
The practical question for B2B teams: if your target account's marketing director spends 40 minutes on TikTok tonight and 30 of those minutes are AI-generated filler, how does that change the way you should build a pipeline off social? Below is the read from our own dataset of 1.6M companies and 175M contacts, plus what we're telling customers to do about it this quarter.
What the 3x figure actually means for B2B
The SEJ report is about volume of AI-generated content on the platform, not about intent quality. But volume changes intent quality by proxy. When feeds get noisier, three things happen: organic reach for genuine brand content drops, CPMs on the surviving high-quality inventory rise, and users get faster at bailing from anything that looks templated. In our own dataset, the median B2B buyer we track visits 4.2 vendor sites before booking a demo. When we cross that with public LinkedIn activity, we're seeing those same buyers spend 22% less time on short-form video landing pages than they did in Q1. The scroll is getting harder to interrupt.
Concretely, we watched a mid-market SaaS customer (a Notion-style workspace tool) burn through a $48K TikTok campaign in six weeks with a 0.3% CTR and zero attributed pipeline. Same team, same offer, ran a two-week LinkedIn Document Ads test against a matched company list from our database and closed $112K in influenced pipeline. The delta wasn't creative. It was audience noise.
Related: Google Reveals AI Visibility Inside SEO Tool, Aims to Boost B2B Sales & Marketing Efforts
Where the buyers actually are (and where they aren't)
The slop problem doesn't kill TikTok as a channel. It changes what TikTok is for. If we look at how identified B2B visitors show up across referrer data on the sites we track, the pattern is consistent: TikTok drives awareness clicks, but very rarely the second or third visit that precedes a demo request. Companies like HubSpot, Ramp, and Klaviyo are all still spending on the platform for brand — but the pipeline math lives elsewhere.
Here's how we're advising customers to reallocate this quarter:
- Treat TikTok as top-of-funnel only. Measure lift on branded search and direct traffic, not demo requests. If branded search from your ICP regions doesn't climb within 30 days, the spend isn't working.
- Move mid-funnel to LinkedIn + retargeting on your own site. Identify the companies visiting after a TikTok view and hit them with account-based ads. Our customers using website visitor identification are catching 32-38% of otherwise-anonymous B2B sessions.
- Kill any short-form ad that looks AI-generated. In a slop-heavy feed, the pattern-match is instant and the click-through collapses. Real faces, real screens, real product footage beat stock imagery by a factor of 4-6x in our customers' TikTok data.
- Feed the CRM, not the vanity dashboard. Impressions and reach numbers on TikTok are inflated by non-buyer accounts. Filter every reported number by whether the account is in your ICP company list before you celebrate.
The identification problem gets worse before it gets better
The awkward truth: AI slop isn't just competing with your ads. It's competing with your identifiability. As feeds fill with generated content, platforms respond by tightening tracking, adding friction to click-outs, and shortening cookie windows. That means the analytics stack most B2B teams built in 2022 is quietly getting worse at telling them who's actually engaging.
This is where deanonymizing your own site traffic becomes non-negotiable. When Zapier, Cloudflare, or Snowflake employees land on your pricing page after seeing a TikTok clip, you need to know that — even if the platform won't tell you. Across 1.6M tracked companies, we resolve about 68% of B2B website sessions to a named account, and roughly 41% to a specific individual with verified email and LinkedIn. That's the delta between a lead you can call and an anonymous session ID.
Related: Where Clicks Go, What Agents Skip, Who's Leaving Bing – SEO Pulse via @sejournal
What to do this week
If you're running paid social in a slop-heavy environment, the fix isn't more spend, better hooks, or a fresh creative agency. The fix is closing the identification gap on the traffic you already have. Three things to do before Friday: audit your last 30 days of paid social traffic and calculate what percentage resolved to a real account; pause any campaign where less than 20% of sessions match your ICP; and route the remaining identified accounts into a sales workflow that treats them as warm, not cold.
If you want to test that identification stack against your own traffic, our pricing starts at a level where a single closed deal from a previously anonymous visitor pays for the year. The 3x AI slop number is a symptom. The underlying problem — anonymous engagement across noisy channels — is the one worth fixing.
