Marketing teams have AI now. They still miss their own deadlines
A July 2026 survey of 300+ enterprise marketers found 85% missed a campaign launch date despite widespread AI use. The bottleneck moved, it didn't disappear.
By David Lara, Founder
Founder-reviewed ·How we research and correct articles
Knak published a report on July 28, 2026 that should have been good news for marketing teams and instead reads like a warning. Surveying more than 300 enterprise marketing leaders, it found that 70% have AI deployed in production — drafting, generating variants, speeding up the first pass on nearly everything a team produces. That part isn’t surprising; by mid-2026 it would be more notable if a marketing org hadn’t adopted some AI drafting tool.
The part that should surprise you is what happened to the thing AI was supposed to fix. Despite that adoption, 85% of the surveyed teams missed at least one planned campaign launch date in the past 12 months, and 1 in 10 miss launches more than five times a year. AI got faster. Launch dates didn’t move.
This isn’t an isolated finding from one vendor with an agenda, either — it lines up with the broader pattern showing up across enterprise AI-adoption research all year: organizations are deploying AI tools faster than they’re redesigning the processes those tools sit inside. Bolting a faster first step onto an unchanged pipeline doesn’t make the pipeline faster; it just moves the visible bottleneck to whichever step used to be second in line. Knak’s report is one of the first to put a hard number on exactly how little that first-step speedup has moved the metric marketing teams actually care about: whether the campaign ships when it was supposed to.
Where the time actually goes
The report’s explanation for the gap is the useful part. AI is getting teams to a first draft, not to a shipped campaign — 88% of respondents say AI output still needs moderate to substantial human editing before it can go out, and the surveyed workflows are heavier than “write, then send” implies: 60% of teams involve four or more people in producing a single campaign, 54% juggle three to five separate tools to do it, and 69% need two to three rounds of revisions before anything launches.
None of the top three causes are writing problems. Approvals and sign-off top the list at 47%, ahead of design and creative production at 38% and cross-team coordination at 36%. AI attacked the one stage of the pipeline — the first draft — that was never actually the slow part for a team of any real size. The slow part was always getting four-plus people to agree, across three-to-five disconnected tools, in two-to-three review rounds, before anything ships.
Why “needs editing” isn’t a small caveat
It’s tempting to read “88% needs moderate to substantial editing” as a rounding error on the way to full automation — a gap that closes as the models get better. That misreads what the editing is usually for. A first draft can be fluent and still be wrong about the thing that actually matters: whether the claim in paragraph two is one legal actually cleared, whether the tone matches what this specific segment responds to, whether the personalization field pulled the right account name instead of a stale one from three re-orgs ago. Those aren’t fluency problems a better model fixes. They’re judgment calls that need a person who knows the account, the brand, and the compliance line — which is exactly why the editing step survives even as the drafting step gets faster, and why 85% of teams still miss launch dates despite 70% already running AI in production. The report’s finding isn’t “AI isn’t good enough yet.” It’s that drafting speed and shipping speed were never the same bottleneck.
The same pattern, with a sharper edge, in cold email
If you run outbound, this should sound familiar, because it’s the same failure mode with a shorter runway. AI can write you a passable cold email sequence in seconds now — personalized first lines, variant subject lines, a full multi-step cadence from a one-paragraph brief. None of that speed matters if the campaign still can’t launch on time, and in cold email the things that actually gate launch are rarely the copy: it’s whether the audience is built from a clean, current segment instead of a stale spreadsheet, whether the sending domain has enough warmup headroom to take the volume, and whether someone actually reviewed the AI’s first-draft output for the tells that make it read like AI before it goes to a few thousand real inboxes.
The stakes are also higher than a missed blog post or a delayed social calendar. A missed marketing launch date is a scheduling embarrassment. A cold-email campaign that launches without the audience properly segmented, or before the sending domain is actually ready, doesn’t just launch late — it can burn the exact reputation you need for every campaign after it. Speed at the drafting stage that isn’t matched by speed at the audience-building and infrastructure stage doesn’t get you to launch faster. It gets you to a launch date you still miss, for the same coordination reasons Knak’s survey describes, except now with a domain-health cost attached if you rush around the gap instead of closing it.
Picture the actual week this usually plays out over. Monday, the AI tool drafts a five-step sequence and three subject-line variants in minutes — genuinely impressive, genuinely fast. Tuesday and Wednesday go to legal and brand review, the same 47%-of-the-time bottleneck Knak’s survey names. Thursday, someone discovers the audience list was last pulled from the CRM two weeks ago and needs to be re-segmented before it’s usable. Friday, someone else notices the sending domain hasn’t seen this volume before and either has to throttle the launch or risk the reputation hit. The campaign ships the following Tuesday — a week late, and the AI draft that took four minutes on Monday had nothing to do with any of the four days that actually slipped.
What actually closes the gap
The fix isn’t a faster AI drafting tool — most teams already have one, and it’s already fast enough. It’s removing the coordination tax around the draft: one place to build the audience instead of a spreadsheet a different person maintains, one place to see whether a sending domain is actually ready for volume instead of assuming it is, and one workspace where the campaign, the audience, and the senders live together instead of across three-to-five disconnected tools that Knak’s respondents describe fighting through on every launch.
| Typical stack Knak's respondents describe | One consolidated outbound workspace | |
|---|---|---|
| Audience source | A spreadsheet a different person maintains | A live, filterable segment inside the same workspace |
| Sending readiness | Assumed, not checked before launch | Domain and warmup status visible before you hit send |
| Tools involved per launch | 3 to 5, per Knak's survey | 1 |
| Review rounds before launch | 2 to 3, per Knak's survey | As many as you want, with nothing to hand off between them |
That’s the specific gap Norbelys is built to close for outbound. Audience segments, the program builder, sender warmup status and A/B testing all live in the same workspace a campaign actually launches from — not a spreadsheet feeding an AI drafting tool feeding a separate sending platform feeding a separate reporting dashboard. If your team can already draft a cold-email sequence in seconds and still watches launch dates slip for the same reasons Knak’s 300 marketers describe, the bottleneck isn’t the writing. See Norbelys pricing and get your next campaign from draft to a domain that’s actually ready to send it.
AI drafting speed vs. campaign launch speed
If AI drafts a campaign in seconds, why does Knak's survey still show 85% missing launch dates?
Because drafting was never the slowest stage for a team of any real size. Knak's respondents cite approvals and sign-off (47%), design and creative production (38%), and cross-team coordination (36%) as the actual causes — none of which get faster just because the first draft arrived sooner.
Does this apply to cold email specifically, or just broader marketing content?
The same pattern applies, with higher stakes. A cold-email campaign's launch is gated by audience quality and sending-domain readiness at least as much as by copy — and unlike a delayed blog post, a rushed cold-email launch can damage a sending reputation that takes weeks to rebuild.
What actually fixes the coordination bottleneck, if not a better AI writer?
Removing handoffs between the tools involved in a launch, not speeding up any single one of them. Knak's respondents juggle three to five separate tools per campaign on average — consolidating audience, sending, and reporting into one workspace removes the handoffs themselves rather than making each one faster.
Should a small team without a big approval process still worry about this?
Yes, in a smaller form. Even a two-person outbound team hits a version of the same handoff tax: pulling a list from one place, checking domain readiness in another, and drafting in a third. Fewer people doesn't remove the coordination cost, it just makes one person absorb all of it.