On July 28, marketing production platform Knak published a survey of more than 300 enterprise marketing leaders. Two numbers from it belong side by side. 70% of these teams have AI deployed in production. And 85% of them still missed at least one planned campaign launch date in the past twelve months.
Read that again. The teams with the biggest budgets, the newest tools, and dedicated AI lines in their spend — and five out of six of them still can't ship on schedule. One in ten missed launches more than five times last year.
The survey also priced what all that machinery costs. At 60% of these companies, producing a single marketing email involves four or more people. At 69%, it goes through two to three rounds of revision. At 54%, it travels across three to five separate tools on the way out the door. Using Bureau of Labor Statistics pay data, Knak put the internal labor at roughly $300 per send. For one email.
The delay was never the draft
Here's the part that matters for anyone running something smaller. When these teams named what actually caused their missed launches, AI's job — the writing — wasn't on the list. The causes were approvals and sign-off (47%), design and creative production (38%), and cross-team coordination (36%). Half of these teams still route final approval through email, Slack, or Teams threads. The draft takes minutes now. The thread where three people go quiet on it takes two weeks.
Which explains the other awkward number in the report: 88% of these leaders say AI output still needs moderate to substantial human editing before it can be used. So the sequence goes AI draft, human rewrite, design queue, revision rounds, approval thread, missed date. The fast step got faster. The slow steps didn't move.
AI adoption measured in drafts is a vanity metric. The only number that tells the truth is what shipped.
There's one more detail worth noticing. Asked where they actually use AI, 64% of teams said first-draft copy and 56% said image generation — the easiest work to hand off. Only 25% use it for building and coding the emails and landing pages themselves, which is the step that eats the most production time. They pointed the machine at the pleasant part of the job and left it out of the expensive part.
Being small is the advantage here
If you're a business of one, or a marketing team of two, every number in that survey describes a problem you don't have. You don't have four people per email. You don't have an approval thread, because you are the approval thread. The entire $300 pipeline — the sign-offs, the coordination, the three-to-five tools — collapses into one person reading one draft and deciding.
That's not a consolation prize. It's a structural edge. The enterprise can outspend you on every tool you'll ever touch, but it cannot approve an email faster than you can. When AI makes drafting nearly free, the winner isn't whoever generates the most drafts — it's whoever has the shortest path from draft to sent. Right now, that's you.
The edge only holds if you don't recreate their pipeline in miniature. Every extra tool the work passes through, every "let me sit on this a day," every revision loop past the second is you rebuilding the enterprise bottleneck at small-business scale, voluntarily.
The one habit to steal from this report
Track your marketing by what actually went out. Not drafts written, not ideas in the notes app — sends. Emails delivered, posts published, follow-ups fired. Once a week, count them. If AI has you producing more drafts but the shipped count hasn't moved, you've found your own version of the approval thread, and that — not another writing tool — is the thing to fix.
And point the AI at the step that's actually slow for you. If writing was never your bottleneck, an AI writer won't change your week. If the slow part is building the email, formatting the post, or the follow-up you keep not sending — automate that. The enterprises in this survey got that backwards at $300 a send. You get to learn it from their invoice.