What You Told Us About AI Slop at Drive 2026

I lost my voice at Drive, and it was worth it.

Three days in Stowe, Vermont, and I spent nearly all of them at our booth talking with marketers about the least glamorous part of their job: what happens to AI-generated content after the draft is done. If you were one of those people, thank you. This post is mostly for you.

Markup AI sponsored Exit Five’s Drive conference September 8 to 10, and I want to start with the thank-you. Thank you to Dave Gerhardt and the Exit Five team for building an event that marketers actually want to attend. Thank you to everyone who stopped to tell us what their week really looks like, especially the people who were honest about the messy parts. And thank you to the group who stayed after our workshop to keep talking!

Here’s what you told us, what we did with it, and what you can take with you.

What Drive is, and who it’s for

Around 400 in-house B2B marketers took over the Lodge at Spruce Peak for three days. The event included keynotes, peer-led workshops, morning hikes, speed networking, and a talent show.

What sets Drive apart is who’s on stage. The agenda is practitioner-led, so the sessions come from marketers doing the work right now at real companies, with real budgets and real constraints. The topics tracked closely to what’s on marketing teams’ plates this year: brand strategy, AI in the content workflow, demand gen, LinkedIn ads, email, and ABM. Seth Godin and Ann Handley anchored the keynotes, joined by Kara Hardin of the Practice Lab and Louis Grenier. 

What you told us

We came to Stowe expecting to explain the problem we solve. Instead, you explained it to us.

One marketer told me she felt like she was being “claude-er-boarded” by AI content. A constant stream of drafts landed in her queue. All were technically finished, but all still needed her to clean them up before anything could go live. I haven’t stopped thinking about that word. It captures the shape of the problem better than anything in our messaging does. The volume isn’t the win everyone expected. The volume is the new work.

Four themes came up again and again:

• You’re generating plenty of content, and you still can’t get it to sound like your brand without heavy human editing. The draft arrives fast. The rewrite takes just as long as it always did.
• The speed and quantity of content built for AI search has created a real feeling of lost control. Things are moving too quickly to review the way you used to.
• Nobody wants to be the person who published the slop. The reputational risk sits with you, not with the model that wrote it.
• The tool glut is exhausting. You see dozens of shiny options every quarter and have to guess which ones matter.

That last one is a challenge for us as much as for you. Plenty of things at a conference are interesting. Very few survive the trip back to the office.

The pattern underneath all four

Here’s what struck me after dozens of conversations: Almost nobody described a writing problem. Every single person described a content control problem.

The drafting is handled. What’s missing is the step between “the draft exists” and “this is safe to publish?” The moment where somebody has to decide whether the claim is accurate, whether the terminology is right, whether it sounds like the brand or like every other brand in the category. That step used to be small because the volume was small. Now it’s the bottleneck, and it lands on one or two people who are also expected to do everything else.

If that’s your week, you’re not behind. You’re describing the same thing hundreds of other marketers described.

What we covered in the workshop

Our session, “Stop the AI Slop: How to Scale Content Without Sacrificing Quality,” sold out, which tells you something about where marketers’ heads are right now. If you couldn’t get a seat, here’s the short version.

1. Put the gate before publishing, not after. Most teams review at the end, when the content is already scheduled and the pressure is on to approve it. Moving the check earlier, right after generation, turns a judgment call into a routine step.
2. Separate the mechanical checks from the judgment calls. Terminology, banned claims, tone drift, and accuracy checks can run automatically on every asset. Positioning, strategy, and the decision to publish belong to a person. Mixing the two is what makes review feel endless.
3. Write your standard down, then make it enforceable. “It should sound like us” isn’t reviewable. A documented brand voice profile is, and it’s the difference between a reviewer’s opinion and a consistent standard your whole team can apply.
4. Consistently check what you publish, so you can see drift coming. Teams that measure content against a standard catch brand drift in weeks. Teams that rely on gut feel catch it during a rebrand, which is expensive.

None of that requires additional headcount. It requires deciding what “ready to publish” means at your company and then enforcing it the same way every time.

About the brand voice profile you took home

Every workshop attendee left with a personalized PDF of their own brand voice profile, built from their real content. We also ran them live at the booth, which turned out to be the best conversation starter we’ve ever brought to an event. You’d tell us your website, we’d pull your profile up on screen, and we’d look at what your content actually sounds like today versus what you think it sounds like.

If you have one, here’s what to do with it:

• Share it with everyone who writes for your brand, including contractors and agencies. Most voice inconsistency comes from people who never saw a guideline, not from people ignoring one.
• Paste the relevant sections into the prompts and custom instructions your team already uses. It won’t solve the problem on its own, but it moves the starting point.
• Use it as the reference in your review step, so feedback stops being “this doesn’t feel right” and starts being specific.
• Revisit it after your next positioning change. A voice profile written against old content will quietly enforce old messaging.

If you didn’t get one and want one, reach out. We’ll build yours. No strings attached.

The postcards, the beanies, and the bracelets

Marketers know good swag. It mattered to me that we bring something you’d actually be glad to take home, so we built the booth around one idea: give people something to do, not something to scan.

More than 100 of you sat down at our postcard station, picked a Vermont postcard, and wrote a note home. The beanies turned out to be exactly right for a Vermont morning. And I was shocked to see how popular the bracelets were! Which one was your favorite?

Thank you, Drive. Thank you to every marketer who told us the truth about their workflow instead of the polished version. That’s the part we’ll remember, and it’s already changing how we build.