🗝️ Keepers of the standard

Over the years, I’ve pulled plenty of people from outside marketing into marketing work.

Salespeople told us which questions buyers ask in meetings, to inform the content plan. Engineers explained why a SaaS feature was built the way it was, for a launch campaign’s behind-the-scenes angle. E-comm category managers gave us this year’s “it” toy to add in the Christmas gift guide.

Every one of those pieces still went through me. I’d get the insight out of the expert’s head, then write the whole thing myself. That made me the bottleneck for all of it.

So when OpenAI’s CMO Colin Fleming published an essay last month proclaiming that now everyone can do marketing, my first reaction was: yes, please.

The idea comes with real issues, though. As Robert Rose of the Content Marketing Institute points out in his response, it can turn marketing into the “AI cleanup crew,” spending time fixing drafts, when marketing has a far more important role to play.

When smart people argue, I usually listen—and this time, I’ve got Opinions, too.

The argument lies in the sequence

Fleming’s argument is that AI removes one of marketing’s oldest constraints: “how far someone can take an idea before they have to hand it off.” Marketing still owns the customer insight and “the final call,” and its job moves “from controlling who gets to participate to setting the standard for what deserves to ship.”

Rose mostly agrees, but worries about how that plays out in practice. Put the standard-keeping job at the end, and marketing gets demoted to the approval desk. An executive generates a white paper over the weekend and hands it over on Monday to “clean up.”

His fix is to flip the sequence: marketing writes the brief first, setting the standard for “what deserves to be started.”

I’m with both of them on the core. Marketing sets the standard, and much of that happens up front.

Here’s what I’d add. Both essays focus on where marketing sets the standard: as a verdict at the end, or as a brief at the start. I’d say “both, and.” Once people across the company are making things, marketing also has to keep it: before the work starts, while it happens, and after it ships.

Before: give the standard somewhere to live

Where does your quality standard live right now?

If it lives only in the heads of senior marketers, it won’t travel far. Memory is a lossy format, and every new person making content adds another version.

As the strategist, you work out what the standard is: collecting input from sales, product, and leadership, and seeing where their views fit and clash. You make sure the calls nobody can settle alone get settled, so everyone works from one agreed baseline.

As the tech architect in an AI world, you encode it: the messaging, the voice rules, and the quality checks, written down and built into AI tools you give people to draft with. Fleming’s numbers suggest marketers are already heading there: 39% of their chat exchanges with AI involve working on code, and only 12% involve actual writing tasks.

During: governance in disguise

If this is starting to sound like governance, that’s because it is.

Governance used to mean training workshops, a monthly update in the company newsletter, a slot at the all-hands. But if everyone is doing marketing now, as Fleming puts it, who from marketing is in the room when a product manager sits down on Thursday afternoon to write about the new release?

With the standard encoded in AI ops, governance happens under the guise of individual work. While helping that product manager draft, the tool shows them the standard in practice and teaches them how to think about it as they go.

A check that explains the issue and why it needs fixing (say, “we don’t call ourselves all-in-one, because our positioning rests on doing one thing well”) does what team marketing training used to do. Only now it’s one-on-one instruction, on their own draft.

However, proper “talking to people” governance is still here. None of it works if the AI setup arrives through the “Godspeed” method. Someone hands you a login and says, “Here, use it. Godspeed!” Getting people on board is part of building the process: show them what the AI workflow is, how quality checks are embedded, and what they get out of it—it makes them much more likely to use it all.

After: guarding the gate (and deciding how tightly)

Some checking still happens once the work is done. There are two ways to do it: a strict check before anything goes live, or letting people publish and nudging them afterward with ideas on what would’ve improved the piece.

Neither is perfect. With the strict gate, marketing becomes the bottleneck again and starts to feel like a watchdog, which strains its relationship with the other teams. The looser gate has its own risk: a critical error in judgment goes out and puts you in hot water. (I still believe that happens way less often than people fear, but can be an issue.)

Which one fits your company? I don’t have a clean answer. It’s one of the judgment calls I still can’t operationalize. What I do know is what I’d consider when making the judgment:

• Track record. How the people publishing have handled the standard so far.
• A culture of ownership. Whether you trust people to uphold the standard without skipping steps to ship faster.
• Company maturity. A startup’s messaging is still moving, so it’s hard to tell what holds, and a tighter gate usually makes more sense. A mature company has stable messaging and enough existing content to show what good looks like.

***

Rose is right that simple cleanup is a demotion. For me, what makes it one is how it happens: in the back room, alone, simply assuming “marketing will take care of it.”

Keeping a standard runs the other way. You’re moving judgment out of your own head and into other people’s, and you do it with them. Nobody on the team should end up as the support crew for someone else’s work, including the marketers.

To get this process going, think about this. Is there someone outside marketing you’d love to see making content? That’s the conversation I’d have first: what they’d need from you to do it well.

Autonomy matters,

👌 Handpicked stories for you

Here are some interesting things I stumbled upon this week while reading, writing, researching, and being a generally curious marketing person on the internet:

💼 Three for work

How We’d Market a $400 Skincare Device Before Black Friday | I love SparkToro’s audience research examples, and this one is a treat. Amanda Natividad maps where LED face mask buyers spend their time online (they over-index on Claude by 221%, of all places). A great model for building strategy around where your audience already is.

The Creative Velocity Gap: What B2B Content Teams Should Steal From How Performance Marketers Iterate | What can we borrow from performance marketers, who ship 30 to 50 creative variants a week? I’m fully on board with two-week sprints (I already recommend them in my Marketing Ops work with clients) and briefs written like user stories. I’m less sold on testing angles on social and email first, since B2B volumes rarely allow a clean read. Anyway, the piece raises interesting points.

How We Use AI for Every Article Without Making AI Slop | When Ahrefs first shared their AI pipeline that drafts a nearly publish-ready article “in minutes,” I rolled my eyes. This piece explains what happens around it: the ideation work up front, plus four quality gates (idea, outline, evidence, draft). They slow the process down, which is exactly what keeps it from producing slop.

👀 Two for fun

This ranking of IKEA furniture by assembly pain | A fun research and data visualization project. Take a look before you get into a row over the exact way to assemble your Fågelfjället bedroom. (My marriage narrowly survived the chart-topping PAX wardrobe!)

This live journal of things people are putting off | If you’re a procrastinator, join your fellows here. Log what you’re avoiding and collect an official Bureau of Procrastination permit to delay it. (But actually, you know, don’t.)

🐈 Weekly cat

Food critic in training…