The best insight I’ve heard about AI at work compares it to Saturday Night Live. That image landed hard with me. The show is packed with talented people chasing airtime. Without a strong editor, the chaos wins. My take is simple: AI doesn’t need more makers—it needs better editors.
On Marketing Against the Grain, HubSpot leaders Kipp Bodnar and Kieran Flanagan nailed the current problem. Companies are racing to try every idea an AI model can spit out. The result is a flood of quick builds, shallow wins, and busywork dressed up as progress. I agree with their core point: creativity without constraints wastes time and confuses teams.
The SNL Lesson for AI Teams
Kipp and Kieran point to Lorne Michaels as the model for how AI should be managed. His job is to shape and limit the show’s creative force so it becomes good television. Without that role, you get noise.
“Using AI in a company is a little bit like Saturday Night Live… [Lorne Michaels’] whole job is to basically… edit creative people and stop them from getting in their own way.”
That’s the job AI leaders need to do now. Not ship every idea. Not greenlight every test. We need structure that filters, prioritizes, and kills weak bets fast.
Stop “Token‑Maxing.” Start Managing.
They also called out a habit I see daily in crypto, marketing, and software teams. Folks try ideas just because tools make it easy. They build a chatbot here, a script there, then move on. No strategy. No owner. No outcome.
“We’re basically just letting everybody take every idea… and just do it… what you actually need is… structure… constraints to get some of the really dumb token‑maxing out of the way.”
I’ve done this myself. So have many smart people I know. When creation is cheap, shallow work feels productive. It isn’t. Speed without purpose is still waste.
What Works Instead
I’ve built businesses around new tech for decades. The teams that win use guardrails. They don’t kill creativity. They guide it into outcomes that matter. Here’s how to reset your AI program this week.
- Define one to three business problems worth solving. Tie each to a real metric like lead quality, activation rate, or cost per ticket resolved.
- Appoint an editor. One accountable owner decides what ships, what dies, and why.
- Set a small budget of time and tokens per idea. Force trade‑offs.
- Demand a measurable result in two weeks. Keep what moves the metric. Cut the rest.
- Standardize winning workflows. Turn ad‑hoc scripts into team playbooks.
These steps turn experiments into a system. They also reduce context‑switching and the “thirty minutes here, thirty minutes there” trap that Kipp and Kieran warned about.
But Isn’t More Experimentation Good?
Fast testing is good. Random testing is not. The counterargument says we should try everything because AI is new and cheap. That sounds bold. It drains focus.
AI is like capital. Cheap money still needs due diligence. Cheap tokens still need editing. Ideas don’t earn airtime. Results do.
How Leaders Should Talk About AI Now
Teams want freedom. Give it, inside a plan. Be clear that shipping isn’t the goal; outcomes are. Kieran’s candid aside—“that is you and I”—is the culture cue. Even leaders fall for toy projects. Call it out. Fix it.
“Oh well, I can build this thing now, so I’m just going to… whether I should or not.”
I’ve seen this movie during the app boom, the ICO surge, and the creator wave. The pattern never changes. Tools spike. Noise floods. Editors win.
The Editor Era of AI
My stance is firm: AI needs fewer experiments and more decisions. Not less creativity—better curation. Be the Lorne Michaels of your AI program. Protect the stage. Cut the weak bits. Promote the hits.
Pick one business metric. Appoint an editor. Ship only what moves that number. Retire the rest. That’s how you turn AI from a toy box into a growth engine.
Now, make the call. Say no to the next clever demo. Say yes to the boring workflow that saves hours every week. Then repeat. That’s how winning shows stay on the air.
