This line is from Arthur Weasley, warning his daughter Ginny (and Harry Potter) about an enchanted diary at the end of the second book. I love this line because the principle applies perfectly to the AI conversation nearly every CMO is having right now.
The real question isn't whether AI can think. It's whether you know what it's doing with that thinking — and whether you've built guardrails smart enough to keep it pointed at revenue and pipeline generation.
AI is not coming for your job. It is, however, coming for your excuses.
The calculator didn't replace accountants. The HP-12C didn't replace finance professionals. What those tools did was kill the manual busywork — and force the humans holding them to get sharper, faster, and more strategic.
(As an aside, AI didn’t invent the use of em dashes either. I’ve used and abused them since high school and winning writing awards, and hate that they are now associated with AI slop.)
AI is the same evolutionary moment, just moving at a speed that makes the spreadsheet revolution look like it happened in cursive (which, incidentally, is also increasingly becoming extinct).
Roughly speaking, companies that have embedded AI into their GTM workflows are generating roughly twice the net new revenue per head compared to light adopters. That's not theoretical. That's a unit economics problem for anyone sitting on the wrong side of it.
So why is that? Agentic AI — systems that can actually execute multi-step workflows, not just answer questions — can power up to two-thirds of current marketing activities. Content generation. Media planning. Synthetic and A/B audience testing. Automated optimization across digital channels. Not "might someday." Now. Not perfect, but still very cool.
But you also hear of organizations getting stuck in running pilots, and not getting to full production because the business case doesn’t warrant it. So companies test these disconnected pilots: More images generated, more copy variations spun up, more early-concept (and soulless) decks produced — and almost zero movement on the bottom line. The early assumption is the "gen AI paradox": The technology is everywhere except the P&L.
It feels like many AI pilot efforts are surface-level. The fix isn't more tools. It's rethinking workflows (entire workflows) from scratch, with AI baked in from the beginning — not tacked on at the end, or to simply summarize something already covered or done.
It’s about having experienced, unique lenses and workflows designed for you. At BrownRobinson, that's exactly what we're building. Our own MCPs (Model Context Protocols) let us run a client's current-state GTM and desired future state through multiple experienced marketing and revenue lenses simultaneously. What used to take weeks of intake and analysis now surfaces instantly with our focused inputs. It's our thinking. It's our frameworks, our experience, our strategic judgment. Just amplified.
That's the right frame for AI in marketing right now: Amplification, not replacement.
There's a lot of AI slop out there — generative junk cranked out to fill a content calendar rather than build a pipeline. The companies winning aren't the ones producing the most. They're the ones combining smart AI workflows with human judgment, creative instinct and presentation, and relentless focus on revenue. There's a reason brand and PR is again resonating…because of how AI finds company awareness and relevance…which clarifies overall spend and ROI.
Here's what that focus should look like in practice: Your AI systems need one guardrail — pipeline. Not engagement metrics. Not impressions. Not MQLs for the sake of MQLs. Build your agentic workflows around revenue, measure against it, and ruthlessly cut anything that can't trace a line back to it. The pipeline you had yesterday means nothing if it didn't convert — and the best thing about well-designed agentic workflows is that they're constantly restocking, refining, and getting smarter about what actually moves the needle.
Arthur Weasley (yes, that was Ron's dad) had a point. You should know where your AI keeps its brain. But thinking through marketing like an engineer, you should be the one who put it there.
