But everybody did their job...
Companies are organized into departments. Customers aren’t. The most expensive problems live in the seams between them.
When growth stalls, the executive instinct is almost always to go looking for the department to blame. If pipeline looks anemic, leadership points at marketing. If deals stall out in the pipeline, it’s a sales problem. If customers churn after ninety days, product gets the heat, and if release dates slip, everyone looks at engineering. It’s an intuitive way to diagnose a business, and after twenty years of doing this, I’ve watched it be dead wrong far more often than right.
Because the most expensive problems inside a $10M to $50M business don’t live neatly inside the functional boxes. They live in the seams between them.
I’ve spent most of my career moving between engineering, product, marketing, and operations. And the issues that actually cost a company twelve to eighteen months are almost always cross-departmental handoffs that went sideways. A marketing crisis that’s actually an unaddressed product flaw. A product roadmap disaster that started with an over-promising sales pitch. Or a retention leak that was baked into the onboarding flow before the buyer ever wrote their first check.
Every single department can be hitting its local numbers while the business as a whole runs right off a cliff.
Everybody did their job
Take sales and product at a travel SaaS company I worked alongside. The executive team kept hearing a persistent complaint from the sales floor: nobody will buy our platform without AI. So leadership did what seemed logical. They stood up an initiative, hired engineers, built recommendation systems, and burned through seven figures of capital.
And prospective customers still refused to buy.
The company had taken a sales objection and translated it straight into a product requirement without digging into the root cause. Engineering built exactly what was specified. Product could proudly point to the new features on the roadmap. Sales had the exact talking points they asked for. On paper, everybody did their job. In reality, the business had solved the wrong problem.
Put a revenue stall in front of five executives and you’ll get five completely reasonable explanations. Marketing insists they need bigger budgets. Sales demands better-qualified leads. Product wants clearer positioning. Engineering needs finalized requirements, and customer success swears the onboarding experience is broken. And look, they aren’t lying. Every one of them is describing a real symptom they can see from their own window. But nobody in the room is looking at the system as a single machine.
Marketing might still be spending money to acquire the buyer persona that worked two years ago, while sales has quietly figured out that a totally different customer is the only one actually signing contracts. Meanwhile, customer success knows exactly why churn is spiking, but that knowledge never once travels back across the hallway. That gap is the seam nobody owns.
A room full of eager interns
That coordination problem was already hard. Generative tools just turned up the volume.
The way I picture it, AI dropped a room full of hyperactive, dirt-cheap interns underneath every single department. They can produce a massive volume of work in an afternoon. But none of them know what’s actually true this quarter.
Your marketing team has five years of old strategy decks, outdated positioning docs, and legacy collateral sitting in Google Drive. Ask an AI model to whip up a new go-to-market campaign, and it’ll synthesize all of that historical material into something that sounds completely polished. It looks like magic right up until you read the substance. The market moved. The product changed. But the model doesn’t know which parts of your company history are dead unless an operator explicitly tells it. So the new strategy comes out as an echo of last year’s mistakes, wrapped in fresh corporate prose.
The obvious errors are easy to catch. It’s the plausible ones that slip through: the slightly stale messaging, the target account profile that drifted three degrees off course. Then those outputs get fed into sales ops, summarized by support, and handed over to product. We used to worry about misalignment as a slow, human drift. Now companies can automate it. It’s drift on steroids.
Pretty-good output ships
When producing work was expensive, companies naturally inspected it before it left the building. When a task takes thirty seconds, the temptation to fire and forget is almost irresistible. Terrible output gets rejected immediately. But pretty-good output ships, because nobody has the time or the friction to stop it.
Mind you, this didn’t start with generative AI. A while back, I audited a marketing agency reporting an 11.7× return on ad spend for an e-commerce brand. On the executive slide, that’s an absolute banger. Nobody on the board questioned it, because why would you challenge a number that looks that good?
When I traced that 11.7× back to the actual merchant bank deposits, the entire campaign had produced $918 in net sales. The reporting platform was taking credit for organic transactions the ads happened to touch, not cause. The company had been burning cash for months on an illusion.
AI simply accelerates that dynamic across every department at the exact same time. Generation gets dirt cheap, while real discernment becomes the rarest commodity in the building.
Be suspicious when every dashboard is green
One of the strangest situations you can walk into is a company where every single department can prove it’s winning, but the business as a whole is missing its growth targets.
Marketing hit its MQL goals. Sales crushed its meeting quotas. Engineering’s sprint velocity is at an all-time high, and product shipped the entire quarterly roadmap. Yet revenue hasn’t budged an inch. Which is wild, and happens far more often than leadership likes to admit.
Arguing harder about those individual dashboards won’t fix it, because none of those numbers are fraudulent. The question is whether the handoffs between them are coherent. Are you marketing to the people sales can close? Are customers actively using what engineering built? Does using the product actually produce the business outcome sales promised during the demo?
The real leverage is always in the seams. Where does a project leave one team’s hands and enter another’s? Where does a metric quietly change its definition? That’s where companies lose momentum, burn cash, and leak millions in enterprise value without a single red light flashing on a dashboard.
Somebody has to hold the loop
So who actually owns a problem that requires marketing, product, sales, and engineering to all change their behavior at the same time?
There’s no universal corporate title for it. Sometimes it’s the Chief Revenue Officer, sometimes it’s RevOps, and sometimes it’s just the founder refusing to look away. But somebody has to own the issue across every functional boundary it crosses, from the moment the leak is spotted to the day the ledger proves it’s solved.
If nobody owns the seam, the problem just gets bounced around the org chart like an unassigned ticket. Every department closes its tasks, everyone gets their quarterly bonus, and the company still slowly gets screwed.
Sometimes a team really does just need better execution. But when every dashboard looks green and the business is still stalling out, stop staring at the departments.
Look at the space between them.
Know where this has to go, and not how?
Twenty minutes. Tell me where you want this company to be in two years and I’ll tell you what I’d go look at first.