Ask any current model for a campaign email with three subject-line variants and you will have a draft in under a minute. Then ask your team how long that same email takes to actually reach the list.

Days. Sometimes weeks.

Michael Hammer measured that gap in 1990, a long time before anyone had a language model. If you run a marketing team that just bought a stack of AI tools, his article is worth twenty minutes of your week.

17 minutes inside 22 days

The piece ran in Harvard Business Review in July 1990 under a title that leaves no doubt: Reengineering Work: Don’t Automate, Obliterate. Hammer’s claim was that heavy investments in information technology had delivered disappointing results, largely because companies used technology to mechanize old ways of doing business. The process stayed intact. Computers just ran it faster.

His best example was Mutual Benefit Life. An insurance application went through as many as 30 discrete steps, 5 departments and 19 people. Typical turnaround was 5 to 25 days, and most of that time was spent passing information from one department to the next.

Then, almost as an aside in parentheses, he dropped the number I cannot stop quoting. Another insurer estimated that an application spent 22 days in process and was actually worked on for 17 minutes.

Make those 17 minutes ten times faster and the application still takes about 22 days.

What Mutual Benefit Life changed

They did not buy each department a faster system. They removed the handoffs. A new role, the case manager, owned each application from arrival to issued policy, backed by a workstation that pulled together what used to live in five departments. Specialists like the senior underwriter or the physician became advisers to the case manager, who never gave up control of the case.

Average turnaround dropped to 2 to 5 days. Some applications went out in four hours. Each case manager handled more than twice the volume the company used to process.

Technology helped. The gain came from changing who touches the work and how many times.

Your campaign is an insurance application

Follow one asset through a mid-sized marketing team. Brief comes in. Someone writes. Design picks it up. Comments come back. Brand review. Legal, if there is a price or a promotion in it. The product manager, if it describes a product. Leadership signs off. Every one of those people has their own queue, their own meetings, their own vacation.

AI went after one slice of that path, the writing and the first design pass. The 17 minutes.

Chamath Palihapitiya said it well on September 25 in his deep dive on why AI is booming but productivity is not: “AI can hand back those three minutes, but it can’t decide what happens to them. If they flow into another meeting or another week waiting on legal, the return is zero.”

I have argued before that your AI marketing needs a verifier, and I still think so. Review is necessary. What hurts is five reviews in a row, each with its own wait, when some could run in parallel and one owner could cover most of them.

Time it for two weeks

I do not have a study that says how many days the average marketing asset waits in Central America, and I am not going to make one up. I do know how you can measure it for your own team.

Take the next ten assets your team ships. For each one, write down when it enters and leaves every stage, brief to publication, and how many minutes of hands-on work each person put in. You end up with two numbers per asset, elapsed time and worked time.

The difference is your wait. It will usually pile up in one or two stages.

Now the questions change. Does every asset need this approval, or only the ones that mention pricing? Can brand and product review side by side instead of one after the other? Could one person on the team own each asset end to end, like Hammer’s case manager, and pull legal in only when it matters?

A new license answers none of those.

Where IQ Source comes in

When a marketing team asks us for help with AI, the stopwatch is the first thing we propose. On Sunday I wrote about MIT’s finding that AI without redesign goes with weaker margins, and a marketing workflow is one of the places where redesign shows up fastest, because the stages are few and visible. If AI makes sense after measuring, it goes where the wait is, with agents that prepare the brand or legal review before a person ever opens it. If the problem is structure, we will say so, even if that means you do not need to buy anything.

Send me your ten timed assets and I will point out where your Mutual Benefit Life is hiding.

See how we work with marketing teams