Essay 08

Having the tool is not the project

You paid for the seats. Months later you are not sure anyone uses them, and it feels a little awkward to ask.

The common stall at midsize companies is not “we don’t have AI.” It is “we already bought it.”

Seats are assigned. A vendor ran a kickoff. Someone in ops or IT can log in. And the work still happens in the old spreadsheet, the old inbox, the old Friday export.

That is not a technology failure. It is a project that ended at procurement.

A tool dies in this size of company for a few boring reasons.

The old process was never retired. People got a second place to do the work. Under deadline they use the familiar one.

Nobody owns the new way. IT owns the license. The vendor owns “enablement.” The department owns the outcome. Those are three different jobs, so nobody changes the Friday routine.

Success was defined as “deployed.” Deployed means the login works. It does not mean the weekly report stopped existing, or the coordinator stopped retyping the same email, or the handoff between sales and ops got shorter.

The tool was chosen before the bottleneck. A platform that can do forty things gets dropped on a company that needed one step to get faster. People wander around in the product and go back to email.

If you already have something, do this instead

Name the one job the tool is supposed to take. Write it as a sentence: “This replaces the person rebuilding the open-order list every Monday.” If you cannot write that sentence, the tool is a library card, not a project.

Put a single operator on it, not a steering group. That person already does the work. Their job is to make the old step go away, not to explore capabilities.

Give it a kill date. In 30 or 60 days, either the old step is gone or you stop paying for seats you are not using. A quiet license nobody cancelled is not a win.

Do not add a second tool on top of the unused one. That is how you get a stack and still have the spreadsheet.

If the company already “has AI,” the useful audit is not a feature comparison. It is a list of weekly steps that were supposed to change and did not. Start with one of those. Retire the step, or drop the tool. Sitting in the middle is what kills the project.