Product strategy + commercial judgment
B2B products that fit the workflow compete on value. Everything else competes on price.
When I took over a healthcare notification product, the ROI was proven on paper. Providers weren’t using it, because the output landed in an sFTP folder that required IT credentials to access. Nobody had asked whether nurses use sFTP. They don’t.
One conversation with a Chief Nursing Officer revealed the real problem. The data was right. The delivery mechanism was wrong. We built email-style interfaces and status indicators so nurses could divide and conquer their work every shift. End-user organizations grew from 20 to 2,500+. The ROI that was always there finally materialized.
The product hadn’t changed. The workflow fit had.
In B2B, the economic buyer writes the spec sheet and wishes for the value to come. The end user routes around friction to complete the job at the acceptable level. Those are two different people with two different goals. The gap between them is where adoption, and subsequently value, dies.
The buyer feels this late. The response is almost always training or a switch. Both are expensive. Neither fixes the root cause.
Same pattern, a decade later, at AWS. Operational notifications were reaching customers, but the teams responsible for acting on them never saw them. The notifications landed in a dashboard nobody checked regularly, and those teams worked in their own ticketing tools. Routing the data into the tools people already used changed the outcome. Resolution rates improved. Support cases for planned events stopped.
Product innovation and workflow integration are two separate problems. The buyer-user split is what you get when you only solve the first one.
Customer success and account manager spend is what you pay when you get the shape wrong. You’re smoothing things over with a human touch that doesn’t scale.
That’s already shifting. As customers deploy AI agents to run their workflows, the vendor relationship stops being the buffer. Software that doesn’t fit gets bypassed, and the vendor’s value goes with it.
AI agents don’t smooth. They expose. The human buffer disappears. Poor workflow fit won’t get papered over: it will get automated around, permanently.