AI in Customer Support, Product, and Leadership
I used to applaud teams for “taking risks.” Then I watched the same launch fail three times with three different names.
A 150-location quick-service chain I worked with rebuilt loyalty and ordering so it would play nicely with Toast at the counter and DoorDash off-premise.
Leadership asked for bold bets, so the team measured output: new rewards logic, new checkout, new push campaigns for Apple and Google users.
Every miss got filed as “learning” and the roadmap kept moving.
But nothing changed because the assumptions never got challenged. The team was failing fast while protecting the story that the plan was still right.
Meanwhile each “experiment” created real drag: store retraining, comped meals, angry App Store reviews, and more edge cases in payments when we layered in Stripe for gift cards.
And this January, app store compliance can quietly remove your ability to course-correct at all - Apple’s updated age-rating questions have a January 31, 2026 cutoff for submitting updates, and new Apple Accounts in Texas now require age assurance starting January 1, 2026.
Once we started doing weekly, brutal re-assessments (kill the feature, not defend it), we cut the roadmap in half and fixed order pacing first.
Support tickets dropped by roughly a third and the rating climbed without another “big launch.”
It makes me wonder whether most “risk tolerant” cultures are just uncomfortable with the replan.