Filed under
Methodology field notes.
All Joust field notes tagged Methodology.
Filed under
All Joust field notes tagged Methodology.
Most AI snake oil starts with something real; the claims just get ahead of it. Ten questions that make an AI product's promise specific enough to evaluate, from a $20 tool to a company-wide system.
Most of the AI cases I am asked to look at are true and useless at the same time. The technology really can do the thing, but that was never the question. One test, run in both directions, separates a business case from a bet on a price you do not control.
The thing that makes your company hard to copy usually lives undocumented in a few people's heads. It is your most valuable asset and your largest single point of failure, and most companies only price it the week someone gives notice.
Most teams do not run on a plan. They run on whoever's inbox is loudest. It is the most expensive operating habit we see, and it stays invisible because the cost is spread across everyone's day.
A walkthrough of a Joust Operating Review. Sector-anonymized findings from a $200M B2B distributor: $1.84M in operations tax mapped across three workflows, with a 12-month implementation roadmap.
Companies that have tried AI usually share a pattern: a high-energy pilot, real engineering effort, and then nothing in production six months later. The model is rarely the bottleneck.
There is a recurring cost inside every operating business that technology could replace. It hides in salaries, not vendor invoices, which is exactly why it never shows up in a procurement review.
There are 8,000 AI tools on the market. Picking one before you understand the work is how you end up with eight subscriptions and zero hours saved.
Capture rate is one of the most misunderstood numbers in AI ROI math. Here's what it is, what it isn't, and how to set it realistically.
The build-vs-buy decision changed when LLMs got cheap. The new question isn't which one to choose. It's which parts of the stack each one belongs to.