AI
How to get started with AI in your company without rebuilding it
Start with the data you already have, pick one repetitive decision, and measure it before you automate anything else.
JS
Julián Soto14 Jul 2026 · 5 min read
The most common mistake is starting with the technology and looking for somewhere to put it. Start instead with one decision your team makes repeatedly, with a known cost when it is made badly: which lead to call first, which supplier price to accept, which ticket to escalate.
One decision gives you a baseline, a clear success metric, and a scope small enough to finish.
A first project that works
Four constraints keep a first AI project from stalling.
It uses data you already hold, so nothing waits on a new integration.
It has a measurable before and after, agreed on before the build.
A human approves the output for the first weeks, so errors are visible and cheap.
It ships inside an existing tool, so nobody has to change how they work to use it.
/ AI // Business // Adoption /