David Ortiz · · 4 minute read
Most failed automations are undocumented processes with software bolted on top. The tool was fine. Nobody could say exactly what it was supposed to do.
Here's the sequence I follow instead, and the questions I'll actually ask you.
1. Look at your week through two lenses
The time lens finds the jobs that are repetitive, slow or easy to get wrong. That's where most people stop.
The income lens asks a different question: what would break first if your leads doubled tomorrow? Whatever you name is the real constraint, and fixing it is worth more than saving a few hours on something that was never in the way.
2. Run the delegation test
Could a new hire follow your written process without asking a single question? If not, it isn't ready to automate.
The fastest way to find out is to hand it to someone and watch where they get stuck. Those sticking points are exactly where an automation would have failed silently.
3. Score it together
Every candidate gets three scores from one to ten: impact, confidence and difficulty. You and I score them together, in the same room. I never score them for you. It takes about twenty minutes, and it means the first project is a decision you made rather than a recommendation handed to you. That matters more in month three than it does on the day.
4. Draw it, then build it
The workflow gets drawn in front of you before anything is built. Every box is a step someone does today, and every arrow is a handoff. If the drawing surprises you, good. That's the conversation we needed to have.
Then it goes live in your own systems, on accounts you own, and I stay to keep it running.
Try this today
Answer the income-lens question on paper: if your leads doubled tomorrow, what breaks first? Then write the process for that one thing as if a new hire had to follow it on Monday. You've just done the first half of an audit.
David
Advisor AI Partners