You are the bottleneck in your practice.

The reason a growing book of clients eats all your time isn't the amount of work. It's that you're still the system every task has to pass through. Better prompts don't fix that. Designing the workflow once does.

The work isn't the problem. You being the system is.

Pulling comparables, rebuilding the same cash flow model from scratch, chasing a document. None of it is hard. It just all waits for you, which is exactly why you can only handle a handful of clients or deals at once. Your capacity isn't set by how good the opportunities are. It's set by how many things you personally have time to touch.

A sharper prompt still runs through you. You have not removed the bottleneck. You have just made it type faster.

Sort. Rank. Design.

01

Sort every recurring task into three piles

Automate (runs without you), Assist (AI drafts, you decide), or Manual (keep it yourself, at least for now).

02

Rank what's in the automate pile

Score by frequency, hours eaten, how well-defined the output is, and how fast you can check the result. Highest totals go first.

03

Describe the workflow once

Name the output, the inputs, when it runs, and the one decision point you want to stay in. That's the whole spec.

04

Start with three

Research (comparables), underwriting (the model), and admin (document chasing) are usually the three that surface first for a growing practice.

What it still gets wrong.

i.

Don't automate the high-judgement calls first

Those belong in the Assist and Manual piles. Automate the jobs that repeat, have a clean output, and are quick to check.

ii.

The hard part isn't the tool

It's working out the formula yourself first, then describing it clearly enough that the AI can repeat it.

iii.

This is the point of the whole library

Every playbook in The Aigent Lab is one job taken off your plate for good, until your capacity is set by judgement, not by how many tasks you personally have time to touch.

Next Playbook