# AI Bandwidth Audit - free audit for solo experts | Colossal

Source: https://colossal.nz/audit/

# How much of your week does your business not need you for?

 The AI Bandwidth Audit is a structured look at where a one to three person business spends its week, and which of those hours can come back. One form, one blueprint, one short call. It is free while it runs as a capped cohort.

 [Start the audit](https://tally.so/r/44RNZO)See what you get

## Built for a business with one person in it.

 Solo experts and micro-consultancies. Coaches, consultants, fractional executives, specialist freelancers. One to three people, no departments, no org chart.

 If you have staff or more than three people, this is the wrong audit and I will tell you so rather than run it anyway. There is a longer version built on stakeholder interviews, and the short one against a real team produces a thin blueprint dressed up as a thorough one.

## How it works.

 1

### You fill in one form

 It asks where your week actually goes, area by area, and what you already use. About 15 minutes, and rough answers are more useful than precise ones you had to go and look up. Do it in one sitting if you can, because a half-finished form is worth nothing to either of us.

 2

### A blueprint lands inside a week

 I write it from that form alone, with no interview round and no discovery call. If something critical is missing I ask exactly one question by message rather than booking time with you. It arrives as a single file that prints cleanly and opens offline.

 3

### Then 30 minutes on a call

 You get the blueprint before the call rather than on it, so you arrive having already read the thing. We spend the time on your questions instead of me reading my own work out loud.

 That is the whole process.

## What the blueprint says.

 One number goes on the cover: the hours a week you spend on work that does not need you, and how many of them you can realistically get back. The second number counts the quick wins only, which makes it smaller and much more likely to be true. A blueprint that promises 20 hours and delivers 4 is worse than useless.

 Under that, at most six opportunities. Three quick wins, two strategic, one larger. A solo operator handed twelve recommendations does none of them, so the list gets cut to what you might actually do this month.

 Effort is scored for someone with no developer. A lot of things that are easy for a company with an engineering team are not easy for you (this is where most generic AI advice falls over), and where that is true the honest answer written down is that it is not worth it yet.

 If your hours are guesses, the blueprint says they are guesses.

 And if there are only three hours in your week worth chasing, it says that too. Your week being already tight is a legitimate finding, and probably a more trustworthy one than a padded list.

## Three ways this ends.

### You do it yourself

 Everything on your list turns out to be a tool you should already be using or a better way to prompt one you have. The blueprint says so plainly and you go and do it.

### You want the whole thing rather than the parts

 The six opportunities are really one problem wearing six hats, and what you actually need is a working AI layer across the business instead of six disconnected automations bolted on at different times.

### One specific build

 A single automation worth quoting for, and we talk about that one.

 The audit ends in a recommendation and sometimes that recommendation is something we build. Naming it here seems better than letting you find it yourself in the first paragraph.

## Why it is free.

 It is running as a capped cohort while the process gets tested on real businesses rather than on mine. You keep the blueprint either way, including the case where your business turns out not to need much doing to it.

 What I ask for in exchange is feedback: three questions on the call about what was confusing in the form, what was most useful in the blueprint, and what was missing or wrong. Then a written review once you have seen the work, never before it.

 When the cohort is full the next one starts a list.

## Questions.

### Do I need to be technical?

 No. The form asks about your week rather than your stack, and every recommendation is scored on the assumption that you have no developer and no appetite for becoming one.

### What if I do not know my hours?

 Estimate. Most people are guessing to some degree, and a rough number you recognise beats a precise one you invented on the spot. The blueprint flags where the estimates are load-bearing.

### How long does the form really take?

 About 15 minutes. It is 13 questions and the longest is a table where you put rough hours against each area of your week.

### Will you try to sell me something?

 The audit ends in a recommendation, and I write it before I know which of the three endings it lands on. Sometimes that is a Colossal build. Sometimes it is that you should go and do this yourself this afternoon.

### What happens to my answers?

 They stay private and they are what I write your blueprint from. If anything from your audit ends up in a public write-up it gets anonymised and you see it first.

### What if I have a team?

 Then this is the wrong track, and I will point you at the full audit, which uses stakeholder interviews and takes four to six weeks.

## Start the audit.

 One form, about 15 minutes. The blueprint comes back inside a week.

 [Start the audit](https://tally.so/r/44RNZO)[See what we build](https://colossal.nz/ai-operating-system/)