Audit Compass: what our free AI audit asks

The audit in short

The audit is a free chat with an agent we built called Compass, and it ends with a score for each process you describe. You answer questions about how your work runs today. Compass writes down what you say and rates the processes you name.

There is no call to book and no card to enter. The right person to take it is whoever can describe the work step by step, which is often the person who does it rather than the person who owns it. You can stop partway and come back.

The audit covers one thing: how your work runs now and which part of it is worth building first. It does not connect to your systems, and it does not price a build.

The questions Compass asks

Compass asks about one process at a time, and it asks the same things we would ask sitting across a table from you.

• What starts this work, and what it delivers when it is done.
• What the steps are, in the order they really happen.
• Who does each step, using which data and which systems.
• What outcome you want to be different.
• Where the work repeats, where it needs judgment, where a skilled person spends hours on work that does not need their skill, and how much money or risk rides on it.

The first three questions describe the work as it stands. A description with steps and no trigger is a list of tasks. A description with a trigger and an outcome and no roles cannot be scored, because nobody knows where the know-how sits. A process is mapped when all three have a plain answer in your own words, and Compass keeps asking until they do.

The last two questions decide the ordering. A workflow can only be rated against an outcome somebody has named, so Compass asks what better looks like before it rates anything. Then it listens for four kinds of evidence: the same steps recurring, judgment or branching or many inputs, a skilled person spending hours on work below their skill, and a large cost or gain riding on the result. One process rarely shows all four. Hearing which ones it does show is the rating.

Compass asks these in the same order we use in Discovery, so the free step and the paid step line up and nobody repeats themselves.

What comes back

At the end, Compass hands back the processes you described, in your words, with a rating and a reason for each one.

The reason matters more than the number. A rating with no reason gives you nothing to argue with. A reason tells you which part of the answer moved the score, so you can go and check it with the person who does the work.

Compass also says plainly when a process is a weak first candidate. Work with no clear input and no clear output is weak. Work that is fully deterministic and already automated is weak, because ordinary software is the better tool. Work with no tolerance for an error and no realistic human check is weak as a first project, because you need room to learn on something safer.

Two limits are worth saying before you start. Compass scores what you tell it, so a vague answer earns a vague score. And the audit stops at a ranked list. Turning that list into a build plan is the next step, and it takes a person.

A worked example

Here is what those questions look like with real answers in them, from a build that runs live at Genesco Sports Enterprises.

The trigger and the outcome: month end, with exports out of Sage 50, and the outcome is reconciled job-profitability numbers the finance team trusts. The steps: importing the files, reconciling totals across spreadsheets, and re-keying prior-month profit figures. Every month, the finance team rebuilt job-profitability numbers by hand from Sage 50 exports. The roles, data and systems: the finance team, the Sage exports, and a stack of spreadsheets.

The outcome they wanted was the days back. The evidence was easy to hear. The same steps ran every month. A senior finance person spent that time on importing and re-keying, which is work below that skill. And the close sat in front of every decision that waits on the numbers.

That process scored well for a first build because the inputs were clear, the effort was large, and most of the rules could be written down. The reconciliation took the CFO about 4 days by hand and now runs in about 60 seconds, as an upload, review and download flow. The judgment calls stay with the finance team.

Genesco's CFO put it this way: "Every month our team burned hours rebuilding job-profitability numbers from Sage by hand. Plainpath built us a tool that does the heavy lifting and leaves us just the judgment calls. It has saved us real time and headaches, and they were quick to adjust it when we asked."

Those figures belong to one company and one process. A close where judgment enters at every step compresses far less. We wrote about where we drew the line between the software and the team in keeping judgment calls with the finance team.

What happens next

The audit is free, and nothing happens after it unless you ask for it.

Plenty of people take the ranked list and work on it themselves. That is a fine outcome, and we wrote the long version of how to pick between candidates in which process to automate first.

When you want the list checked by a person, that is Discovery. Someone from our team goes through what Compass heard, digs into your top processes with the people who run them, and writes the plan for building them. The plan names the data to be structured, the decisions to be written as rules, the measures, and the guardrails. We wrote about what a plan has to answer in what is in an AI build plan.

The build is quoted from that plan, so the number comes after the work is scoped.

The plain take

Compass asks what starts the work, what the steps are, who does them with which data, what outcome you want, and where the effort and the repetition sit. It gives back a rated list with a reason for each rating. It is free, it takes one sitting, and the honest input is the whole value of it.

Frequently asked questions

What is Audit Compass?

Compass is the agent that runs our free AI audit. It asks about how one of your processes works today, writes down your answers, and rates the processes you describe so you can see which is the best place to start.

What does the free AI audit ask about?

What starts the work and what it delivers, the steps in the order they happen, who does each step with which data and systems, what outcome you want to be different, and where the work repeats or eats a skilled person's hours.

Do I need to prepare anything before the audit?

No. Bring one process you can describe step by step. The person who does the work daily usually gives better answers than the person who owns it, so send them if you can.

What do I get at the end of the audit?

The processes you described, in your words, with a rating and a reason for each one, including a plain note when something is a weak first candidate. The audit stops at that ranked list.

What happens after the free audit?

Nothing, unless you ask. If you want the list checked by a person and turned into a build plan, that is Discovery. The build is quoted from the plan, so you see the number before you commit.

The plain take

Compass asks what starts the work, what the steps are, who does them with which data and systems, what outcome you want to be different, and where the repetition and the effort sit. It gives back the processes in your words with a rating and a reason for each one, and it says plainly when something is a weak first candidate. It scores what you tell it, so the value of the audit tracks the honesty of the answers. It is free, there is no call and no card, and the next step only happens if you ask for it.