Read Your Whole Team at Once: The Folder Method

Everybody takes the assessment. Everybody gets a PDF with their type on it. Everybody says that's so accurate. Then it goes in a folder and nobody opens it again.

The newsletter version runs the three prompts end to end. This guide goes deeper on the third one, which is the one that actually changes how you manage: reading the whole team at once instead of one person at a time.

Individual assessments tell you about people. A team read tells you about the shape of the group and how you can best work together.

Why the third prompt needs a folder

The first two prompts fit in a normal chat window. You paste one person's results, you paste their job description, you ask your questions.

The team version does not fit. Six people times a full assessment PDF plus a job description is more than you want to paste, and pasting it means re-pasting it every time you want to ask a follow-up.

So you put it in a folder and point the AI at the folder. Then the context is standing rather than something you rebuild every session.

Step 1: Build the folder

One folder. Inside it, one subfolder per person.

team-read/
  alex/
    disc-results.pdf
    strengthsfinder.pdf
    job-description.md
  jordan/
    ...

What goes in each person's folder:

Every assessment result you have for them. The full report, not the one-page summary. The summary is the horoscope. The full report has the behavioral detail.
Their actual job description. Not the job posting you hired from. What they really do now.
Their roles and responsibilities, if you keep that separately.

What does not go in: performance reviews, salary, anything from a private conversation. More on that in the boundaries section, and it matters.

If you have no home for a folder like this, the second-brain guide covers setting one up properly.

Step 2: Point an AI at the folder

You need a tool that reads files off your machine or your drive rather than one you paste into. Either Cowork or Claude Code does this, and both live inside the Claude app, so neither one is a terminal exercise.

Cowork is the easier starting point. If you already work in Claude Code, use that instead. The prompt below is identical either way.

Open it in the folder you just built. It can now read every file in there without you uploading anything.

Step 3: Ask the team question

Here is the prompt, in full.

These are the assessments and job descriptions for my full team. Put together an assessment of us as a team. Where are our skill gaps? Where are we strongest? Where are the friction points, and where could collaboration break down?

Context: we're a fully remote team.

That last line does more work than the rest of the prompt. Add whatever is actually true about how you operate:

Fully remote, hybrid, or in one room
Time zones, if they are spread
Contractors versus employees
Whether you are growing fast, holding steady, or contracting
What you sell, because a team shipping software fails differently than a team running client accounts

Without that context you get a generic personality summary. With it you get something about your actual operation.

Step 4: Push on what comes back

The first answer is the headline. The useful material is one question deeper.

Good follow-ups:

Who on this team is most likely to burn out first, and what would the early signs look like?

If I hired one more person, what should they be, based on what this team does not currently have?

Where are two people likely to duplicate each other's work without noticing?

Which of these gaps is a training problem and which is a hiring problem?

That last distinction is the one worth the whole exercise. Most managers treat every gap as a hiring problem because hiring feels like action.

Step 5: Turn one finding into an operating standard

A finding you do not act on is a slide. Pick one and convert it into something that changes how the team works.

Mine came back highly creative with a strong pull toward perfectionism, flagged as a combination that could make shipping hard. That has not been our problem, and reading it still made me realize I had never told anyone what finished looks like.

So I defined it. What "done" means and what "good enough" means for the category of work where good enough is the correct call.

That is the conversion: a personality observation became a written standard that anyone can point at. Not a workshop, not a slide, one paragraph that settles an argument before it happens.

Do that once and the whole exercise has paid for itself.

How to read it without turning it into a verdict

This is the part that goes wrong, so be deliberate about it.

An assessment describes tendencies under normal conditions. It does not describe capability, potential, or worth. The moment "high in conscientiousness" becomes "she's the detail person, don't give her the strategy work," you have used a helpful tool to build a smaller job for somebody.

Three rules that keep this useful:

Read for what to change about the work, not for what to conclude about the person. The output should generate decisions about scope, process, and support.
Let people see their own read. If you would be uncomfortable showing someone what the AI said about them, you have crossed from managing into assessing.
Treat it as a hypothesis. You are looking for questions worth asking in a one-on-one, not answers that replace the one-on-one.

Where the boundaries are

You are putting personal psychological data about employees into an AI tool. Decide these before you start, not after somebody asks.

Where do the files live? A private folder you control. Not a shared drive the whole company can browse.
What is it never used for? Mine is hiring and firing. This exists to manage people better, and it will never be used to build a case about someone.
Does the team know? Tell them before you run it. "I'm putting our assessments together to look at how we work as a group" is an easy sentence in advance and a bad one to be caught not having said.
What did they consent to? Somebody who took a DISC assessment in an onboarding two years ago did not consent to an AI analysis of their working style. Ask again.

Deleting the folder afterward is also allowed. The finding is what you needed, not the archive.

Common mistakes

Using summaries instead of full reports. The summary is written to feel accurate. The full report has the behavior.
Skipping the context line. Without "we're fully remote," you get a generic team profile that would fit anybody.
Taking the first answer. The headline is rarely the insight. Ask the follow-ups.
Running it and doing nothing. Which is the exact failure the assessments were already sitting in a folder committing.
Adding performance data. The moment reviews go in the folder, this stops being a management aid and becomes a file on someone.