AI agents · readiness check
Is your business ready for an AI agent?
Six questions about one task, about five minutes. You get a score out of 18, what it means, and the first step that fits. The check doesn't save or send your answers.
The scoring guide is public. No email needed.
The question
What is an AI readiness check?
An AI readiness check scores how ready one part of your business is for an AI agent. DH Studios scores six things: a task that repeats, data software can reach, steps that are written down, a person who owns the work, a measure of how long it takes today, and rules for what the agent may touch. Each scores 0 to 3, for a total out of 18. The check on this page is you scoring yourself, free. The deeper version is the Drawing: about two weeks inside your operation, where we check the same six things against your real work and draw the agent with a fixed scope and price, for $2,500 counted toward the build.
The check · 6 questions
Score one task, not the whole company.
Pick the one task you'd most like off your team's plate, and answer for that task. Choose the line closest to how it runs today. The number beside each answer is its points.
01 · The task
Is there one task your team repeats every week?
02 · The data
Can the information that task needs be reached by software?
03 · The steps
Could a new hire do the task from what's written down?
04 · The owner
Is there one person who owns this task and could approve the agent's work?
05 · The measure
Do you know how long the task takes today?
06 · The controls
Do you know what the agent may and may not touch?
Your score · 0 of 6 answered
0/ 18
Answer the six questions to see your score.
- The task—
- The data—
- The steps—
- The owner—
- The measure—
- The controls—
Your answers stay in this browser tab. The check doesn't save or send them.
The scoring guide
How the score works.
Six checks, each scored 0 to 3: 0 is not yet, 1 is started, 2 is ready, 3 is strong. The total is out of 18, and the total sets the band.
| Check | What we look for | Why an agent needs it |
|---|---|---|
| The task | One task that comes back on a schedule and follows a pattern. | An agent earns its keep on work that repeats. A one-off isn't worth building for. |
| The data | Where the task's information lives, and whether software can read it. | An agent can only work with what it can read. Paper and memory have to be written down first. |
| The steps | Written steps, including the judgment calls and the reasons for them. | An agent follows rules. If the rules live in one person's head, they have to come out first. |
| The owner | A named person with the time and the know-how to check the work. | A person approves every hand-off. Without an owner, nobody signs, and the agent stalls. |
| The measure | How often it runs, how long each run takes, and who does it. | Today's time is the number the agent is judged against. Without it, nobody can say it worked. |
| The controls | What it may read, what it may send, who approves, and who can switch it off. | Clear limits are what make it safe to let an agent do real work. |
The three bands
0 to 7
Groundwork first
Not ready for an agent yet, and that's worth knowing before you spend anything. Pick the task, name its owner and time it for two weeks. Those steps cost nothing and raise the score fastest.
8 to 13
Ready for one task
Enough to start a first agent on this task. The gaps are normal, and the Drawing is where they get closed: we check the six things on site, then draw the agent with a fixed scope and price.
14 to 18
Ready to build
A strong first agent. The Drawing mostly confirms what you already know, then fixes the scope and the price.
One rule overrides the total: if the task or the owner scores 0, the result is Groundwork first. An agent needs a task that repeats and a person to approve its work.
Example result · not a client
What a result looks like.
An invented company, Example Co., scored the weekly status report its managers build by hand. Every number here is a point on the scoring guide, chosen to show how a result reads.
Example Co. · the weekly status report
10/ 18Ready for one task
Enough to start a first agent on this task. The gaps are normal, and the Drawing is where they get closed: we check the six things on site, then draw the agent with a fixed scope and price.
- The task3 · Strong
Built every Monday, the same way each week.
- The data2 · Ready
The numbers come from two systems that both export to a spreadsheet.
- The steps1 · Started
One manager knows which numbers to flag. It isn't written down.
- The owner2 · Ready
The operations lead would review each report before it goes out.
- The measure1 · Started
Everyone says it takes "most of Monday". Nobody has timed it.
- The controls1 · Started
Nobody has said who may see the report before it's sent.
Raise this first · The steps
Have the person who does the task write the steps once, including the calls they make and why.
Illustrative. Example Co. is invented, and its scores aren't from a real company.
How to read a result
- The big number is the total, out of 18.
- The band beside it says what to do next.
- Each bar is one check, filled from 0 to 3.
- Raise this first names the lowest check and the one step that lifts it.
Two ways to check
The free check here, and the Drawing on site.
| The check on this page | The Drawing | |
|---|---|---|
| Who scores | You. | We do, with the people who run the task. |
| Time | About five minutes. | About two weeks inside your operation. |
| What it rests on | Your best answer to six questions. | Your real work: we watch the task, test the data and map who approves what. |
| What you get | A score out of 18, its band and one next step. | The same six scores with what we saw, and a one-page drawing of the agent with a fixed scope and price, signed before anything is built. |
| Price | Free. | $2,500, counted toward the build. |
Prices as of October 2026.
Questions about the check
What people ask about AI readiness.
What is an AI readiness assessment?
It's a check of whether a business is ready to put AI to work. DH Studios checks one task at a time, on six things: a task that repeats, data software can reach, written steps, an owner, a measure of today's time, and rules for what the agent may touch.
How is the AI readiness score worked out?
Each of the six checks scores 0 to 3, for a total out of 18. 0 to 7 is Groundwork first, 8 to 13 is Ready for one task, and 14 to 18 is Ready to build. If the task or the owner scores 0, the result is Groundwork first, whatever the total.
What does an AI readiness assessment cost?
The check on this page is free. The on-site version is the Drawing: about two weeks inside your operation for $2,500, which counts toward the build if you go ahead.
What score do I need before building an AI agent?
8 or more out of 18 on the task you want to hand over, with no 0 on the task or the owner. Below that, the cheapest next step is the groundwork: pick the task, name its owner and time it.
Does the check work for any kind of business?
Yes. The six checks are about one task, not about your industry, so they read the same for any team that repeats work by hand.
Do you keep my answers?
No. The check runs in your browser and doesn't save or send your answers. If you'd like to talk them through, copy your score into the note when you book a call.
Specified. Built. Signed off.
Start with one workflow.
Book a 30-minute call. We look at one workflow together, and you leave with a yes, a no, or a smaller first step.
The first call is free. Pricing is built around your workflow.