What you should get from an AI assessment

A $2,500 assessment should tell you what to build first and what would make you walk away. You should be able to spot the recommendation that depends on perfect data or a time saving your team doesn't believe.

Compound's assessment takes 14 days. You get:

• A ranked opportunity map: what to work on first and why.
• A cost model: the assumptions behind the expected return.
• Tool recommendations: which options fit the work.
• A fixed-price scope: the boundaries of the first build.

You keep the plan even if someone else does the implementation.

I want you to be able to read the recommendation and challenge it. If the plan says a task is costing you six hours a week, you should be able to find the volume and time estimates that produced that number. If it recommends a build, you should be able to see why it won over the other options.

From the assessment

What the assessment gives you

The plan shows how the work happens, what each opportunity is worth, and which one to build first.

1. 01

   Workflow X-Ray

2. 02

   Opportunity Map

3. 03

   True-Cost ROI math

4. 04

   First Win recommendation

5. 05

   Build-ready scope

6. 06

   Tool recommendations

We start by walking through the work

You don't need to write a process manual before the assessment. The Workflow X-Ray is a guided working session to get the process out of your head and into a form we can inspect.

Take an incoming invoice, for example. Someone receives it, finds the right customer or job, checks the amount, and enters the information somewhere else. Then check the awkward versions:

• The customer name doesn't match your records.
• The photo is too blurry to read.
• The invoice has already been entered.

Those exceptions are part of the work. They affect what we can automate, what a person still needs to check, and whether the time saving is worth paying for.

Bring a few examples you can share, including an awkward one. Access to the people who do the task matters more than having a polished explanation of it.

The plan gives each opportunity a reason to be there

The Opportunity Map is a 10–15 page plan of the opportunities identified across your business, ranked by payback. It should explain why the first recommendation comes first and what would need to change before the others make sense.

Here's an illustrative excerpt, using made-up workflows to show the decisions a map should support. It isn't a client's assessment.

The completed map uses your records and your working week. It also names the First Win: the automation with the best combination of value and a realistic path to getting it working with your team.

Sometimes that recommendation starts with cleaning up customer records. That's work worth identifying before anyone builds a faster way to move the wrong information.

The cost model makes the assumptions visible

The True-Cost ROI (return on investment) calculation should show:

• Setup costs: the build and the work needed to get it running.
• Running costs: software subscriptions, usage fees, and maintenance.
• Remaining work: what people still need to check or do by hand.
• Expected benefit: expenses you can remove, separately from capacity you can use elsewhere.

Compound scopes and prices builds around a target of 5–10x return within the first year. In the model, that means expected first-year benefit divided by total first-year investment. It's a target to test, not a promised outcome. The pricing guide explains the calculation and how it differs from net return on investment.

Ask to see the result with a lower time saving or slower adoption. You need to know whether the recommendation still makes sense when the team takes longer to use it and more exceptions come back for review.

Tool recommendations belong beside this math. If an existing product handles the work, the plan should say so and include its costs. The buy-or-build guide covers how to compare the options.

A builder should be able to work from the scope

“Automate our paperwork” leaves too much undecided. For a document-to-record workflow, the scope should get down to details like these:

• Inputs: agreed document types from one named inbox or upload location.
• Output: a draft in the chosen destination, with the source document available for review.
• Exceptions: a named person handles unreadable documents, uncertain customer matches, and missing fields.
• Duplicates: flag them before creating a second record.
• Approval: customer-facing actions wait for a person to approve them.
• Acceptance: agree on test examples and expected results before building starts.

This is an illustrative scope, not a description of a purchased engagement. A real scope also names what's excluded, who owns the workflow, and what counts as accepted work. It gives you a fixed price for those boundaries.

The assessment stops at planning and scoping. Implementation is a separate purchase, with the first automation live within 30 days of the build starting. That clock starts with the build.

What happens if the assessment finds no good solution?

If I don't find a durable, reliable solution to the bottleneck you're facing, you get the full $2,500 back. You keep the plan.

If you proceed with Compound and the build starts within 30 days of assessment delivery, the full assessment fee comes off its price. After that, the credit expires.

For a finished example of focused automation, the Newgarden case study shows a check-photo-to-invoice workflow and roughly six hours recovered per week. The sample map and scope above are illustrations; they don't reproduce Newgarden's private documents.

The free discovery call comes before the paid AI Assessment. Bring the process you keep having to chase. We'll start there.