AI Automation Case Study Template (Fill-In Version)

An AI automation case study shows a real client's problem in their numbers, what you built, and the before-and-after result measured against a baseline you recorded at kickoff. Keep it to one page, use only real figures the client approved, and end with what's next.

The case study is the most persuasive sales asset an agency has, and the easiest to get wrong. Most weak ones fail for one reason: nobody measured the starting point, so there's no honest "before".

Plan the case study at kickoff, not at the end

You can't write a credible result without a baseline. Record it before you switch anything on.

Measure the same way both times: same source, same time window, same definition. "Average time between form submission and first reply, taken from CRM timestamps over 30 days" is a measurement. "It felt much faster" isn't.

The seven sections

1. Headline

One line: the client type, what changed and by how much, using the real number. "A plumbing company's web leads now get a reply in under two minutes, down from next-day."

2. The client

Industry, size and location in a sentence. Use the name only with written permission. Otherwise describe them ("a five-person bookkeeping firm in Ohio").

3. The problem, in their numbers

What was happening before and what it cost, using the baseline. Quote the owner if they'll let you.

4. What you built

The automation in plain language, step by step. A simple diagram helps. Name the tools only if the reader cares.

Name the workflow you chose to automate, so readers can match it to their own use cases.

5. Rollout

How long it took, what the client's team had to do, and how you handled the first weeks (for example, staff approving messages before they sent).

6. Results

Before and after, side by side, with the measurement window. Add what the client noticed that you can't measure, attributed to them.

7. What's next

The next automation or change you're working on with them. It shows the relationship continues.

The fill-in template

# [Client type]: [what changed, with the real number]

**Client:** [Name, with permission, or "a [size] [industry] business in [region]"]
**Built:** [Automation name]   **Live since:** [Month Year]

## The problem
Before, [what happened, e.g. "web leads waited until the next working
day for a reply"]. Over [time window] the team [measured baseline,
e.g. "took an average of 19 hours to reply to 84 web leads"].

"[Quote from the owner about the problem, with permission]"

## What we built
1. When [trigger], the system [first action].
2. Within [time], it [second action].
3. [Anything a person still checks or approves.]

## Rollout
Live in [number] weeks. For the first [number] weeks, [who] approved
[what] before it went out.

## Results ([measurement window], measured from [source])

| | Before | After |
|---|---|---|
| [Metric 1] | [baseline] | [result] |
| [Metric 2] | [baseline] | [result] |

"[Quote from the owner about the result, with permission]"

## What's next
[Next automation or improvement.]

Rules for honest case studies

1. Real numbers only. Every figure comes from a measurement you can show the client. No projections presented as results, and no "up to" figures.
2. Written permission. Get approval for the name, the numbers and every quote, in writing. Offer an anonymous version if they'd rather.
3. Show the window. "Over the first 90 days" or "September vs June". Results without a time window can't be compared.
4. Don't overclaim causation. If they also hired a receptionist the same month, say so. Credibility is the point of the document.
5. Keep their data private. No customer names, phone numbers or screenshots with personal data.
6. Label agency results as agency results. If you're using results from an agency you work for or partner with, say whose client it was.

Where to use it

• On your website, one page per case study, written for the niche it belongs to
• In outreach: "Here's what this looked like for another roofing company" with a link
• In proposals, as the proof section (see the AI automation proposal template)
• On discovery calls, when the prospect asks "has this worked for anyone like us?"

A good case study doubles as a business case: a prospect can hand it to whoever signs off on the spend.

No case study yet?

Use a demo instead. Build the automation for a sample business in your niche, record a three-minute video of it running, and say plainly that it's a demo. Then measure your first client's baseline properly, so the demo becomes a real case study within 90 days. How to find clients covers getting that first one.

Learn the sales side

The Stride AI Sales Process course covers the outreach, scripts and offer structure used to close AI deals at Executive Stride, Josh Pocock's AI agency. You can bring a draft case study to the weekly live calls for feedback, or start with the free guide library.

FAQ

How do I write a case study using AI?

Give an AI tool your baseline numbers, results, rollout notes and the client's approved quotes, along with the template above, and ask for a one-page draft. Then check every number against your measurements and send it to the client for approval.

What should an AI automation case study include?

A headline with the real result, a short client description, the problem in the client's numbers, what you built, how it rolled out, before-and-after results with the measurement window, and what's next. Keep it to one page.

How do I get permission to publish a case study?

Ask in principle at kickoff, then send the finished draft for written approval of the name, numbers and quotes. Offer an anonymous version as an alternative. Many clients agree when they know they'll see and approve every word.

What if the results aren't impressive?

Publish the honest version if the client is happy, or keep it as an internal reference. A modest, specific, measured result is more believable than a big vague one. Don't round up, extend the window to find a better number, or leave out caveats.