Case study template¶
Do not publish anything from this template until every box in the consent checklist is ticked. This file contains placeholders only. It holds no real or example results; if you add a sample for illustration, mark it "Example" in the heading and in the text, and do not present it as a customer's outcome.
Consent checklist¶
Get each permission in writing, from a named person with authority at the customer, before drafting.
- [ ] Permission to publish the company name. Named by:
[NAME, ROLE], date[DATE] - [ ] Permission to publish the numbers below, each one individually reviewed by the customer. Named by:
[NAME, ROLE], date[DATE] - [ ] Permission to publish the quote, with the exact wording approved. Named by:
[NAME, ROLE], date[DATE] - [ ] Permission to name the person quoted (or agreement to attribute to the role only).
[NAME / ROLE ONLY] - [ ] The customer has seen the final text and the pilot report it is drawn from.
- [ ] No end customer's personal data appears (names, numbers, message text).
- [ ] The customer knows they can withdraw consent; we will remove the page within
[N]days of a request.
If any box is unticked: stop. A case study with a missing permission is not published.
Title¶
[COMPANY NAME]: [ONE JOB, for example "recovering missed calls"] over 30 days
The situation¶
[Two or three sentences, in the customer's own terms: what the job was and why it mattered. Use only what the customer told you.]
What we set up¶
- Job:
[ONE OF: missed-call recovery / appointment reminders / payment follow-up / lead follow-up] - Channel(s):
[CHANNELS] - What the agent did and did not do:
[COPY FROM THE SIGNED PILOT AGREEMENT] - Approval rules:
[COPY FROM THE PILOT AGREEMENT] - Period:
[START DATE]to[END DATE]
What was measured¶
Every figure below is copied from the pilot report generated for this agent and period. Attach the PDF.
| Measure | Before the pilot (customer-entered) | During the pilot (recorded) |
|---|---|---|
| Monthly volume | [BASELINE] |
[CONVERSATIONS] in [DAYS] days |
| Resolved without a person | not measured | [N] ([X]%) |
| Handed to a person | not measured | [N] ([X]%) |
| Corrected by a person | not measured | [N] ([X]%) |
| Estimated model cost | not applicable | [AMOUNT] ([at least] if the report says the cost is a lower bound) |
| ROI, based on the customer's assumptions | not applicable | [PERCENT] using [VALUE PER RESOLUTION] per resolved conversation |
State the assumptions beside the ROI, in the same sentence or the same table: "ROI is based on the customer's own
estimate of [VALUE] per resolved conversation." Do not publish an ROI without them.
In the customer's words¶
[EXACT, APPROVED QUOTE]
[NAME, ROLE, COMPANY](only as approved in the checklist)
What did not work¶
[At least one thing, from the pilot notes. A case study with no problems is not credible and probably not complete.]
What this does not show¶
- It shows what was recorded while the agent ran, not that the agent caused the change.
- It covers one customer, one job and one month.
[ANY LIMIT THE REPORT STATED, for example conversations outside the connected channels]
Sign-off¶
| Name | Date | |
|---|---|---|
| Customer approved the final text | [ ] |
[ ] |
| We checked every number against the pilot report | [ ] |
[ ] |