Canonical: https://alcub3.com/resources/ai-project-decision-workbook
Updated: 2026-10-24

#
AI Project
Decision Workbook
Choose one workflow. Find the unknowns. Design a useful next experiment.
Get the free 11-page PDF, three editable spreadsheet worksheets and a one-page experiment template. Version 1 · October 2026.
Email
What are you exploring? (optional)
Choose if useful
An AI or automation workflow
AI agents and tools
A team decision
Just exploring
Leave this empty
We use your email to deliver this resource. This does not subscribe you to marketing emails or book a consultation.
[Privacy details](https://alcub3.com/privacy)
.
Get the free workbook
##
Your workbook is ready.
Choose the PDF, or download the complete package with editable CSV worksheets and a text template. These links work for seven days.
Download PDF
Download PDF + worksheets (ZIP)
##
A practical path from question to experiment.
Describe the current work, define a useful result, check the inputs and compare realistic routes before choosing what to build.
Name the decision
Describe the current work
Define a good result and a bad one
Check inputs and permissions
Compare three practical routes
Worked example: drafting service replies
Worked example: a field inspection assistant
Write a one-page experiment brief
Read the result honestly
Bring a useful question to a consultation
##
Read a complete worked example.
Fictional illustration; no customer result or measured ALCUB3 performance is claimed.
A small service business receives recurring questions about appointment preparation. Its coordinator believes drafting is slow, but no timing study exists. Several versions of the preparation instructions are circulating.
The decision is whether a draft assistant would help the coordinator answer from one approved instruction sheet. It is not whether the business should automate its entire inbox.
The first useful action is to settle the source document and observe the current process. The team then prepares a small set of synthetic questions, including a missing detail, a conflicting request and a request outside the service's scope.
Three routes:
1. A maintained template library may solve the recurring questions with little setup.
2. An AI drafting experiment may help with varied wording, if it preserves the approved facts and makes exceptions visible.
3. Deferring the prototype is reasonable if nobody owns the source instructions or review.
The proposed test compares templates and AI-assisted drafts against the same examples. A reviewer checks accuracy and records total preparation plus correction time. The prototype cannot send mail, access the real inbox or change appointments. Any invented instruction or attempted external action triggers review before testing continues.
A successful test would support another limited step. It would not establish production reliability, customer satisfaction or a business-wide return on investment.
##
Bring a useful question to a conversation.
You can request a free 30-minute consultation at any point. The workbook is optional; bring one workflow and the decision you need to make.
[Request a free consultation →](https://alcub3.com/consultation)