AI Project Submission Form

Section 1

Submitter and Idea

Identifying information and the one-line pitch. Section 1.6 is the manager sign-off.

1.1 Project name * Working title is fine.

1.2 Submitter and BU * Your name, role, business unit.

Submitter email * We send a copy here and replies go to this address.

1.3 Accountable owner If different from submitter.

1.4 One-line pitch * Fill in: "We will use Claude to ____ so that ____."

1.5 Users and volume * Who uses it, how many people, how often.

1.6 Manager sign-off (required)

I have reviewed this submission. The submitter has time allocated to do this work without compromising their core role. The use case is real. I support sending this to the AI Council.

The manager named below has reviewed and approves this submission. *

Manager name *

Date *

Section 2

Problem and Bet

What is broken, what success looks like, what we will stop doing if it doesn't work.

2.1 Problem today * Hours, dollars, or misses currently being lost. Be specific. "We spend X hours per week on Y" beats "it's slow."

2.2 Prize if solved * Dollars saved, dollars earned, CSAT lift, hires avoided, time back. A number, not an adjective.

2.3 Why now * What changed? Why is this the right time vs. six months ago or six months from now?

2.4 Spend ask (30 / 60 / 90 days) * Seats, tokens, engineering hours, total dollars over each window.

2.5 Kill criteria * Fill in: "If by Day 30 we don't see ____, we stop." Be honest. This protects the submitter as much as the Council.

2.6 First deliverable (due in 2 weeks if approved) * Concrete thing the Council can look at on Day 14. Not a status update.

2.7 Options considered * At least two paths. Cheaper vs. more capable, hosted vs. on-prem, build vs. buy. State the trade-offs. "Only one obvious path" usually means we haven't looked hard enough.

Section 3

Data Sources

What systems we touch, how we get the data in, and how sensitive it is. Section 3.4 controls everything else.

3.1 Source system(s) * Check all that apply.

Zoho Smartsheet Acumatica Other

3.2 Ingestion method * How the data gets into the AI.

REST / GraphQL API MCP server Scheduled file feed (CSV / XLSX) Webhook or event Manual upload Direct DB connector Other

3.3 Refresh cadence * How often the data updates.

3.4 Data sensitivity * What's the most sensitive data this project touches? Check the highest-applicable level. Public — Industry data, public docs, info you'd say at a conference. Proceed to next section. Internal — Internal docs, drafts, code. WMS-owned, non-customer. Proceed to next section. Confidential — Pricing, financials, IP, methodology. Requires CSO approval before submission. Complete fields 3.5 to 3.7 below. Customer data / PII — Customer names, contracts, account info, screenshots, recordings. Requires written HR approval coordinated with the customer executive sponsor before submission. Complete fields 3.5 to 3.7 below.

3.5 Who approved this data being used? * Name and role. Date of approval. Attach the email or written confirmation below.

Attach approval (email or written confirmation) PDF, image, .docx, .msg, .eml, or .txt. Up to ~7 MB total.

3.6 Retention and disposal * When does this data get deleted? How? Default: within 24 hours of session, per the retention schedule.

3.7 Audit trail * How will we prove (later) that data was handled correctly? Logs, screenshots, deletion evidence.