Template

AI Opportunity Hypothesis & Scoping Template (Interactive)

An interactive scoping worksheet that captures AI opportunity hypotheses, business value, data readiness, technical approach, safety & ethics checks, prototype plan, and priority scoring — plus a reproducible prioritization formula for quick pipeline decisions.

Interactive Tool

AI Opportunity Hypothesis & Scoping

Use this scoping worksheet to turn an AI idea into a defensible, prototype-ready hypothesis.

Capture the problem, desired customer outcome, estimated value, required data, basic model approach, safety and ethics flags, and a minimal prototype plan. Use the prioritization scales to compare candidates quickly; a suggested weighted scoring formula is provided in the Help for PriorityScore.

Tip: Keep each submission focused on one discrete opportunity (one business process or customer outcome). If an idea has multiple distinct outcomes or data sources, create one submission per candidate so scoring stays meaningful.

A short, specific name for this candidate (e.g., 'Invoice fraud detection - high risk vendors').
Person or role who will sponsor the prototype and champion adoption.
Describe the current problem, who experiences it, and how it is measured today. Be concrete about pain, waste, risk, or missed opportunity.
What change do we expect for customers, users, or internal processes if this works? Tie to measurable outcomes (e.g., reduce processing time by X%, reduce false positives, increase win rate).
A best-effort annual benefit estimate. Enter 0 if unknown and explain assumptions in the next field.
List the assumptions behind the estimate (volumes, unit economics, conversion uplift, error rates, cost savings). Cite data sources where possible.
Include prototype build, data engineering, cloud costs, and expected integration effort. This is a planning-level estimate.
Optional: estimated months to pay back prototype and early rollout costs.
List data tables, files, streams, or APIs needed. Be specific about identifiers, timestamps, and joins. Note any third‑party or sensitive data.
E.g., 'Order database (orders schema) — Data Engineering team / data.owner@example.com'
Check all that apply or use this as a checklist during discovery.
Select the best-fit modeling approach you expect to explore in prototyping.
Note integrations, latency requirements, compute constraints, on‑device needs, or existing infra that helps/hinders feasibility.
Mark any that apply and expand in 'ethics_notes' if needed.
Describe potential harms, stakeholders to consult, necessary mitigations, and regulatory constraints.
Describe process changes, role changes, retraining, or approvals needed for the solution to be adopted.
Describe the smallest useful prototype: inputs, outputs, validation method, success criteria, and what 'done' looks like for a prototype that proves value.
List 3–6 milestones (e.g., data extract ready, baseline metric measured, model trained, offline validation, integration smoke test) with expected dates or weeks and owners.
List concrete metrics and targets (e.g., reduce false positives from 8% to 3%; improve throughput from 20 to 80 claims/day). Include how metrics will be measured and who owns them.
List the three biggest risks that could block prototyping or rollout and possible mitigations.
How substantial is the expected business impact if successful?
1.0 10.0
How feasible is a prototype given current infra and expertise?
1.0 10.0
How ready is the data for modeling (availability, label quality, access)?
1.0 10.0
How much operational change (people/process) is required for adoption? Higher scores mean more change required.
1.0 10.0
Overall risk including safety, legal, privacy, and adversarial concerns. Higher is worse. Use to reduce priority.
1.0 10.0
Optional: enter a calculated priority. Suggested weighted formula: Priority = (Impact*0.4 + Feasibility*0.25 + DataReadiness*0.2 + (10 - Risk)*0.1) — normalize to 0–10. Example: Impact=8, Feas=7, Data=6, Risk=4 => Priority ≈ (8*0.4 + 7*0.25 + 6*0.2 + (6)*0.1)=3.2+1.75+1.2+0.6=6.75.
List immediate actions (data pull, stakeholder meeting, engineering spikes) and an ask for the sponsor (time, budget, access).
You can explore this tool now. Sign in or create an account to save your responses and return to them later.
Sign In to Save Create an Account
Make this tool part of your work

Save a personal copy, bring it to your team, or tailor the questions and workflow to fit what you are hungry to improve.

Member customization and team collaboration are coming soon.
Discuss Ask the Librarian

Discussion

Comments and conversation will live here.