PreviewYou're previewing CPO AI Launchpad. Sign in to use it.Sign in
← Tools

Signature tool · CPO AI Launchpad

AI feature opportunity canvas

List candidate AI features with the user problem, AI fit, data, quality bar, risk, cost per use and pricing model. The canvas scores each one in code (weights are shown and documented), computes gross margin, ranks them and places them on a quick wins / big bets / maybe later / avoid grid. Example rows are pre-filled and labelled; replace them with your own.

Candidate AI features (4)

Rows marked Example are illustrations to show how the canvas works, not recommendations. Edit or remove them and add your own.

  • Outcome-based
  • Included in plan
  • Per-seat add-on
  • Included in plan

Quick win

2

Big bet

1

Maybe later

1

Avoid for now

0

Portfolio gross margin 26% (revenue 33.65 vs cost 24.86 per active user per month, summed across features).

Ranked portfolio

AI features ranked by priority score
#FeatureQuadrantPriorityValueEaseMarginFlags
1Meeting summaries and action itemsexampleIncluded in plan · cost/user/mo 0.75 · revenue/user/mo 2.00Quick win71.572.572.563%
  • Gross margin 63% is under the 70% target.
2Help-center answers with sourcesexampleOutcome-based · cost/user/mo 0.06 · revenue/user/mo 1.65 · break-even 0.018Quick win66.682.55096%
  • High quality bar: agree an eval set and a pass mark before launch.
3Rewrite in brand toneexampleIncluded in plan · cost/user/mo 0.05 · revenue/user/mo 0Maybe later54.147.592.5n/a
  • No price or attributed value: the cost is real but the margin is unknown. Decide how it pays for itself (plan price, retention, usage limits).
4Autonomous reorder agentexamplePer-seat add-on · cost/user/mo 24.00 · revenue/user/mo 30.00 · break-even 24.00Big bet22.970020%
  • Gross margin 20% is under the 70% target.
  • High harm if wrong: needs a human checkpoint, clear limits and a legal/brand review.
  • The data it needs is not ready (missing, messy or not permitted).
  • High quality bar: agree an eval set and a pass mark before launch.

Portfolio view

AI feature portfolio: value against ease1. Meeting summaries and action items: Quick win (value 72.5, ease 72.5); 2. Help-center answers with sources: Quick win (value 82.5, ease 50); 3. Rewrite in brand tone: Maybe later (value 47.5, ease 92.5); 4. Autonomous reorder agent: Big bet (value 70, ease 0)BIG BETSQUICK WINSAVOIDMAYBE LATER1234
  1. 1Meeting summaries and action items
  2. 2Help-center answers with sources
  3. 3Rewrite in brand tone
  4. 4Autonomous reorder agent
How the scores are calculated
  • Value (0–100) = weighted mean of problem severity, reach and AI fit.
  • Feasibility (0–100) = weighted mean of data readiness, the quality bar (inverted: a lower bar is easier) and AI fit.
  • Risk penalty = 0 at risk 1, rising linearly to max_risk_penalty points at risk 5.
  • Ease = feasibility − risk penalty. It is the x axis of the portfolio grid; value is the y axis.
  • Unit economics per active user per month: cost = cost per use × uses. Revenue = seat price (seat), price × uses (usage), price × uses × success rate (outcome), or the value you attribute (included).
  • Gross margin = (revenue − cost) ÷ revenue. Economics score = 100 at or above target_margin, 0 at or below zero margin, linear in between; no_revenue_score when there is cost but no revenue; free_and_costless_score when there is neither.
  • Priority = weighted mean of value, feasibility and economics, minus the risk penalty (0–100). Ties: higher value first, then name.
  • Quadrants: value ≥ 50 and ease ≥ 50 → quick win; high value, low ease → big bet; low value, high ease → maybe later; both low → avoid for now.
  • Weights: value 0.45, feasibility 0.3, economics 0.25; max risk penalty 30; target margin 70%.
The scores are calculated in code. AI only explains them — check anything important.

Sign in to save your work privately and come back to it. Export works without an account.

The other signature tool →🧩 Product Builder team designer