Where AI Helps (and Where It Shouldn’t)
Appropriate Uses
Not Appropriate
You are using AI as a thinking amplifier, not a substitute for collaboration.
DO THIS EXERCISE
Step 1: Select One “Too Big” Story
Example:
“Build a new user dashboard with analytics.”
This is oversized, multi-featured, and vague.
Step 2: Use This Vertical Slice Prompt
Copy and use:
PROMPT TEMPLATE — Vertical Slice Generator
You are an experienced Product Owner and Agile coach.
INPUT
User Story: {paste oversized story}
Constraints: {tech constraints, sprint length, dependencies if known}
TASK
Propose 3 vertical slices that:
-
Deliver user-visible value
-
Can be completed within one sprint
-
Are independently testable
-
Avoid architectural layering splits
For each slice:
-
Provide a short title
-
Explain the user value
-
List 3–5 acceptance criteria
-
Explain why this is a true vertical slice
Keep responses concise and practical.
Step 3: Example Output (For the Dashboard Story)
Slice 1 — “View Basic Metrics Summary”
User Value:
User can see top 3 KPIs on login.
Acceptance Criteria:
-
Displays revenue, active users, churn
-
Data refreshes on page load
-
Handles empty data state
-
Works on desktop layout
Why Vertical:
End-to-end data retrieval, rendering, and validation.
Slice 2 — “Filter Metrics by Date Range”
User Value:
User can view metrics for last 7, 30, or 90 days.
(With criteria…)
Slice 3 — “Export Dashboard Snapshot as PDF”
User Value:
User can share dashboard externally.
(With criteria…)
Step 4: Bring One Slice to the Team
This is critical.
Do not accept AI output as final.
With the team:
The team must own the rewritten story.
Rewrite Template (With the Team)
Once a slice is selected:
Final Story Format
As a {user}
I want {capability}
So that {measurable benefit}
Acceptance Criteria:
-
…
-
…
-
…
Definition of Done Additions:
Why This Works
AI reduces:
The team retains:
-
Ownership
-
Technical judgment
-
Commitment authority
That balance preserves collaboration while increasing throughput.