Skip to main content
Resource · Framework

AI-Augmented Scrum

The Scrum framework enhanced with AI agents at every touchpoint. Agents augment human roles, while Notion AI Meeting Notes turns every transcript into intelligent summaries with actions, risks, and conflicts surfaced automatically.

Who it's for

Built for Scrum teams that already ship

Product and delivery teams running Scrum on Notion plus Jira or Linear, with 1-2 week sprints, that want AI augmentation without changing tools, ceremonies, or roles. Agents sit alongside the team - they don't replace the Product Owner, the Scrum Master, or the engineers.

NOTION + JIRA / LINEAR·1-2 WEEK SPRINTS
The framework

Four events, three agents, one meeting AI

The Sprint Cycle stays the same. What changes is what happens inside each event - Notion AI Meeting Notes processes transcriptions in real time, classifies what it hears, and routes risks, conflicts, and action items to the right custom agent.

Sprint Planning
Daily
Sprint Review
Sprint Retrospective
Each event has a dedicated Notion AI instruction. Risks, conflicts, and action items are classified live and routed to the right agent.
Agents

Three custom agents augment human roles

Each agent owns a narrow set of skills tied to a Scrum role. They run inside Notion, reach into Jira and Linear over MCP, and only act when Notion AI hands off a classified signal - so the human stays the decision-maker.

01

Product Owner Agent

Role

Reviews goal quality against SMART criteria and story readiness against INVEST. Drafts product and sprint goals from raw context, and flags stories that aren't ready before they hit Sprint Planning.

Skills
  • review_product_goals
  • review_sprint_goals
  • check_sprint_stories_readiness
  • write_sprint_goals
02

Scrum Master Agent

Role

Generates sprint status and reports straight from Jira, Notion, and Linear over MCP. Reviews user stories for clarity, and seeds retrospectives with patterns it has noticed across the sprint.

Skills
  • generate_sprint_status
  • draft_sprint_report
  • review_user_story
  • generate_retro_ideas
03

Engineer Agent

Role

Enforces quality gates on every increment. Runs security and code-quality checks against the codebase, monitors build health, and translates acceptance criteria into Gherkin scenarios.

Skills
  • security_codebase_review
  • build_health_check
  • code_quality_check
  • write_gherkin_scenarios
Meeting AI

One instruction per Scrum event

Notion AI Meeting Notes runs a different instruction for each event. The instruction sets the lens, the classification rules, and the escalation path - so Daily output looks different from Retrospective output, but both feed the same agents.

  • Sprint Planningsprint_planning_instruction

    Captures the sprint goal, scoped stories, and assumptions. Flags stories that fail INVEST and routes them back to the Product Owner Agent.

  • Dailydaily_instruction

    Listens for blockers, slipped commitments, and silent dependencies. Posts a daily digest and triggers the Scrum Master Agent when a risk is classified.

  • Sprint Reviewsprint_review_instruction

    Summarises what shipped, what didn't, and stakeholder feedback. Flags Definition-of-Done gaps and routes them to the Engineer Agent.

  • Sprint Retrospectivesprint_retrospective_instruction

    Clusters recurring themes from the transcript into keep / drop / try, and pre-populates next sprint's improvement actions with owners.

  • Backlog Refinementbacklog_refinement_instruction

    Captures items refined, breakdowns, sizing, and open questions. Routes new estimates and acceptance-criteria changes to the Product Owner Agent for INVEST checks.

What it detects
RisksConflictsAction items

Notion meeting instruction templates

Each instruction tells Notion AI what shape the meeting summary should take. Below are the templates we use - copy them as-is, or tune the headings to match your team's vocabulary.

Sprint Planning
sprint_planning_instruction
# Sprint Planning Summary
## Sprint [name] Overview
- Sprint Goal agreed upon by the Scrum Team
- Business value or user outcome this Sprint targets
- Sprint Backlog Items (Key Stories, Work Items mentioned)
- Capacity considerations or constraints discussed
- Dependencies identified
## Risks & Assumptions
- Key risks, unknowns, or assumptions flagged
## Decisions Made
- All decisions taken during the session
## Action Items
- Task, owner, and due date for any follow-ups
Daily
daily_instruction
# Daily Scrum Summary
## Progress Toward Sprint Goal
- Key updates shared by each team member
- What was accomplished since the last Daily Scrum
- Catch Jira issue IDs or any other tasks identification
## Plan for the Next 24 Hours
- What each team member plans to work on today
- Any re-ordering or changes to the Sprint Backlog
## Impediments & Blockers
- Issues raised that slow progress toward the Sprint Goal
- Who is affected and suggested next steps
- Any new dependencies mentioned during the call
## Decisions Made
- Any quick decisions taken during the sync
## Follow-Up & Parking Lot Items
- Conversations to continue after the Daily Scrum
- Who needs to collaborate on what
Sprint Review
sprint_review_instruction
# Sprint Review Summary
## Sprint Goal Outcome
- Was the Sprint Goal met? (Yes / Partially / No)
- Brief explanation
## Increment Demonstrated
- Features or Product Backlog items demonstrated as Done
- Items not completed and why
## Stakeholder Feedback
- Key feedback, reactions, and suggestions from stakeholders
- New ideas or feature requests raised
## Product Backlog Updates
- Changes to priorities or scope based on feedback
- New items to add to the Product Backlog
## Progress Toward Product Goal
- Current status and trajectory
- Market, budget, or timeline updates discussed
## Decisions Made
- Agreements reached during the review
## Action Items
- Task, owner, and due date for follow-ups
Sprint Retrospective
sprint_retrospective_instruction
# Sprint Retrospective Summary
## What Went Well
- Practices, behaviors, or outcomes the team wants to continue
## What Didn't Go Well
- Challenges, pain points, or problems encountered
- Root causes identified
## Improvement Actions
- Most impactful improvements agreed upon
- Owner and target date for each improvement
- Items to add to the next Sprint Backlog (if applicable)
## Team Dynamics & Collaboration
- Observations about communication, trust, and self-management
## Process & Tools
- Feedback on tools, workflows, or ceremonies
- Definition of Done - any proposed changes
## Decisions Made
- Agreements and commitments for the next Sprint
## Appreciation & Kudos
- Team members recognized for contributions
Backlog Refinement
backlog_refinement_instruction
# Backlog Refinement Summary
## Items Refined
- List of Product Backlog items discussed (classify Jira issue ids or others identifications)
- For each item: updated description, acceptance criteria, and sizing
## Items Broken Down
- Large items split into smaller, deliverable pieces
- Rationale for the split
## Estimation & Sizing
- Estimates assigned (story points, T-shirt sizes, or other)
- Items with high uncertainty or disagreement flagged
## Clarifications & Open Questions
- Questions raised and answers provided by the Product Owner or Developers
- Remaining open questions and who will resolve them
## Dependencies Identified
- Cross-team or external dependencies discovered
## Priority or Ordering Changes
- Any re-ordering of the Product Backlog discussed
## Decisions Made
- Agreements on scope, acceptance criteria, or approach
## Action Items
- Task, owner, and due date for follow-ups