Ceremony Adjustments

Every agile ceremony needs adaptation when a team adopts AI-assisted development. The changes are not dramatic -- they are targeted adjustments that surface AI-specific information, address new team dynamics, and ensure that quality and learning keep pace with accelerated delivery. This section provides specific, actionable adjustments for each ceremony, aligned with Pillar 4: Continuous Improvement.

Daily Standup Adjustments

Modified Format

Replace the traditional three-question format with a four-question format during the first 3 months of AI adoption:

Question	Purpose	Time
"What did I complete yesterday?"	Standard progress tracking	30 sec
"What am I working on today?"	Standard planning	30 sec
"Are there any blockers?"	Standard impediment surfacing	30 sec
"Any AI observations?"	Surface tool issues, quality concerns, or learning moments	30 sec

The fourth question explicitly creates space for AI-related observations that might otherwise go unmentioned. After 3 months, this question can be merged into the standard blocker/progress questions as the team develops the habit of surfacing AI-related information naturally.

What to Listen For

As the scrum master, listen for these signals during standup:

Signal	What It Might Mean	Your Action
"I spent a lot of time rewriting AI output"	Poor prompt quality or unsuitable task for AI	Facilitate a prompt improvement conversation; see Prompt Engineering
"I'm still waiting on review for my PR"	Review queue building up; generation outpacing review	Rebalance; assign additional reviewer; see Sprint Adaptation
"The AI kept generating the wrong pattern"	Tool limitation or insufficient context	Log as impediment; may need tool evaluation or team knowledge sharing
"I'm not sure the AI code is correct but tests pass"	Insufficient test quality or review confidence	Facilitate pair review session; engage QA Lead
"I just used AI for everything today, it was great"	Possible over-reliance or insufficient review	Check quality metrics; have a private conversation about review rigor

Standup Anti-Patterns to Address

• AI show-and-tell. Standup devolves into sharing cool AI tricks. Keep it brief; save extended discussions for dedicated sharing sessions.
• AI shaming. Team members who code manually feel pressured to justify not using AI. Protect autonomy.
• AI excusing. "The AI generated a bug" used to avoid accountability. Reinforce that developers own their output regardless of how it was generated.

Sprint Planning Adjustments

Pre-Planning Preparation

Add these steps to your pre-planning routine:

1. AI suitability pre-assessment. Before planning, review the top candidate stories and tag them with estimated AI suitability (High/Medium/Low). This accelerates the planning conversation.
2. Review capacity update. Calculate the team's review capacity (not just development capacity). Ensure the sprint plan does not generate more code than the team can review. See Sprint Adaptation for capacity formulas.
3. Tool status check. Confirm all AI tools are functioning and licenses are active. Tool outages during a sprint cause unexpected impediments.

Planning Techniques

Technique 1: AI Suitability-Weighted Planning Poker

Modify planning poker to include AI suitability:

1. The product manager presents the story
2. Team members silently estimate both the traditional effort AND the AI suitability score
3. Reveal estimates simultaneously
4. Discuss divergences in both dimensions -- disagreement about AI suitability is especially valuable
5. Calculate the adjusted estimate using the technique from Estimation in an AI World

Technique 2: Generation-Validation Split

For larger stories, break the estimate into generation and validation components during planning. This makes the review overhead visible and prevents it from being forgotten.

Example story card:

Story: Create user notification service
Estimate: 8 points total
  - Generation (coding + tests): 3 points
  - Validation (review + security + integration): 5 points
AI Suitability: 4/5

Sprint Goal Framing

When framing the sprint goal, account for AI learning and improvement:

• Delivery goal: "Deliver the order management API with full test coverage"
• Quality goal: "Maintain defect density at or below the sprint 4 baseline"
• Learning goal: "Team completes prompt engineering workshop, module 3" (first 3 months only)

Including a quality goal and learning goal alongside the delivery goal reinforces that speed without quality is not success.

Sprint Demo Adjustments

Demo Structure

When demoing AI-augmented work to stakeholders:

1. Lead with value. Show the working feature, the customer impact, the business outcome. AI is the means, not the message.
2. Show quality evidence. Display test coverage, security scan results, and code review metrics. This builds confidence that AI acceleration is not compromising quality.
3. Share one learning. Include one brief (2-minute) insight about how AI contributed to or challenged the sprint work. This educates stakeholders and builds organizational knowledge.
4. Avoid tool demos. Unless specifically requested, do not demo the AI tool itself. Focus on the product, not the process.

Stakeholder Questions to Prepare For

Question	Suggested Response Framework
"How much of this was AI-generated?"	"AI assisted throughout, but every line was reviewed and validated by our engineers. We measure quality, not origin."
"Can we go even faster now?"	"We are seeing [X%] improvement in throughput. We are also investing in quality assurance to ensure sustainable acceleration." See Velocity & Quality Trade-offs.
"Are we at risk with AI-generated code?"	"We apply enhanced review processes per our standards. Our defect metrics show [trend]. See Quality & Risk Oversight."

Retrospective Adjustments

AI-Focused Retrospective Formats

Run an AI-focused retrospective at least monthly during the first quarter of adoption, then quarterly thereafter.

Format 1: The AI Partnership Review

Divide a whiteboard or digital board into four quadrants:

1. AI Amplified -- Where did AI make us genuinely better?
2. AI Complicated -- Where did AI create more work or confusion?
3. AI Missed -- Where could AI have helped but we did not use it (or could not)?
4. Human Essential -- Where was human judgment clearly the right choice?

This format normalizes both the benefits and limitations of AI tools, creating a balanced team perspective.

Format 2: The Quality Balance

Focus on the quality dimension of AI-assisted development:

1. Each team member shares one quality win (caught an AI issue, wrote better tests, etc.)
2. Each team member shares one quality concern (something that worried them, a near-miss, etc.)
3. Team identifies the top 3 quality improvements to make next sprint
4. Team identifies the top 3 quality practices that are working well and should continue

Format 3: The Confidence Check

Focus on team health and skill development:

1. Anonymous confidence vote: "How confident are you in your AI-assisted development skills?" (1-5)
2. Discuss: "What would increase your confidence by one point?"
3. Identify specific actions: training, pairing, practice exercises
4. Feed results into the Team Health Indicators tracking

Retrospective Topics by Adoption Phase

Phase	Key Retro Topics
Month 1	Tool usability, learning progress, initial quality observations, anxiety levels
Month 2-3	Workflow optimization, prompt effectiveness, review process tuning, estimation accuracy
Month 4-6	Sustainable practices, cross-team sharing, advanced techniques, long-term quality trends
Month 7+	Continuous optimization, new tool evaluation, teaching others, industry benchmarking

info

Share retrospective insights (anonymized) with your Development Manager to inform Team Enablement decisions, and with other scrum masters to accelerate organizational learning.