Curs Augmented Agility: AI for Agile Teams

Module 1 – Foundations – AI Technology & the Agile Mindset

AI projects challenge conventional project management assumptions. Unlike deterministic software, AI solutions are inherently probabilistic: they produce results in ranges of likelihood, not guaranteed correct outputs. Understanding these differences is essential for project managers.

What Agile Teams Need to Know About AI

• How large language models (LLM) work
• Overview and comparison of free AI tools: Claude vs ChatGPT vs Gemini vs Copilot
• Strengths and weaknesses of each tool for Agile ceremony and artifact contexts
• AI limitations every Agile team must understand: hallucinations, missing context, bias, confidentiality risks

Introduction to AI Prompt Engineering for Agile Teams

• Basics of an effective Agile prompt.
• AI Prompting patterns for Agile
• Converting Agile project inputs into effective AI prompts
• The ‘AI as Junior Colleague’ mental model

Exercise: Tool Comparison Challenge — Same Agile task run across 3 AI tools. Teams compare outputs, evaluate quality, and identify best use cases for each tool. Teams builds a shared ‘AI Blind Spots’ list.

Module 2 – Product Envisioning using Artificial Intelligence (AI)

Learn how to use AI to accelerate early-stage product discovery — generating Vision Documents, framing the product opportunity, and defining initial scope — while keeping strategic decisions with the Product Owner and stakeholders.

2.1 Agile Project Envisioning — AI-Assisted

• What product envisioning means in Agile: defining the ‘why’ before the ‘what’
• Using AI to synthesize stakeholder inputs, market context, and product goals into a structured vision
• Alignment with the Agile Extension to BABOK®: Business Need, Required Capabilities, and Solution Scope

2.2 Vision Document — AI-Generated Draft

• Why you need a Vision for your Agile project. Vision Document structure
• AI prompts to generate a Vision Document
• What to look for in the AI generated Vision Document: context, business knowledge, and strategic intent

2.3 Initial Scope and Product Goals — AI Prompts

• Using AI prompts to obtain initial product capability and high-level requirements from product vision
• Aligning product goals to measurable outcomes: determine relevant project OKRs / KPIs with AI assistance
• Exercise: Vision Workshop — Each team receives a one-paragraph product brief. Use AI to generate a Vision Document. Team Discussion: what did AI get right? What strategic and market context is missing? Teams present their before/after vision refinement.

Module 3 – Stakeholder Analysis & Obtaining Personas with AI

Learn how to use AI to identify and analyze stakeholders and to build user Personas — the foundation for all User Stories and requirements work that follows.

3.1 Stakeholder Analysis — AI-Assisted

• • Using AI prompts to identify and generate a stakeholder list from project context:
  • AI-generated Stakeholder analysis: Power/Interest Grid, Influence/Impact Matrix
  • How the project teams should check the quality of AI generated deliverable
  • Stakeholder engagement strategy: AI-drafted communication and involvement plans

3.2 Persona Development — AI Prompts

• Agile Persona concept. Content of a Persona template. Why should you use Persona concept?
• Using AI to generate detailed Persona descriptions from stakeholder analysis and product context
• Validating AI-generated Personas. What to look for as a team
• Linking Personas to User Stories.

Exercise: Stakeholder & Persona — Using the product brief, teams use AI to generate a Stakeholder Analysis with Power/Interest grid, then build 2 detailed Personas. Team discussion: which AI assumptions about stakeholders need team verification?

Module 4 – User Stories & Acceptance Criteria with AI

Learn how to use AI to generate, review, and refine User Stories and Acceptance Criteria — ensuring every Persona has a User Story, acceptance criteria, and priority before entering the backlog.

4.1 User Story Creation — AI Prompts

• User Story template: As a [persona] / I want [capability] / So that [outcome] — ensuring AI output is outcome-focused
• Generating multiple User Story variants from a feature description targeting different personas
• Using Personas as AI context to generate richer, more specific stories
• Validating User Story generated by AI. Using Story quality criteria: INVEST (Independent, Negotiable, Valuable, Estimable, Small, Testable)

4.2 Story Structure — User Story + Acceptance Criteria + Priority

• Using AI prompts to generate the mandatory components for every User Story:
  • User Story title
  • User Story description: As a [persona] / I want [capability] / So that [outcome]
  • Acceptance Criteria: as a User Acceptance Test (UAT) scenario or as a Gherkin language acceptance criteria (Given/When/Then)
  • Priority Suggestion: business value rationale, time criticality, risk reduction factor, and recommended MoSCoW or other priority position (WSJF)
• Comparing AI-generated user Story + Acceptance Criteria + priority against team judgment: identifying what AI cannot know about your context

4.3 Acceptance Criteria Quality — AI generated output Review

• Detecting and fixing ambiguous language in Given/When/Then (Gherkin) scenarios
• Comparing AI-generated with team-written Acceptance Criteria
• Alignment with Agile Extension to BABOK®: Acceptance and Evaluation Criteria (10.1) and User Stories (10.48)

4.4 Story Quality Critique — AI-Assisted INVEST Review

• AI INVEST scoring (Independent, Negotiable, Valuable, Estimable, Small, Testable)
• Detecting low quality User Stories: large, technical, no persona, solution-prescriptive
• Comparing AI output with team review to identify context AI misses

Exercise: User Story Development — Part A: Each team brainstorms and generate 3 User Story variants from a Persona + feature brief. For each story, use AI prompts to generate the User Story + Acceptance Criteria + priority. Part B: Teams review the AI generated output— what was right, what was wrong, what was missing?

Module 5 – Epics, Story Decomposition & Backlog Management with AI

Learn how to use AI to decompose epics into sprint-sized stories and to build, maintain, and prioritize the product backlog — while keeping product ownership with the team.

5.1 Epics & Story Decomposition — AI-Assisted decomposition with SPIDR Framework

• What is a good Epic. How to write good Epics
• Alignment with best practices from Agile Extension to BABOK® standard
• How to obtain sprint-sized stories by applying Epic decomposition technique SPIDR (Spikes, Paths, Interfaces, Data, Rules)
• Ordering user stories slices after Epic decomposition
• Validating decomposed stories: check obtained User Stories to be valuable and DoR compliant

5.2 Backlog Development — AI-Assisted

• Extracting candidate stories from stakeholder emails, meeting notes, and requirement documents
• Using AI to detect low quality generated user Stories: INVEST violations, duplicates, missing acceptance criteria, and ambiguous language
• Maintaining backlog quality applying best practices

4.3 Backlog Prioritization — with Artificial Intelligence AI Prompts

• Alignment with Agile Extension to BABOK®: Prioritisation techniques and Backlog Management
• Prioritization frameworks with AI: MoSCoW, WSJF (Weighted Shortest Job First), Value vs Effort matrix
• AI-assisted priority WSJF scoring: User-Business Value, Time Criticality, Risk Reduction, Job Size
• How to monitor and change AI prioritization. Factors that you should use.

Exercise: Backlog Grooming — Teams receive a ‘broken backlog’ with 8 issues (INVEST violations, duplicates, missing acceptance criteria, vague language, mismatched priorities). Step 1: manual review. Step 2: AI-assisted review. Team discussions: what did AI catch that humans missed, and what did humans catch that AI missed?

Module 6 – Story Mapping, Release Planning & Product Roadmap with AI

Learn how to use AI to build story maps, draft product roadmaps, and obtaion realistic release plans — connecting individual stories to the user journey and the business timeline.

6.1 User Story Mapping — AI-Assisted

• Alignment with Agile Extension to BABOK®: Story Mapping technique and Scope Modelling
• What a Story Map is and how to use it
• Using AI to identify the key requirements from a customer journey map or backlog
• Organizing backlog items into story maps
• Identifying candidate releases: AI-assisted definition of Minimum Viable Product (MVP) and next releases

6.2 Release Planning — AI-Assisted

• Ai prompts to create a release plan from a story map, team velocity, and target release date
• AI-generated release plan structure best practices
• Where AI release planning fails: team morale, organizational constraints, strategy, and sponsor decisions
• Validating the AI generated Release plan with team capacity and external dependencies

6.3 Product Roadmap — AI-Generated Draft

• Product Roadmap structure best practices
• Using AI to generate a product roadmap from the product vision and story map
• Communicating the roadmap to different stakeholder personas: AI-adapted executive summary vs. team detail view

Exercise: Story Map & Roadmap— Using a given backlog, team build a Story Map identifying the backbone and 3 releases. Then use AI to generate a project roadmap. Each team presents one roadmap AI decision they would override and why.

Module 7 – Definition of Ready (DrR) & Definition of Done (DoD) with AI

Learn how to use AI prompts to generate, good DoR and DoD definitions — the quality gates that protect sprint integrity.

7.1 Definition of Ready (DoR) — AI Gap Analysis

• DoR purpose: the User Story quality checks before it enters Sprint Planning Meeting
• Running a DoR review with Ai prompts
• Generating context-specific DoR items based on team profile, domain, and releases

7.2 Definition of Done (DoD) — AI-Generated Templates

• DoD purpose: the quality checks an Increment must pass before it is declared Done in the Sprint Review
• AI prompts to generate DoD list tailored to your project requirements and quality standards
• What to look for when checking the quality of AI generated DoD
• Sprint-level DoD vs. Release-level DoD: distinguishing the two with AI-assisted prompts

Exercise: DoR & DoD Refinement — Each team audits a provided DoR and DoD using AI gap analysis prompts. Identify 3 improvements to each. Teams then validate the result: could every item on the DoD list be objectively verified by a different team member?

Module 8 – Sprint Planning with Artificial Intelligence AI

Learn how to use AI to accelerate Sprint Planning and prepare, structure the Sprint Backlog — while the team retains full ownership of every commitment made.

8.1 Sprint Goal Generation — AI Prompts

• Alignment with Scrum Guide: Sprint Goal and Sprint backlog best practices
• identifying candidate Sprint Goals from the top backlog items and stated product objective
• Ensuring Sprint Goals are outcome-focused, testable, and achievable within one sprint
• AI-generated options as team conversation starters: the team validates and owns the final goal

8.2 Sprint Planning Preparation — AI-Assisted Checklists

• AI-generated preparation checklist: backlog items readiness, team capacity
• Capacity-adjusted sprint backlog with AI prompts
• Identifying story dependencies and risks before the Sprint Planning Meeting

8.3 Sprint Backlog — AI-Assisted Decomposition into Tasks

• Using AI to decompose selected User Stories into implementation tasks for the Sprint Backlog
• Task estimation: AI-suggested time estimates for discussion, not as final commitments
• Validating the Sprint Backlog against the Sprint Goal: does the selected work deliver the goal?

Exercise: Sprint Planning Simulation with AI — Teams use a pre-built backlog to generate 3 Sprint Goal options (team votes on one), select stories for the sprint, and decompose them into tasks. Each squad identifies at least one AI suggestion they override with justification.

Module 9 – Estimation & Planning Poker – AI Assisted

Learn how to use AI as an estimation input — never as an estimation decision-maker.

9.1 User Story Analysis — AI Prompts

• AI prompts to analyze complexity factors: technical uncertainty, integration risk, business logic complexity, team familiarity
• AI-suggested story point range on the Fibonacci scale with reasoning
• Team validation of AI estimates. Basis for team discussion, not commitments to accept

8.2 Clarifying User Stories before Planning Poker — AI Prompts

• Generating clarifying questions to resolve unknowns before Planning Poker estimation
• AI identification of stories with potential high estimation variance — and the reasons why
• Using AI-generated questions to shorten planning sessions without losing estimation quality

8.3 AI Estimation Risk — The Anti-Pattern

• Why revealing AI estimates before Planning Poker eliminates genuine team estimation
• Blind-first estimation rule
• How to use productively the divergence between AI and team estimates

Exercise: AI-assited Planning Poker — Blind team Planning Poker on 5 stories. AI estimates generated. Team discuss point gap between AI and Team estimates: why the divergence? What context does AI lack that the team knows?

Module 10 – Daily Standups & AI-Assisted Team Communication

Identifying appropriate and inappropriate AI technology use in Daily Scrum meetings

10.1 AI in the Daily Scrum — What Belongs and What Does Not

• Scrum Guide best practice: Daily Scrum meeting – synchronization and impediment identification, not status reporting
• Appropriate AI use for Daily Scrum meeting
• AI misuse pattern: AI-generated standup summaries and why this fails

10.2 Impediment Detection — AI analysis of Daily Scrum meetings

• Ai prompts to analyze Daily Scrum notes to identify recurring blockers and stalled stories
• Identifying systemic impediments: process gaps, tooling failures, dependency bottlenecks
• AI-generated escalation: structured impediment messages for Product Owner or leadership

10.3 Sprint Progress Monitoring — AI-Assisted Early Warning

• Comparing sprint progress to sprint goal and flagging early warning signals
• AI identification of at-risk stories
• What the Daily Scrum offers that no AI summary can replace: relationship, trust, and real-time human sensing

Exercise: Daily Scrum Audit — Teams analyze a transcript of 5 problematic daily Scrum meetings using AI pattern analysis. Team Discussion and learning opportunities: what would have been lost if the team had read an AI summary instead of meeting? Each participant commits to one Daily Scrum meeting behavior they will NOT automate.

Module 11 – Sprint Review with Artificial Intelligence AI Prompts

Learn how to use AI to prepare, structure, and follow up on the Sprint Review — the Scrum event where the team inspects the increment and adapts the product backlog with stakeholder input.

11.1 Sprint Review Preparation — AI Prompts

• Scrum Guide best practice for Sprint Review
• Using AI to generate a Sprint Review agenda tailored to the sprint goal and key outcomes
• AI-generated increment summary: what was Done / Not Done and value delivered
• AI prompts to adapt the increment Sprint Review presentation for different stakeholders

11.2 Sprint Review Facilitation — AI-Assisted

• Generating stakeholder discussion questions to elicit actionable feedback during the review
• Using AI to capture, structure, and prioritize stakeholder feedback in real time
• Translating stakeholder feedback into candidate backlog items / User Stories
• What AI cannot replace in Sprint Review: conversation, relationship and human judgment

11.3 Post-Review Backlog Update — AI-Assisted

• Using AI to update Backlog based on structured stakeholder feedback from the review
• Prompts to identify changed priorities, new User stories, or change in product strategy
• Ensuring the backlog reflects the Product Owner’s understanding — AI assists, PO decides

Exercise: Sprint Review with AI — Teams receive a fictional sprint outcome (stories completed, bugs found, demo feedback). Use AI to prepare a Sprint Review agenda, an increment summary, and a stakeholder-adapted presentation narrative. Teams present and discuss what AI missed about the stakeholder dynamics.

Module 12 – Retrospectives & Continuous Improvement with AI

Learn how to use AI for retrospective and pattern detection — while actively protecting psychological safety of the team.

12.1 Sprint Retrospective Pattern Analysis — AI Prompts

• Using AI prompts to synthesize past retrospective data to understand systemic patterns
• Ai prompts to identify recurring blockers, improvements, and systemic root causes
• AI vs. human pattern recognition: what the data reveals vs. what team observes

12.2 Custom Retrospective Format Design — AI Prompts

• Generating a tailored retrospective format based on team challenges, energy level, and sprint context
• AI-designed facilitation guides: activity names, timing, facilitator instructions, and debrief questions
• Format options: Start/Stop/Continue, 4Ls, Mad-Sad-Glad, Sailboat— AI recommendation vs. Scrum Master judgment

12.3 Continuous Improvement Tracking — AI-Assisted

• Tracking retrospective action item completion rates and impact over multiple sprints using AI analysis
• AI continuous improvement assessment: did it produce improvement, or did it fail?
• Psychological safety commitment for effective Retrospective meetings

Exercise: Retrospective Pattern— Teams get 5 sprints of retrospective data with patterns. Part 1: Team analysis — teams identify themes manually. Part 2: AI analysis — compare findings. Part 3: design a targeted retrospective to address the most critical systemic theme found.

Module 13 – AI Risks and Anti-Patterns

Learn how to use AI to recognize risks and avoid the seven AI anti-patterns most common in Agile settings.

13.1 The Four AI Risks in Agile Contexts

• Hallucination: AI generating plausible but wrong Agile artifacts — detection and prevention techniques
• Bias amplification: AI inheriting biases embedded in prompts or in its training data
• Privacy erosion: pasting confidential data — PII, financial data, IP — into public AI tools
• Over-reliance: skill atrophy, reduced collaboration, and teams that cannot function without AI

13.2 The Seven AI Anti-Patterns — Recognizing and Avoidance

• AI Oracle: treating AI output as authoritative without team ratification
• Backlog Dump: AI-generated backlogs without Product Owner review for intent and context
• Silent Standup: replacing the Daily Scrum with AI-generated summaries
• Retrospective Surveillance State: inappropriately using retrospective information with AI, and destroying psychological safety
• Estimation Anchor Lock: eliminating genuine team estimation
• Credential Leak: using financial data, or proprietary IP into public AI tools
• Skill Atrophy Spiral: AI replacing human practice until the team cannot function without it

Exercise: AI Risks — Each team is assigned a different AI use case (ranging from low to high risk). Identify how it could go wrong. Design a specific, practical mitigation for each risk. Present and defend the safeguard against challenges from other teams.

Module 14 – Course Review & Continued Development

14.1 Key Takeaways & Best Practices

• What Agile teams must never automate and why
• Standards alignment: use best practices from Scrum Guide and Agile Extension to BABOK®

14.2 Q&A and Closing

• Open discussion: one AI practice I will start, one Agile practice I will protect from automation
• Training evaluation and feedback

La finalul acestui training participanții vor fi capabili să:

• utilizeze inteligența artificiala (AI) în mod responsabil în toate ceremoniile Scrum, fără a compromite valorile Agile
• folosească un catalog de prompt-uri Ai specializate pentru proiecte Agile SCRUM
• realizeze activitățile de proiect mai rapid (de 3-5 x) și mai bine (cu cele mai bune practici)
• creeze livrabile Agile îmbunătățite cu AI: User Stories, Epics și structura priorităților
• folosească AI pentru backlog management, story mapping, release planning și roadmaping
• genereze Definition of Ready și Definition of Done cu ajutorul Inteligentei Artificiale (AI)
• aplice AI în Sprint Planning, Daily Scrum, Sprint Review și Sprint Retrospective
• evalueze răspunsurile AI în ceea ce privește completitudinea, acuratețea si sa evite răspunsurile greșite (halucination)
• înțeleagă și sa evitate riscurile utilizării Inteligentei Artificiale (AI)

• Scrum Masters (toate nivelurile de experienta);
• Product Owners și Product Managers;
• Business Analysts, care fac livrabile de analiza pentru proiecte Agile
• Dezvoltatori IT implicați în activitatea de implementare a soluțiilor de proiect;
• Manageri de Proiect, Agile Coaches, Team Leads si Managers implicati in proiecte Agile;

• experiență cu metodele Agile (Scrum sau Kanban)
• familiaritate cu User Stories și participarea la ceremoniile SCRUM (Sprint Planning, Daily Scrum, Sprint Review, Sprint Retrospective)
• Familiaritate cu livrabilele din SCRUM Agile: Backlog, Product Goal, Sprint Goal, DoR, DoD
• Acces la un cont gratuit pentru un instrument AI — Claude, ChatGPT sau Gemini — creat înainte de training
• Nu este necesară experiență anterioară cu Inteligenta Artificiala (AI) sau machine learning
• Activitate pregătitoare (daca este fezabil): participanții aduc 5 cerinte (User Stories) reale din backlog-ul echipei lor pentru a le putea discuta

De preferat:

• experiență de business analyst sau familiaritate cu standardul Agile Extension to BABOK®;
• participarea anterioară la trainingul Trilex: “Scrum Agile Framework” https://trilex.ro/training-curs/scrum-agile-framework/

Notă: Experiența în proiectele Agile SCRUM aduce un plus in înțelegerea livrabilelor de proiect realizate cu ajutorul inteligenței artificiale (AI).
Acest training se bazează pe experiența participanților existentă în Agile — nu introduce conceptele de bază Agile, ci se concentrează exclusiv pe îmbunătățirea cu Inteligența Artificiala (AI) a tehnicilor cunoscute deja, pentru a lucra mai eficient, mai rapid și cu rezultate mai bune în proiectele reale.
Conținutul si formatul trainingului este ușor accesibil atât persoanelor din business cat si persoanelor mai tehnice