A few months ago, a friend called me.

“Dejan, I need a simple shift scheduling app. A table, some names, who works when. Nothing complicated.”

An hour and a half later, we had uncovered:

• A priority algorithm that distinguishes “greedy” from “modest” shift requests

• An automatic displacement mechanism when a higher-priority worker enters a full slot

• Completely different logic for weekdays vs. weekends

• Time-locked access windows (Fridays only, 8AM–12PM)

• Two user roles with different permissions

• 12 sections of functional specification

From one phone call.

Here’s what actually happened: I used AI - five structured prompts - to pull all of that out of a casual conversation. What normally takes three meetings and two weeks took one working session.

That process is what I built a course around.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

THE AI-POWERED BACKLOG REFINER

From Vague Request to Sprint-Ready User Story in 10 Minutes

https://www.whatisscrum.org/ai-powered-backlog-refiner/

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

C.R.A.F.T. is a 5-step prompting framework for Product Owners and Scrum Masters. Five copy-paste prompts that turn any vague stakeholder request into a sprint-ready User Story, with Acceptance Criteria, Gherkin scenarios, and an INVEST quality check, in under 10 minutes.

Not “paste into ChatGPT and hope.” A framework. In order. Every time.

What you get for $29:

• 34 video lectures (2.5 hours) — use the framework the same afternoon

• 7 PDF toolkits (~45 pages) — prompt library, Product Bible template, team adoption playbook, 30 real stakeholder requests with full C.R.A.F.T. outputs

• The full Moj Raspored case study — watch the exact session that became this course, live on screen

• 5 quizzes + capstone exercise + certificate of completion

• BONUS: User Story Mastery — my full companion course on writing User Stories from scratch, normally $59, included free

Works with ChatGPT, Claude, Copilot, Gemini, or any on-premise AI your company allows.

Who this is for:

Product Owners and Scrum Masters with 1–5 years of experience who already know what a User Story is, but spend 30 to 90 minutes writing each one, watch them get rejected in sprint planning for missing acceptance criteria, and sit through 2-hour refinement meetings that feel like writing workshops.

If that’s your sprint right now, this course pays for itself before the sprint review.

▶ Enroll now — $29 early-bird

https://www.whatisscrum.org/ai-powered-backlog-refiner/

Questions? Just reply to this email.

— Dejan Majkic

CIO, Agile Educator, whatisscrum.org

140,000+ students taught · 4.4 stars across 7,600+ reviews

P.S. The 10-Story Promise: use C.R.A.F.T. on your next 10 stories. If they don’t save you at least 5 hours, email me your prompts and outputs. I’ll personally rewrite them for your domain, free. If they still don’t save you time, full refund within 30 days. You keep everything regardless.

Your Scrum team doesn’t have a Scrum problem. It has a PEOPLE problem.

One person complains in every retro but never owns an improvement.

A talented developer turns every code review into a fight.

Two or three good people carry the whole team… until they burn out and leave.

Scrum didn’t create this. Scrum exposed it.

In this video I share the simple but brutal “Fire, Train, or Keep” framework I use with Scrum teams and leaders to separate high-ownership people from the ones who quietly destroy culture and performance.

Whether you’re a Scrum Master, Engineering Manager, Agile Coach or team lead — this is the real leadership work most frameworks avoid.

Great Scrum is not about perfect ceremonies.

It’s about low ego, high ownership, and hunger to learn and deliver value every sprint.

👉 If you recognized your team in this video, the hard part is doing something about it.

I built the Agile and Scrum Accelerator for leaders who are ready to have the real conversations, protect psychological safety, and build teams with genuine ownership.

Practical tools. Working agreement templates. Coaching questions. Community of leaders doing the same work.

🔗 Start here → https://www.whatisscrum.org

Drop a comment:

Are you the leader who waited too long… or the one who acted early?

Someday is not a day of the week…

Dejan

In this video, we break down the 3 critical mistakes Product Owners and Scrum Masters make when bringing AI-generated user stories into Jira, and how to fix them in 5 seconds.

Learn why AI answers structured questions, while your team answers reality questions (featuring a real-world case study from the “My Schedule” project).

Get the Full C.R.A.F.T. Framework Here: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Dejan Majkic

This post reveals why your best ideas take months to launch and shows how replacing rigid project checklists with a true Agile culture can shrink your delivery time from months to weeks.

READ IT HERE.

Dejan Majkic

Wondering why your team’s AI workflow falls apart after a few sprints? Most AI systems work great for new features, but completely break down the moment a bug, tech debt ticket, or UX tweak enters the backlog.

In this video, I show you how the C.R.A.F.T. framework handles all four work types without forcing you to relearn a single thing.

Using a real-world case study from “My Schedule” —a custom WordPress scheduling plugin built for a production crew, we walk through how to adapt your AI prompts dynamically. By changing just a few lines in Step 3 (Ask for Structure), you can seamlessly generate perfectly formatted stories for:

Stop asking “Can AI handle this type of story?” and start running your entire backlog through a single, unified process. You remain the editor; the AI just learns to draft in four different shapes.

If you found this valuable, make sure to like, subscribe, and drop a comment below about how your team currently handles AI in your sprints!

🔗 Get the full C.R.A.F.T. Framework & Prompts here: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Using a real-world WordPress shift-scheduling plugin (”My Schedule”) as a case study, you’ll see exactly how to prevent AI from hallucinating features you don’t need (like SaaS billing or complex notifications) and how to make it write user stories that actually sound like your team.

👉 Download the Full 9-Section Product Bible Framework Here: https://www.whatisscrum.org/ai-powered-backlog-refiner/

If you want to master AI-powered backlog refinement, scale your Agile sprints, and hold AI to your exact Definition of Ready, this is the workflow you need.

If you found this helpful, please LIKE, SUBSCRIBE, and leave a comment with how you currently manage your AI prompts for product management!

Dejan Majkic

www.whatisscrum.org

In this video, we dive into Step 5 of the C.R.A.F.T. Framework: Test and Validate. Using a real-world complex shift-coordination priority algorithm as our case study, watch exactly how to run a dual-pass validation check on a user story in under 3 minutes using Gherkin Syntax and the INVEST Criteria.

We map out 3 critical behavioral scenarios live:

1️⃣ The Happy Path (Standard execution)

2️⃣ The Displacement Path (Handling complex edge cases and business rules)

3️⃣ The Rejection Path (Fallback rules like first-come, first-served)

By running this quick validation, you ensure your development team writes the code right the first time, aligns stakeholders instantly, and catches hidden dependencies before they cost you time and money.

If you are a Product Owner, Scrum Master, or Agile Coach looking to turn messy stakeholder phone calls into flawless functional specifications in a single working session, this video is for you.

Resources Mentioned:

Get the Version 2 Functional Specification PDF & 5 Bonus Case Studies: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Subscribe for more practical, non-tech Agile and Scrum walkthroughs!

The “happy path” in a user story is a trap. You think your backlog item is ready for development, but what happens when the logic bends sideways?

In this video, we dive into Step 4 of the C.R.A.F.T. framework: Forcing Edge Cases. Using the AI-Powered Backlog Refiner, we stress-test a complex priority algorithm story.

Watch as the AI surfaces 4 critical logic gaps in 7 seconds, invisible flaws that would have caused production bugs, emergency hotfixes, and broken deployment windows.

Learn how to use AI for pattern recognition, failure modes, and boundary conditions so you can write bulletproof acceptance criteria before a single line of code is written.

🚀 Master the full framework and get the exact prompts here: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Dejan Majkic

You’ll see live how to take a chaotic pile of 47 stakeholder answers and instantly decompose it into 6 distinct, independent feature areas, including a complex priority displacement algorithm that would usually take weeks of back-and-forth meetings to refine.

GET THE FULL C.R.A.F.T. FRAMEWORK & CASE STUDY PACK HERE:

👉 https://www.whatisscrum.org/ai-powered-backlog-refiner/ (Download the complete My Schedule Case Study PDF with all prompts, raw requests, and worked solutions)

What You Will Learn:

The 2-Minute Pre-Processing Prompt: The exact 4-instruction paragraph to copy and paste under your messy raw data.

The Real Project Math: Why a 2-minute untangling step combined with 10 minutes per story collapses weeks of refinement meetings into just 1 hour and 10 minutes.

Step-by-Step Breakdown: Watch Context Injection, Raw Request Capture, and Asking for Structure map out a bulletproof Acceptance Criterion for a production-breaking algorithm.

How to Spot Edge Cases & Risks: Identify race conditions, dependencies, and unconfirmed assumptions before a single line of code is written.

If you want to build faster, eliminate ambiguity for your developers, and master AI product management, hit that Subscribe button!

Dejan Majkic | MA in CS&IT | Scrum Master | Product Owner | Agile Coach

www.whatisscrum.org

Many Agile product managers, Scrum product owners, and business analysts struggle with missing business logic and vague stakeholder requirements that derail sprints.

Here, we break down Step 2 of the C.R.A.F.T. framework: Raw Request Capture. You will see exactly how feeding a messy stakeholder conversation into an AI prompt can surface hidden complexity, edge cases, and unspoken rules before a single line of code is written.

How to Prevent Scope Creep Using AI:

• AI requirements gathering works best when you do not clean up the input.

• By pasting a raw stakeholder transcript in quotes and instructing the AI to surface uncertainty rather than hallucinate answers, you turn vague feelings into hard, code-ready operational logic.

• We dive deep into the 4 buckets of hidden complexity uncovered by the model and reveal the exact prompt engineering framework you can use to turn fuzzy words like “fair,” “simple,” and “intuitive” into precise development thresholds.

Stop discovering requirements during development one costly Slack thread at a time. Learn how to make software complexity visible in advance.

Master the complete system and download the full C.R.A.F.T. framework blueprint here: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Dejan Majkic | MA in CS&IT | Scrum Master | Product Owner | Agile Coach

www.whatisscrum.org

What You’ll Learn in This Video:

The 5 Essential Sections: Product Definition, Users/Personas, Hard Constraints, Definition of Ready (DoR), and Team Conventions.

Why Context is greater than Prompts: How to strip away the AI’s “freedom to invent” so it stops hallucinating features (like unwanted notifications) and starts matching your team’s exact workflow.

Side-by-Side Comparison: Watch the AI transform a generic, useless template story into a precise, localization-ready user story using terms like “cameraman” and “radnik” within a strict Friday submission window.

Resources & Links:

🎓 Get the C.R.A.F.T. framework: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Dejan Majkic | MA in CS&IT | Scrum Master | Product Owner | Agile Coach

www.whatisscrum.org

In this video, we break down the case study of Vladimir’s “simple” scheduling app for 27 TV cameramen. What started as a basic request for “a table and a few workers” quickly spiraled into a complex web of priority algorithms, displacement mechanisms, and unknown edge cases.

Stop falling into the compressed complexity trap! Learn the 5 crucial places where complexity hides and how to shift your mindset from solving a “scheduling problem” to solving a “resource allocation” problem.

By the end, you’ll see how to map out Users, Constraints, and Scheduling Windows to save your team hours of Slack ping-pong and broken assumptions.

What You’ll Learn in This Video:

1. The Backstory: Vladimir’s “Simple” Request

2. The “Simple App” Trap & Compressed Complexity

3. The 5 Places Complexity Hides (Checklist)

4. Breaking Down the Case Study: Rules, Exceptions & Edge Cases

5. The Golden Rule: Competing Demand vs. Constrained Supply

6. Preview: Building a 12-Section Product Bible in 5 Minutes

Mentioned in this video:

AI-Powered Backlog Refiner is an online course that teaches Product Owners and Scrum Masters a 5-step AI prompting framework called C.R.A.F.T. https://www.whatisscrum.org/ai-powered-backlog-refiner/

Dejan Majkic | MA in CS&IT | Scrum Master and Product Owner | Agile Coach

www.whatisscrum.org

In this video, we walk through the exact 3-pass review system used to bulletproof C.R.A.F.T. outputs and protect the product from bad AI assumptions.

We cover the 80/20 rule in practice: 1 minute of prompting, 7 minutes of structured, strategic product thinking.

In this debrief, you will learn:

1. Pass 1: Spec Alignment - How to spot and eliminate AI hallucinations instantly.

2. Pass 2: Completeness - How to handle AI-surfaced edge cases before they derail your Sprint Planning.

3. Pass 3: Scope Integrity - How to apply real Product Owner judgment to split or bundle stories correctly.

4. The 3 Fatal AI Mistakes - Hallucinated features, missing constraints, and ambiguous scope.

Subscribe for more strategies on mastering Agile, Scrum, and AI-driven product ownership.

AI-Powered Backlog Refiner: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Dejan Majkic | MA in CS&IT | Scrum Master | Product Owner | Trainer

www.whatisscrum.org

In this talk, Dejan, CIO at the Tax Administration of Republika Srpska, shares a practical roadmap for tax administrations that have not implemented AI yet.

This video explains how to start with AI in government without falling into the common traps: buzzword pressure, vendor noise, weak data, unclear ownership, and missing governance.

You’ll learn:

1. Why being deliberate is not the same as being late

2. The 3 questions every tax administration should answer before an AI pilot

3. What the SyRI case teaches public institutions about transparency and trust

4. Why the best first move is one narrow, low-risk pilot

5. 3 safe starting points: taxpayer chatbot, OCR for archives, and basic case routing

This video is for tax leaders, CIOs, public-sector digital transformation teams, tax technology professionals, and anyone working on responsible AI in government.

Key topics:

AI in tax administration, artificial intelligence in government, tax technology, public sector AI, AI governance, taxpayer services, OCR, case routing, digital transformation, and responsible AI.

Core message:

Being deliberate is not the same as being late. If your administration has not implemented AI yet, you still have the chance to start clean, avoid other people’s mistakes, and build the right foundations first.

www.whatisscrum.org

I ran the brand-new C.R.A.F.T. framework on a real stakeholder request in real time and turned vague Slack copy into:

• 6 bulletproof Acceptance Criteria

• 8 edge cases auto-discovered

• 4 Gherkin scenarios

• Full INVEST scorecard

Watch the exact prompts + the one moment every junior PM misses (AI hallucinated email notifications even though the context block said “no notifications in v1”).

You’ll see:

• Context Block injection (paste once, reuse forever)

• Raw request → structured story in seconds

• How to catch AI hallucinations instantly

• Edge-case prompting that finds risks BEFORE dev starts

• One-line correction prompt that fixes everything

Perfect for Product Owners, Product Managers, Scrum Masters, and anyone tired of 30-minute refinement meetings.

Timer starts at 0:00 → Story sprint-ready at 8:12.

Next module drops soon: when the stakeholder sends the messy version.

👉 Want AI + Product Management framework? https://www.whatisscrum.org/ai-powered-backlog-refiner/

#ProductManagement #AIforPM #Jira #UserStories #Agile #ProductOwner #scrum

Learn how AI helps uncover hidden requirements, edge cases, and stakeholder ambiguity before software development starts.
CLICK HERE TO READ

Dejan Majkic

👉GET THE FULL FRAMEWORK: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Spot issues like “FAIL: Too big—split it!” before they waste your team’s time. Quality gate: If it passes INVEST + Definition of Ready (DoR), it’s sprint-ready. No more “probably ready” stories!

Perfect for Product Owners, Agile Coaches, QA teams, and Scrum Masters building features like worker management in tools like Jira.

🔥 Full course: https://www.whatisscrum.org/ai-powered-backlog-refiner/

#UserStory #INVEST #Gherkin #Agile #Scrum #ProductOwner #SprintPlanning #QA

(Example Gherkin: “Given admin logged in, when submit valid worker form, then worker appears in list.”)

This is the exact moment the C.R.A.F.T. framework goes from “faster stories” to “BETTER stories.” One single AI prompt surfaces 6–12 real edge cases, failure modes, and boundary conditions your team would have discovered the hard way in Sprint Planning.

You’ll see the exact prompt + the 6 edge cases it found for Vladimir’s real worker-management story (deleting active workers, self-deletion, username collisions, and more). Then I show you the 3-bucket system (A = Add as AC, B = Separate Story, C = Flag as Risk + Discard) so you make smart scope decisions instead of blind ones.

✅ After this step you’ll walk into refinement saying “We already considered that and consciously left it out”, instead of “I didn’t think of that.”

This is Step 4 of the complete C.R.A.F.T. framework for writing user stories with AI.

Previous steps:

• Step 3 → Structured Stories & Acceptance Criteria https://www.whatisscrum.org/ai-powered-backlog-refiner/

If you’re a Product Owner, Scrum Master, or Agile coach who wants user stories that actually survive development, drop a 🔥 in the comments and tell me which C.R.A.F.T. step you want next.

#UserStories #AcceptanceCriteria #Agile #Scrum #AI #ProductOwner #CRAFTFramework #EdgeCases #SoftwareDevelopment

In this video (ABOVE), I show how we cut refinement from two hours to thirty minutes without hiring more people or adding new tools.

The shift was simple: we stopped drafting and debating at the same time. Before the meeting, every vague request gets run through AI to produce a structured story, acceptance criteria, edge cases, and dependency flags. What used to take endless discussion now becomes an 80% draft in about ten minutes per story.

That changes everything. The meeting is no longer a blank-page argument. It becomes a focused decision session about scope, trade-offs, and priorities, the part that actually matters.

I am building a framework around this: https://www.whatisscrum.org/ai-powered-backlog-refiner/

Are you still writing Agile user stories from scratch?

In this video, I reveal Step 3 of my AI product management framework. You’ll see exactly how to combine stakeholder context, raw requests, and clarifying answers into a single, highly structured AI prompt that does 80% of the heavy lifting.

Vague prompts give you vague stories. But with this exact copy-paste template, you’ll instantly generate the Title, User Story (As a/I want/So that), Acceptance Criteria (Given/When/Then), Dependencies, Assumptions, and Estimates. Stop writing, start editing, and reclaim your time as a Product Owner!

The Exact AI Prompt Used in This Video:

“Using the context from above, the stakeholder request in quotes, and these clarifying answers, write a user story in this exact format: Title. User Story: As a, I want, so that. Acceptance Criteria: 4-7 testable criteria in Given, When, Then format. Dependencies. Assumptions, flagging anything that needs stakeholder confirmation. Story Points Estimate Range. Risks.”

Level Up Your Agile Career:

Ready to master Scrum and Product Management? Check out my comprehensive training resources, including exam walkthroughs (like the PSM1) and complete Agile course bundles to help you lead high-performing teams: https://www.whatisscrum.org/ai-powered-backlog-refiner/

#Scrum #ProductOwner #UserStories #ChatGPT #AgileFramework #ScrumMaster #AIforBusiness