DEV Community: Denis Glushko

The New Agile Team Structure Nobody Talks About

Denis Glushko — Thu, 20 Aug 2026 05:30:18 +0000

Opening

Traditional Agile Teams:

Product Owner
Scrum Master
Developers
QA

Modern Agile Teams:

Product Owner
Developers
QA
AI Assistants

The AI assistant is becoming a real delivery participant.

What AI Is Doing Today

AI already assists with:

Story generation
Backlog refinement
UML diagrams
Test case generation
Coding
Code reviews
Documentation
The Hidden Problem

Most organizations deploy AI without governance.

That creates:

Hallucinated requirements
Unsafe code
Compliance risks
Technical debt
New Operating Principle

A rule we adopted:

AI assists. Humans approve.

AI can suggest.

Humans remain accountable for:

Business problems
Architecture decisions
Security approval
Value realization

This mirrors emerging AI-assisted delivery governance discussed in enterprise Agile workflows.

Conclusion

The most successful Agile teams will not have the most AI.

They will have the clearest human decision gates.

I Let GitHub Copilot Build an Entire Sprint. Here's What Happened.

Denis Glushko — Thu, 20 Aug 2026 05:29:02 +0000

Last year AI mostly generated code.

Now it can generate:

User Stories
Acceptance Criteria
Test Cases
API Documentation
SQL Scripts
Release Notes

So I decided to test something.

Could AI help build an entire Sprint?

Experiment

Input:

Plain Text
1
Business Requirement
Show more lines

AI Generated:

Epic
Features
User Stories
Acceptance Criteria
Test Scenarios
Technical Tasks
Findings

AI was amazing at:

✅ Documentation

✅ Boilerplate work

✅ Missing scenarios

✅ Edge case suggestions

AI struggled with:

❌ Business context

❌ Organizational politics

❌ Prioritization decisions

❌ Stakeholder alignment

Lessons Learned

AI is not eliminating Scrum Masters.

AI is eliminating administrative work.

The Product Owner role becomes more strategic.

The developer role becomes more architectural.

Final Thought

The best engineer in 2026 might not be the fastest coder.

It might be the person who asks the best questions.

Kaspa Toccata Hard fork succeed

Denis Glushko — Wed, 01 Jul 2026 05:19:38 +0000

“AI Is the New Scrum Master (And That’s a Problem)”

Denis Glushko — Mon, 29 Jun 2026 05:03:47 +0000

Let me say something controversial:
AI might become your best Scrum Master…
and your biggest risk.

AI Is Already Changing Agile
2026 Agile teams are using AI for:

Backlog prioritization
Sprint predictions
Automated retrospectives

Some teams are already 30–50% more efficient with AI support [10pearls.com]
Sounds amazing, right?
Not so fast.

The Hidden Danger
AI optimizes for efficiency.
Agile should optimize for value.
Those are not the same thing.
AI will:

Suggest faster delivery
Optimize workload distribution
Predict bottlenecks But it won’t challenge bad strategy.

The Real Risk
If you blindly follow AI-driven Agile, you’ll get:

Faster delivery of the wrong features
Perfectly optimized mediocrity
Teams that stop thinking critically

AI removes friction.
But friction is where innovation happens.

What AI Cannot Replace
Agile is fundamentally human:

Negotiation
Empathy
Product intuition

AI can analyze data.
But it cannot answer:
“Should we build this at all?”
The Winning Strategy
Treat AI like:
A junior team member with superpowers
Use it to:

Automate repetitive Agile rituals
Generate insights from data
Improve planning accuracy

But keep humans responsible for:

Vision
Trade-offs
Value decisions

Final Thought
AI won’t kill Agile.
But bad leaders using AI will.

“Agile Is Not Dead — Your Implementation Is”

Denis Glushko — Mon, 29 Jun 2026 05:02:14 +0000

Everyone keeps asking: “Is Agile dead?”
No.
What’s dead is the way you’re doing it.
Agile didn’t fail. Your company just turned it into a checklist.
Let’s be brutally honest.

The Problem Nobody Talks About
Most teams today don’t practice Agile. They perform Agile theater:

Standups where nobody listens
Sprint planning based on wishful thinking
Retros that repeat the same problems every two weeks

You’re not Agile. You’re just busy.
And the worst part?
It feels productive.

Cargo Cult Agile Is Killing Teams
There’s a term for this: Cargo Cult Agile.
Teams copy rituals without understanding principles:

“We have Scrum, so we’re Agile”
“We do sprints, so we’re efficient”

But Agile was never about ceremonies.
It was about adaptability and value delivery [atlassian.com]

What Real Agile Looks Like in 2026
Agile today is evolving fast:

AI assists sprint planning and backlog refinement
Teams focus on product outcomes, not project outputs [10pearls.com]
Continuous delivery replaces rigid sprint cycles

Modern Agile teams:

Kill useless meetings
Measure customer impact, not velocity
Change plans mid-sprint when needed That’s real agility.

Hard Truths You Need to Hear

If your roadmap never changes → you’re not Agile
If stakeholders don’t show up → you’re not Agile
If delivery doesn’t impact users → you’re not Agile

Agile is uncomfortable.
That’s why most organizations avoid doing it properly.

How to Fix It
Start here:

Replace velocity with outcome metrics
Cut 30% of ceremonies immediately
Talk to customers every sprint
Let teams say “no” more often

Agile isn’t about doing more.
It’s about doing less, but better.

Why Enterprise Agile Needs Less Scrum and More Economic Thinking

Denis Glushko — Wed, 24 Jun 2026 11:48:17 +0000

Most Agile transformations focus on process.
But enterprise success depends on decisions.
And decisions are economic.

The Hidden Problem
Teams follow Agile perfectly:

sprint planning
stand-ups
demos

But still:

wrong features are delivered
valuable work is delayed
priorities constantly change

Why?
Because:

Process does not equal value

Agile Without Economics Is Incomplete
Agile tells you how to deliver.
It doesn’t fully define:

what to prioritize
how to compare initiatives
how to allocate capacity

That’s where economic thinking comes in.

The Missing Layer: Portfolio Economics
Every organization has limited capacity.
Every decision is a trade-off.
But most enterprises don’t treat it that way.
Instead:

everything is “important”
no one owns prioritization
decisions depend on influence

What Economic Thinking Looks Like
Instead of asking:

“Can we do this?”

Ask:

“Is this worth doing compared to alternatives?”

Practical Model
Evaluate initiatives across:

business value
urgency
risk reduction
effort

Then compare them consistently.

Why This Changes Everything
It creates:

transparency
alignment
accountability

And removes:

politics
guesswork
noise

Agile + Economics = Real Enterprise Agility
Without economics:

Agile becomes delivery engine

With economics:

Agile becomes value engine

Final Thought
Scrum won’t fix your portfolio.
Ceremonies won’t fix prioritization.
Velocity won’t fix strategy.
But economic thinking will.
Because:

In enterprise Agile,
the biggest bottleneck is not delivery

It’s decision-making

AI in Project Management: Why Faster Delivery Can Make Things Worse

Denis Glushko — Wed, 24 Jun 2026 11:46:56 +0000

AI is everywhere.
It writes tickets, summarizes meetings, generates reports, predicts risks, and even suggests priorities.
And most teams are asking:

“How can AI make us faster?”

But that’s the wrong question.

The Real Risk of AI in Agile
AI doesn’t fix problems.
It amplifies them.
If your system is good → AI makes it better
If your system is broken → AI breaks it faster

Where AI Helps
Let’s be clear — AI is incredibly useful:

summarizing backlog items
generating acceptance criteria
analyzing risks
improving reporting
automating repetitive tasks

These are massive productivity gains.

Where AI Fails
But here’s what AI cannot do:

understand real business context
resolve stakeholder conflicts
define product strategy
make trade-offs across initiatives

And most importantly:

AI cannot decide what actually matters

The Dangerous Pattern
When organizations adopt AI blindly, this happens:

more tickets created
more backlog items generated
more documentation produced

Everything looks:
✅ structured
✅ efficient
✅ professional
But underneath?
❌ priorities are still unclear
❌ value is still undefined
❌ decisions are still weak

Faster ≠ Better
AI reduces effort, not complexity.
If you build the wrong thing faster:

rework increases
costs increase
frustration increases

AI doesn’t change that.

What High-Performing Teams Do Differently

Use AI for Execution — Not Strategy AI supports:

writing
analyzing
structuring

Humans own:

decisions
priorities
trade-offs

Focus on Decision Quality Before asking:

“Can AI generate this?”

Ask:

“Should this exist?”

Strengthen Product Ownership AI will expose weak Product Owners. Because:

writing stories → can be automated
defining value → cannot

The New Role of the Product Owner
From:

backlog manager

To:

decision-maker
value owner
alignment driver

Final Thought
AI will not replace project managers or Product Owners.
But it will expose who is:

thinking
vs
just managing tasks

Because:

AI makes execution faster
But only humans can make it meaningful

Agile Is Not Broken — Your Portfolio Is

Denis Glushko — Wed, 24 Jun 2026 11:44:57 +0000

Let’s be honest.
Agile isn’t failing in your organization.
Your portfolio management is.
Most Agile transformations focus on teams:

Scrum ceremonies
backlog grooming
sprint velocity
Jira dashboards

And teams do improve.
But the business outcome?
Still disappointing.

The Real Problem: Agile Solves Delivery, Not Decisions
Agile is great at answering:

“How do we deliver this faster?”

But enterprise environments struggle with a more important question:

“Why are we building this at all?”

That’s where things break.

What Actually Happens in Enterprises
At scale, Agile teams operate inside a system where:

priorities constantly shift
multiple stakeholders compete for attention
urgent requests override strategic work
governance sits outside delivery

This leads to a familiar pattern:

teams are busy
backlogs are full
releases happen

…but business impact is unclear.

The Portfolio Bottleneck
The issue isn’t execution.
It’s decision-making before execution.
Typical portfolio problems:

everything is “high priority”
no consistent prioritization model
value is assumed, not defined
trade-offs are avoided

So what happens?
Teams get overloaded with work that shouldn’t exist.

The Shift: From Delivery-Centric to Decision-Centric Agile
High-performing organizations don’t just optimize delivery.
They optimize what enters delivery.
This requires a mindset shift:
From:

“Let’s deliver everything faster”

To:

“Let’s deliver only what matters”

What Works Instead

Make Value Explicit Every initiative should answer:

What problem are we solving?
What measurable outcome do we expect?
What happens if we don’t do this?

No clear answer = no priority.

Introduce Structured Prioritization Stop relying on opinion. Start using criteria like:

business value
strategic alignment
urgency
complexity

This changes conversations from:

“We need this urgently”

To:

“This delivers more value than alternatives”

Limit Demand, Not Teams Most organizations try to increase team capacity. Better organizations:

reduce low-value work
protect team focus
say “no” more often

Align Governance With Flow Governance isn’t the enemy. But if it lives outside Agile, it creates bottlenecks. Bring it inside:

define rules upfront
integrate approval into flow
automate where possible

The Hard Truth
Agile teams rarely fail.
The system around them does.
If your portfolio is broken:

Agile will look ineffective
teams will feel frustrated
stakeholders will lose trust

Final Thought
Agile doesn’t need to be fixed.
Your decision system does.
Because in the end:

Delivering fast doesn’t matter
If you’re delivering the wrong things

Debunking Common Misconceptions About Remote Work

Denis Glushko — Fri, 19 Jun 2026 11:59:20 +0000

According to Forbes Advisor, by 2025, an estimated 32.6 million Americans will be working remotely, which equates to about 22% of the workforce
USA TODAY reports that 36% of U.S. white-collar workers who have reached, at most, a high school degree or equivalent are fully remote
In the European Union, 22% of employed people aged 15 to 64 worked from home in 2023, according to new data by Eurostat. This consisted of 9% who usually worked from home and 13% who worked from home only occasionally
The Netherlands had the highest share of people who worked from home at least occasionally in 2023 at 51.9%, followed by Sweden (45.3%), Iceland (42.6%), Norway, and Finland (around 42%)

Remote work has become a significant part of the modern work landscape, yet several misconceptions persist. Let’s explore and debunk some of the most common myths surrounding remote work.

Myth 1: Remote Workers Are Less Productive

A widespread belief is that remote workers are less productive due to the lack of direct supervision. However, this is far from the truth. Numerous studies have demonstrated that remote workers often outperform their in-office counterparts. The absence of typical office distractions, such as in-personal meetings and casual conversations, allows remote workers to focus better on their tasks. Additionally, remote workers often have more control over their work environment, which can lead to increased productivity and job satisfaction.

Myth 2: Remote Work Leads to Isolation

While it’s true that remote work can sometimes result in feelings of isolation, many remote workers actively seek social interactions through virtual meetings, coworking spaces, and social media. Companies can also play a significant role in mitigating isolation by organizing regular virtual team-building activities and promoting open communication. By fostering a sense of community and ensuring that employees feel connected, organizations can help remote workers maintain a healthy social balance.

Myth 3: Communication Breaks Down in Remote Work

Another common misconception is that communication suffers when teams are not physically together. In reality, remote work can enhance communication through the use of digital tools like Slack, Zoom, and Microsoft Teams. These platforms facilitate real-time communication and collaboration, often making interactions more efficient than traditional methods. With clear communication channels and regular check-ins, remote teams can maintain strong communication and collaboration.

Myth 4: Remote Workers Are Always Available

There’s a belief that remote workers are constantly on call and can work at any time. This misconception can lead to burnout if not addressed properly. It’s crucial for remote workers to set clear boundaries and for employers to respect these boundaries to maintain a healthy work-life balance. Establishing set working hours and encouraging employees to disconnect after work can help prevent burnout and ensure that remote workers remain productive and satisfied.

Myth 5: Remote Work Is Only Suitable for Certain Jobs

While some jobs are inherently more suited to remote work, the range of roles that can be performed remotely is expanding. With the right tools and processes, many traditionally office-based jobs can now be done remotely. Advances in technology have made it possible for a wide variety of tasks to be completed from anywhere, allowing more employees to enjoy the benefits of remote work.

Myth 6: Remote Work Erodes Company Culture

There’s a fear that remote work will weaken company culture. However, with intentional efforts, companies can maintain and even strengthen their culture. Regular virtual meetings, recognition programs, and opportunities for social interaction can help keep employees engaged and connected. By prioritizing company culture and making a concerted effort to include remote workers in all aspects of the organization, companies can ensure that their culture remains strong.

The Future of Remote Work

As remote work continues to evolve, it’s clear that it is here to stay. Hybrid work models, which combine remote and in-office work, are becoming increasingly popular. These models offer the best of both worlds, providing flexibility while maintaining opportunities for in-person interaction. Organizations that embrace remote work and virtual collaboration are likely to see numerous benefits, including increased employee satisfaction, higher productivity, and access to a global talent pool.

However, success in this new environment requires a proactive approach to addressing the challenges and leveraging the opportunities that remote work presents. By debunking these common misconceptions and adopting best practices, organizations can better support their remote workers and create a thriving, productive work environment.

In conclusion, remote work has transformed the way we work. By understanding and addressing the misconceptions surrounding remote work, organizations can better navigate this new normal and leverage the benefits of remote work to enhance productivity and employee satisfaction. With the right strategies and tools in place, remote work can be a highly effective and rewarding way to work.

Challenges in Implementing AI in Agile Projects

Denis Glushko — Fri, 19 Jun 2026 11:54:29 +0000

Integrating Artificial Intelligence (AI) into agile projects offers numerous benefits, but it also presents several challenges that organizations must navigate. Understanding these challenges is crucial for successful implementation and maximizing the potential of AI in agile environments.

Resistance to Change

One of the most significant hurdles in implementing AI in agile projects is resistance to change. Team members who are unfamiliar with AI or fear that it might replace their jobs can be hesitant to adopt new tools and practices. This resistance can slow down the integration process and reduce the effectiveness of AI solutions. To overcome this, organizations need to focus on education and transparency. Demonstrating the value that AI can bring to the team and ensuring that employees understand how AI will enhance their roles rather than replace them is essential.

Data Quality and Availability

AI algorithms rely heavily on data to make accurate predictions and provide valuable insights. However, poor quality data or insufficient data can significantly impact the effectiveness of AI. Ensuring that data is clean, relevant, and accessible is crucial for successful AI implementation. Organizations must invest in data management practices, including data cleaning, validation, and integration, to ensure that their AI systems have the high-quality data they need to function effectively.

Integration with Existing Systems

Integrating AI tools with existing project management systems and workflows can be complex. Compatibility issues, data migration challenges, and the need for customization can pose significant hurdles. This complexity requires careful planning and collaboration with IT teams to address these challenges effectively. Organizations should conduct thorough assessments of their current systems and workflows to identify potential integration issues and develop strategies to mitigate them.

Skill Shortage

There is a notable shortage of professionals with the skills required to implement and manage AI technologies. This skills gap can slow down the adoption of AI in agile projects. To address this issue, organizations need to invest in training and development for their existing team members or hire skilled professionals with expertise in AI and agile methodologies. Building a team with the right skills is essential for leveraging the full potential of AI in agile projects.

Maintaining Code Quality

AI-generated code, especially from generative AI models, needs to be carefully vetted to ensure it meets quality standards. There is a risk of introducing errors or vulnerabilities if the generated code is not thoroughly reviewed and tested. Organizations must establish robust code review processes and quality assurance practices to ensure that AI-generated code is reliable and secure.

Ethical and Security Concerns

The use of AI raises ethical and security concerns, particularly regarding data privacy and the potential for biased decision-making. Implementing robust security measures and ethical guidelines is essential to address these concerns and ensure responsible AI use. Organizations should develop and enforce policies that prioritize data privacy, security, and ethical considerations in their AI projects.

Cost and Resource Allocation

Implementing AI can be resource-intensive, requiring significant investment in technology, infrastructure, and training. Organizations need to carefully consider the cost-benefit ratio and allocate resources effectively to ensure a successful implementation. This includes budgeting for AI tools, infrastructure upgrades, and ongoing training and support for team members.

Strategies for Overcoming Challenges

To successfully integrate AI into agile projects, you can adopt several strategies:

Education and Training: Invest in training programs to educate team members about AI technologies and their benefits. This can help reduce resistance to change and build a culture of continuous learning.
Data Management: Implement robust data management practices to ensure that AI systems have access to high-quality data. This includes data cleaning, validation, and integration processes.
Collaboration with IT Teams: Work closely with IT teams to address integration challenges and ensure that AI tools are compatible with existing systems and workflows.
Hiring and Development: Hire skilled professionals with expertise in AI and agile methodologies, and invest in the development of existing team members to build a strong, capable team.
Ethical Guidelines and Security Measures: Develop and enforce ethical guidelines and security measures to address concerns related to data privacy and biased decision-making.
Resource Allocation: Carefully allocate resources to ensure that AI implementation is cost-effective and sustainable. This includes budgeting for technology, infrastructure, and training.

By acknowledging and addressing these challenges, you can better navigate the complexities of integrating AI into agile projects and fully leverage the benefits that AI has to offer. With the right strategies and tools in place, AI can significantly enhance productivity, decision-making, and overall project success in agile environments.

The Integration of AI in Agile Project Management and Decision-Making

Denis Glushko — Fri, 19 Jun 2026 11:51:52 +0000

In recent years, the integration of Artificial Intelligence (AI) into agile project management and decision-making has revolutionized the way teams operate. AI algorithms are now capable of predicting project trajectories, identifying potential bottlenecks, and providing valuable insights for smarter decision-making. This technological advancement is enhancing productivity and freeing up time for teams to focus on more strategic tasks.

Predicting Project Trajectories

One of the most significant contributions of AI in agile project management is its ability to predict project trajectories. By analyzing historical data and current project parameters, AI algorithms can forecast future outcomes with remarkable accuracy. This predictive capability allows project managers to anticipate potential issues and make proactive adjustments to keep projects on track. For instance, AI can predict delays in project timelines based on past performance data, enabling teams to take corrective actions before problems escalate.

Identifying Potential Bottlenecks

AI is also instrumental in identifying potential bottlenecks in project workflows. By continuously monitoring project progress and analyzing data from various sources, AI can pinpoint areas where work is slowing down or where resources are being underutilized. This real-time analysis helps teams address issues promptly, ensuring that projects run smoothly and efficiently. For example, AI can detect when a particular task is taking longer than expected and suggest reallocating resources to prevent delays.

Providing Insights for Smarter Decision-Making

The insights provided by AI are invaluable for making smarter decisions in agile project management. AI algorithms can analyze vast amounts of data to uncover patterns and trends that may not be immediately apparent to human observers. These insights can inform decision-making at all levels, from strategic planning to day-to-day operations. For instance, AI can analyze customer feedback and market trends to help teams prioritize features that will deliver the most value to users.

Enhancing Productivity with Automation Tools

Automation tools powered by AI are being integrated with agile practices to enhance productivity and free up time for teams. These tools can automate repetitive tasks, such as updating project status, generating reports, and managing workflows. By reducing the manual workload, automation allows team members to focus on more strategic and creative aspects of their work. For example, AI-powered chatbots can handle routine inquiries and support tasks, freeing up human resources for more complex problem-solving.

The Future of AI in Agile Project Management

The future of AI in agile project management looks promising, with continuous advancements in technology and increasing adoption by organizations. As AI algorithms become more sophisticated, their ability to predict, analyze, and optimize project workflows will only improve. This will lead to even greater efficiencies and more successful project outcomes.

Moreover, the integration of AI with other emerging technologies, such as the Internet of Things (IoT) and blockchain, will further enhance its capabilities. For instance, IoT devices can provide real-time data on project environments, while blockchain can ensure the security and transparency of project transactions. Together, these technologies will create a more connected and intelligent project management ecosystem.

In conclusion, AI is transforming agile project management and decision-making by predicting project trajectories, identifying potential bottlenecks, and providing insights for smarter decision-making. The integration of automation tools with agile practices is enhancing productivity and freeing up time for teams to focus on strategic tasks. As AI technology continues to evolve, its impact on agile project management will only grow, leading to more efficient and successful projects.

Agile User story splitting techniques.

Denis Glushko — Fri, 19 Jun 2026 11:48:33 +0000

Why do we need split user stories?
Easier to understand
Smaller stories provide faster feedback
Small stories increase progress

Split conditions:
What this is - if a story has multiple items listed in "what" make each a separate story.

When to use it - words such as "and" or "or" appear.

Helpful questions - Are all of these conditions helpful?

Example: As a customer I want to create an order and pay for the order using ApplePay:

As a customer I want to create an order
As a customer I want to pay for the order using ApplePay

Workflow
What this is - identify the specific steps in a workflow. Implement the workflow in stages, creating multiple stories for the incremental implementation.

When to use it - the initial story describes a workflow or a process

Helpful questions - What steps does a user perform? Are all these steps necessary? Can steps be simplified?

Example: As a customer I want to create an order

As a customer I can view the items on the menu
As a customer I can select the want in my order
As a customer I can review my order
As a customer I can submit my order

Use case scenarios.
What is this - similar to the workflow pattern, one story represents the happy path, different stories for the alternative paths.

When to use it - the initial story refers to an interaction described by a use case

Helpful question - What does the happy/alternative flow look like? Are all the alternative flows necessary? Can the alternative flows be simplified?

Example: As a customer I want to create an order want to create an order

As a customer I want to order 1 item
As a customer I want to order multiple items
As a customer I want to cancel my order

Operations
What is this - splitting a story based on the different operations done on an entity (Create, read, update, delete).

When to Use it - The story is about managing or configuring something.

Helpful questions - What operations does the story entail? All of these operations necessary?

Example: As a restaurant owner I want to manage my restaurant profile so that people can order from my restaurant:

As a restaurant owner I can create a restaurant profile
As a restaurant owner I can update a restaurant profile
As a restaurant owner I can view a restaurant profile
As a restaurant owner Ican delete a restaurant profile

Zero->One->Many
What this is - splitting a story that deals with multiple instances of the entity into stories that deal with one instance, and then multiple instances.

When to Use it - You are working with an entity where multiple instances are allowed

Helpful questions - Do we need to be able to handle multiple instances, or 1 instance is sufficient.

Example - As a restaurant owner, I want to view customer order.

As a restaurant owner, I want to view customer order with no menu items.
As a restaurant owner, I want to view customer order with 1 menu item.
As a restaurant owner, I want to view customer order with multiple menu items.

Core functionality Vs Enhancements.
What this is - Splitting a story to deal with the simple core conditions and then enhancing the functionality with additional user stories.

When to Use it - When the story has a simple core that provides most of the learning.

Questions: What's the simplest version of this? What data types are supported? What parameters are relevant?

Example: As a customer I want to see my past orders so that i can see what I have eaten before

As a customer I can all of my orders
As a customer I can see orders for a given location.
As a customer I can see all the orders in the past 3 month
As a customer I can orders in revers date order

Major effort
What this is? Splitting a story in the way that requires substantial effort for the first story and less effort for all subsequent stories.

When to use it? When you apply the most obvious split, any story you do first is the most difficult.

Questions: Can we create the necessary infrastructure and deliver value at the same time?

Example: As a customer I can pay for my orders with VISA, MasterCard, UnionPay so that I don't have to deal with cash

As a customer I can pay with one debit card type (VISA, MasterCard, UnionPay)
As a customer I can pay with all debit card types, given one card type already done.

Acceptance criteria
What this is? Split a user story into multiple stories to handle different scenarios in the Acceptance criteria.

When to use it? When the team identifies the multiple scenarios for the story, or accurate, but not entirely relevant acceptance criteria identified.

Questions: What tests are used to verify this story? What acceptance criteria apply? What scenarios are relevant? Are all test scenarios are necessary?

Example: As a customer I want to create an order so that i can have something to eat.

As a customer I want to create an order with an approved debit card.
As a customer I want to create an order with a declined debit card.
As a customer I want to create an order with an in-stock item.
As a customer I want to create an order with an item not in stock.

Variations in data
What this is? The story does the same thing to different types of data. Create a story for each option.

When to use it? When a solution has to support multiple options.

Questions: Are these options necessary right now? What is the most common option we need to care for right away.

Example: As a customer I want to view the menu in my native language, so i can decide what I Want to eat.

As a customer I want to view the menu in English
As a customer I want to view the menu in Kazakh
As a customer I want to view the menu in Russian

Data Boundaries
What this is - Splitting a story dealing with several attributes of the same entity. Support a few key elements first and introduce remaining data elements later via additional stories.

When to use it - When s story dealing with several different pieces of data.

Questions - What are the essential data elements we need to have? What data elements are not needed right now.

Example: As a customer I want to view menu item information so I can decide what I want to eat

As a customer I want to view dish name & description
As a customer I want to view dish ingredients
As a customer I want to view nutritional information
As a customer I want to view picture of dish

Interface variations
What this is - Splitting a story dealing with a complex interface with addition stories to incrementally add in complexity

When to use it - When the user story deals with a complex user interface where a simpler one will work in the meantime

Questions - What is the simplest user interface we can use?

Example: As a customer I can view the menu so I can select what I want to eat.

As a customer I can view the menu as .PDF
As a customer I can view the menu as a searchable web-page
As a customer I can view the native in a browser

Platform options
What this is? Split a story adding a new user interface by the various platforms thar are applicable

when to use it? When adding a new user interface that may be accessed by multiple different platforms

Questions - which platforms are supported? Are all platforms required? Are some platforms more difficult to support than others.

Example: As a customer I want to create an order so that i can get something to eat

As a customer I want to create an order on my android device
As a customer I want to an order on my iPad
As a customer I want to an order on my iPhone

Business rules
What this is? Split a user story so that complex business rules are handled by separate business rules. A special case of this is deferring input validation to a later story.

When to use it? When a story has a variety of business rules, often identified by a large number of examples.

Questions - What rules apply to this story. Are all of this rules necessary. Can simpler rules suffice.

Example: As a customer I can use a credit card to pay for my order.

As a customer I want to know when I didn't provide a properly formatted CC number.
As a customer I want to know when I didn't provide a proper expiration date.
As a customer I want to know when I didn't provide a proper expiration date.

Role
What this is? Split a story so that the functionality is implemented at different times for different roles.

When to use it - When a story impacts multiple roles, and the impact is different for each role.

Questions - What roles are involved in this story. Are any roles necessary now.

Example: As a customer I want to create an order, so that I can get something to eat.

As a customer I want to create an order
As a restaurant owner, I want to create an order for a customer.

Defer system qualities
What this is? Split the story to deliver the necessary functionality, then include extra stories to improve performance, scalability, usability or precision.

When to use it? When the base functionality does not exist, implementation is not that hard, the team can learn a lot from it, and the hard work is in making it better.

Questions. Is it necessary to have the optimal performance now?

Example: As a customer I want to know that the restaurant received my order so that I'm sure I'll get some food.

As a customer I want to know that the restaurant received my order.
As a customer I want to know that the restaurant received my order within 5 seconds.

Spikes
What this is? Split the story for some research and investigation on functionality before implementing.

When to use it? When the team finds they are uncertain about the implementation of a story and they need to do some research.

Questions. What are the 1-3 questions you have about the story.

Example: As a customer I want to see the menu in the language of my choice so that I can pick something I want

Spike: Investigate translation functionality
Spike: Research on performance optimization

Low Fidelity/High Fidelity
What this is? Split the story into gradual increase of quality.

When to use it? When getting to the optimal level of quality or usability is too expensive to deliver a solution immediately.

Questions. What is good enough for this functionality

Example: As a customer I would like a recommendation of what food to order so that I can know I'm getting a good dish.

As a customer I want to see a numeric rank of top selling menu items.
As a customer I want to see the aggregate ranking of customers.

Transient then persistent
What this is? Split the stories based on actions to pass data along and those necessary to save data.

When to use it? When the story deals with functionality that includes storing data but does not require storing data.

Questions. Do we have to store this data

Example: As a customer I want to use a credit card to pay for my order so that I don't have to carry cash.

Collect credit card info and use to gain authorization but not store it.
As a customer, I want to save my credit card information to my profile.

Dummy then Dynamic data
What this is? Split stories to first use static data and then add stories to make the data dynamic.

When to use it? When the story covers parameters where the values differ depending on conditions.

Questions. Does the data need to be dynamic.

Example: As a customer I want to search for menu items.

As a customer I want to search for menu items
As a customer I want to search for menu items less than $5 or over $5.
As a customer I want to search for menu items by food type.

Manual then automated
What this is? Split stories that utilize an existing manual process. Implement the functionality that does not exist.

When to use it? When functionality you are developing interacts with an existing manual process, develop the functionality but tin the short term, do not automate the manual.

Questions. Is it necessary to automate the manual process right now.

Example: As a customer I want to use multiple credit cards to pay for my order.

As a customer I want to use a credit card to pay for my order.

Defer Error Handling or logging
What this is. Split a story so that you focus on functionality first, then go back and provide functionality to log errors and handle them

When to use it. When you need to gain feedback on functionality quickly to determine whether to keep it.

Questions. Is it necessary to deal with exception for this functionality.

Example: As a restaurant owner I want to authorize the customer's credit card payment

As a restaurant owner I want to know when the authorization process encountered errors.
As a restaurant owner I want to know what errors the authorization process encountered.