Introducing Vibe Kanban: A Unified Dashboard for AI Coding Agents

IT experts are recommending AGILE methodologies for software development. In parallel we have the hype on Artificial Intelligence. Please show me on paper or a whiteboard a typical AI Requirement and how AGILE helped develop the utility fast.Time and money is immaterial at this point of time. The question is can AGILE methodology be used for development of Artificial intelligence solutions.Show me examples, walkthroughs anything that makes it interesting.

Many of today's software "Best Practices" and formulaic processes are SYSTEM 1 thinking. It appeals to the System 1 brain, because it is repeatable, reducing thinking to repetition, creates immediate short term dopamine, short term visible feelings. It glosses over the deeper thinking aspects of the job. CI/CD, TDD, Scrum - these are all surface layer thinking, it may dip into system 2 occasionally, sometimes kicking and screaming, but it is predominantly the automation of the mind. We will have a stand up, we will write a test to overcome that fuzzy feeling in the mind when we have to think harder. If you want real results, you need to turn to System 2 Domain knowledge, Architecture and Design, Over vs Under-Engineering, vision, alignment, values, culture, deep thinking about getting things right. Teams that focus on system 1, are also automating their mind. Yes, there are cases, where this is blurry and of course sometimes you just cant get away from having to think, but the intent is there - automate it, get it out the way, so that you can automate you, get you, out the way. AI, is just another system 1 tool. It is immediately rewarding and people love a system 1 tool.

This article on the pendulum swing between agile & over documentation and how it fits with AI Code agents and the future of coding is spot on (and short read)

🚀 Back to the (agile) basics — even when working with AI coding agents. I recently realized that my feature and design docs had become too technical too early — and the results often missed what I actually wanted. The fix? Returning to what we figured out in Agile Development years ago: User stories and acceptance tests. By focusing on what the system should do and how we’ll know it does it, my collaboration with coding agents became smoother, faster, and — most importantly — outcome-driven. In my latest blog post, I share how this shift reshaped my process (and even improved agent behavior). 👉 Read it here: https://lnkd.in/gYAPJtfE Shoutout to all my agile peeps (Sam Newman, Alistair Cockburn, Angela Martin, Gabrielle Benefield, Peter Epstein, Joe Walnes, Kent Beck, Michael Feathers, Rachel Davies, Jeff Patton, Bradford Cross, Dave Astels, Paul Hammant, Jutta Eckstein, Nicola Dourambeis and others): are you doing the same when working with coding agents? #AgileDevelopment #UserStories #AcceptanceTests #AICodingAgents #SoftwareEngineering #EngineeringLeadership #AgenticWorkflows

Are We Going Full Circle in Software Development? For decades, software development has swung like a pendulum: *Waterfall Era: “Specs first.” Code was expensive, so we invested heavily in detailed specifications. The mantra: measure twice, cut once. *Agile & XP Revolution: Enter the Agile Manifesto (2001). Suddenly, “code is cheap, specs are expensive.” We embraced iterative delivery, minimal upfront design, and rapid feedback loops. *DevOps & CI/CD: Gene Kim and others championed “always be delivering.” Automation made code even cheaper to produce and deploy. Continuous Integration and Continuous Delivery turned speed into a cultural norm. Specs? Often sidelined in favor of “working software over comprehensive documentation.” Now, Enter the Agentic Era With AI and autonomous agents writing code, the economics have flipped again. Code isn’t just cheap—it’s practically free. But here’s the catch: If you don’t tell the AI what to build, it will happily build the wrong thing—faster than ever. Specs are no longer a bureaucratic artifact; they’re the control surface for your AI-driven software factory. The quality of your specification determines the quality of your outcome. Hallmark Moments That Got Us Here * 1975: The Mythical Man-Month taught us that adding people to a late project makes it later. * 2001: Agile Manifesto flipped the script on documentation vs. working software. * 2013–2016: Gene Kim’s The Phoenix Project and The DevOps Handbook cemented the “always deliver” mindset. * 2025: Agentic systems tip it all on its head—working software is trivial; working software that solves the right problem is the challenge. Where Do We Go From Here? We’ve gone from “specs are cheap, code is expensive” → “code is cheap, specs are expensive” → and now, in the age of AI, specs might be everything. So here’s the question: If code is nearly free, how should we rethink the role of specifications in guiding AI-driven development? Is spec-driven development the next big shift? Or something else entirely? Hashtags for reach: #SoftwareDevelopment #Agile #DevOps #AI #AgenticEra #DigitalTransformation #FutureOfWork #Leadership

https://lnkd.in/g4KBmBqX Forget Co-Pilots. The next leap in software engineering is the Autonomous SDLC. A team of 16 specialized AI agents, from Architect to UAT to Resiliency, are poised to automate, coordinate, and accelerate development like never before. This isn't just faster coding—it's faster development cycles, higher quality, and zero-friction deployments. Dive in to see how the agent-driven revolution works.

Embedding GenAI across our full SDLC - here's what actually worked in my experience (and what didn't): What Worked: - MCP integration (GitHub + Atlassian) for auto-generating architecture docs from legacy code - AI Agents building applications from requirements - GenAI-enhanced Playwright and unit testing improving coverage and speed - Governance framework enabling safe adoption Result: Improved process efficiency What Didn't: - Letting teams adopt AI tools without standards (created chaos) - Assuming AI would "just work" without deeper human involvement - Skipping the governance conversation (security/compliance issues emerged) The lesson: AI augmentation needs engineering discipline, not just enthusiasm. What's your experience with AI in SDLC? What's working for your teams? #AI #SDLC #EngineeringLeadership Here's an interesting article on AI-augmented SDLC. I would add Claude code to the development platforms.

The New Way of Software Development... *** SPEC DRIVEN DEVELOPMENT: HOW AI IS REWRITING THE SDLC *** For years, software teams lived in a code-first world. Developers wrote, shipped, and patched, fast. But somewhere along the way, clarity disappeared. Documentation drifted. Specs fell out of sync. And AI coding tools only made that chaos faster. AI doesn’t fail because it’s weak. It fails because our instructions are vague. When intent isn’t captured precisely, AI fills the gaps, and that’s where risk lives. The fix isn’t more prompting. It’s a new way of building. Spec-Driven Development (SDD) makes specifications the single source of truth. You describe what you want and why it matters. AI agents handle the how: generating architecture, tasks, code, tests, and docs - all aligned to your intent. This shift is massive: 1. 50–80% faster implementation (mobileLIVE, 2025) 2. Traceable, maintainable systems 3. Safer AI workflows with zero spec drift The future engineer won’t be typing lines of code. They’ll be authoring intent. And AI will do the rest. Follow for more on how AI for SDLC

Source: https://lnkd.in/djYf2fSF 🚀 AI + Context Engineering = Smarter Project PlanningDid you know AI can help engineers plan projects by dynamically analyzing team data? The article shows how to use context engineering with templates (like GitHub repos) to prioritize tasks, reallocate resources, and mitigate risks—without manual guesswork. 📊 💡 Example: Use JSON data from Jira to auto-generate prompts for sprint planning.📌 Pro tip: Treat prompts as code! Version control, test them, and reuse across teams. Want to boost your engineering workflow? Try this approach today. 💬 Let me know how it works for you!

Qodo: Agentic code review and integrity testing before & after merge - Qodo provides a code quality platform that coordinates AI agents to map out the workflow context and specific change requirements of each pull request in order to shift testing and code review left within an iterative DevOps release process. https://lnkd.in/eiJfiQCq