Mar 26, 2026 · 12:00 - 13:00 GMT+1
Webinar: AI-Powered Software Development from the Trenches
Online
How far can you actually take AI assistance in a real development workflow? Most devs use AI in one way or another, but greatly underestimate what's possible.
Watch recordingEvent
information
In this webinar Henrik Kniberg demonstrates how the Abundly dev team uses AI assistance in all aspects of development - beyond just coding - and ships a new version of their platform every day. He will share lessons learned, concrete tips, and some thoughts on where software engineering as a craft is headed.
Want to go even deeper? See also Henrik’s full day masterclass on April 16.
What you'll experience
- •Concrete examples using Cursor, Claude Code, and Abundly
- •When and how much to use AI assistance
- •How to balance agent autonomy with human control and oversight
- •Helping your AI navigate large complex codebases
- •How to maintain quality and avoid technical debt
- •Using AI agents beyond just coding — for things like backlog management and release management
- •What this means for team structure, processes, and the software engineering craft in general.
- •Q&A
Who should attend
- Professional developers
- Technical leaders with a background in coding (it is OK if you are super rusty)
- People who used to code once upon a time, and want to get into it again now that the threshold is so much lower
- Engineering managers and tech leads who want to understand how AI changes the software development workflow
Speakers
Henrik Kniberg
Co-founder and Chief Scientist, Abundly AI
Recording
1h
Watch recording
1h
Enter your details to access
Transcript
Webinar Summary: Henrik Kniberg on AI-Powered Development
March 26, 2026
- AI coding has gone through three major shifts — from copy-paste with ChatGPT, to AI embedded directly in the IDE (Cursor), to autonomous agents handling full-stack tasks end-to-end — each step dramatically increasing developer output.
- The engineer's role is being redefined — less time writing code, more time on architecture and design. This isn't a loss; it's an elevation, enabling smaller teams to build more, iterate faster, and reduce technical debt.
- Abundly built internal agents (Backlogger, Releaser, Grace) to automate their own workflows — automating ticket creation, daily releases, and even retrospectives. Grace, a high-level orchestration agent, writes code via Cursor's cloud API and self-improves through reflection.
- Context is the new bottleneck, not prompting — providing AI with the right context (rules files, architecture docs, system descriptions) matters far more than crafting perfect prompts. Context engineering is the emerging skill for developers.
- The right mindset is "superpowers, not fear" — companies that encourage AI adoption to increase productivity will win. The risk isn't AI replacing developers; it's developers who refuse to adapt becoming obsolete.