AI is changing how we write code. Should it also change our Definition of Ready and Definition of Done?
25 Jun 2026
I have been thinking about this topic recently. Most teams I've worked with follow Definition of Ready(DoR) and Definition of Done(DoD).
DoR helps reduce rework and acts as a checklist before the work even starts mostly to determine whether a story or a feature is ready to be picked up for example questions like:
DoR helps reduce rework and acts as a checklist before the work even starts mostly to determine whether a story or a feature is ready to be picked up for example questions like:
- Is the story/feature small enough?
- Are acceptance criteria clear and testable?
- Are dependencies understood? and so on
And DoD is like a shared agreement on what a finished actually means for the team which may have some questions like :
- Is the feature live in production?
- Are monitoring and logging in place?
- Are the tests implemented and passed? and so on
As more teams are moving towards AI-assisted development and specification-driven development, I'm curious to hear how DoR and DoD are evolving in your teams? What does Definition of Ready and Definition of Done look like in your team today? Has AI influenced any part of it? Or there's no DoR and DoD needed anymore?
Parveen Khan
Quality Engineering Manager at CFC
She/Her
Hi, I'm Parveen Khan, Quality Engineering Manager passionate about software testing and quality engineering. I also enjoy sharing my experiences and stories.
Open To
Write
Teach
Speak
Mentor
CV Reviews
Podcasting
Gary Hawkes
For us nothing has changed, and nor should it in my opinion. Do any of the examples you used for DoR or DoD change just because of AI assisted development? If anything, maybe how we evidence some of those things may change.
Ady Stokes
An excellent question and one which I suspect will be heavily context dependent on what, if any, changes are needed
Ujjwal Kumar Singh
I think it should change. Just because something is generated by AI doesn't mean it should be considered done. It still needs to be evaluated against the same quality standards we expect from human-written work.
Today, AI generates code, AI reviews code, and in many cases AI even suggests fixes. If there is no meaningful human evaluation in the loop, who is accountable for confirming that the work actually meets the required quality?
AI has become a crucial part of software development, but that also means our definition of Done should evolve. Just because AI completed it shouldn't automatically mean it's ready.
Sign in
to comment
Better than a generic video, see YOUR test, live, ready to show you what matters most: quality at scale.
Explore MoT
What I learned about influence by becoming a stakeholder
Unleash the power of generative AI to boost your software testing and day-to-day tech tasks
Into the MoTaverse is a podcast by Ministry of Testing, hosted by Rosie Sherry, exploring the people, insights, and systems shaping quality in modern software teams.
