How AI rewrites the lessons of the Agile Manifesto
Does the Agile Manifesto still make sense?
I believe it’s an open question.
Each item of the Manifesto now has more nuance with AI in the picture.
Let’s explore.
Individuals and interactions over processes and tools
What happens when the tools can reason and take action independently? What if, by defining processes, we can enable AI systems to automate and act more autonomously within a bound context and rapidly drive down costs? I still think humans will remain vital for having the conversations and creating the connections necessary to decide what to build, but the ‘individuals’ detailed in this point may no longer be exclusively human.
Working software over comprehensive documentation
It’s now incredibly cheap to generate working software - but generating software that is valuable (i.e. can be easily changed) increasingly relies on comprehensive documentation that AI tools can follow. Also, writing said documentation is now incredibly fast and cheap too. These two may become increasingly fused rather than an ‘either/or’ binary.
Customer collaboration over contract negotiation
One of the things AI really encourages and drives towards is outcome based thinking. It’s why I frame my consultancy services on value based pricing. I think of it almost as a corollary of the ‘bitter lesson’ - if we focus on the outcome, and less on the how we get there, we can use AI in increasingly innovative ways. Thus while customer collaboration might be essential, the clearer up front we are about what the expectations are the more we can leverage AI to deliver on them.
Responding to change over following a plan
Of all the principles, I think this one still holds the most value. It’s true that AI tools will perform best when given a comprehensive plan to follow, but it’s also true that the industry and landscape is shifting fast. AI will be a essential for navigating this turbulent time, but human vision and creativity will be fundamental for adapting and persevering. In the micro AI needs a plan, but in the macro we must preserve optionality in the face of uncertainty.
*
In short, AI collapses many of the binaries the Manifesto for Agile Software Development wrestled with.
We’re entering a time of fusion, where classic Waterfall methodology synthesises with agile to create a new, more powerful framework and way of creating and maintaining software.
The underpinning spirit of agile is that change is constant, and that we must continually adapt.
Agile itself is no exception from this.