As a business coach helping SMB leaders build great companies, I've spent 35+ years in technology - from software developer to COO to consultant. Today, I'm seeing a transformation that will impact every business, not just tech companies. The leaders and teams who understand and adapt to these changes will have exponential advantages over those who don't.
Everyone is debating whether AI is ready to replace software developers.
I think there’s a more important question we should be asking.
Instead of focusing on AI capabilities, we should be asking:
How do we redesign our development processes to work with AI?
For the past 20 years, software development teams have maintained remarkably stable role structures. Whether you're using Scrum, Kanban, Lean UX, or some other methodology, the atomic unit of product development has remained essentially unchanged:
These roles evolved around a fundamental constraint: building software takes time. Features require weeks or months to develop, creating natural coordination rhythms and allowing for sequential handoffs between specialists.
But what happens when AI can build features in minutes instead of weeks?
I've been experimenting with AI-augmented development using tools like Replit, and the speed gains are amazing. Features that used to take a 2-3 week sprint can now be built in minutes. Working systems can be built in hours. This isn't a minor efficiency improvement, it's a sea-change level of shift.
It also reveals that our existing processes and coordination models simply can't keep up.
When a developer can implement a feature faster than a product owner can write detailed requirements, or faster than a designer can create mockups, our carefully designed processes become the bottleneck. Traditional optimization of existing processes won't bridge this gap. We need to rethink the entire approach.
Here's a concept that has helped me think about this:
Everyone needs to move up a level of abstraction.
This isn't about optimizing existing processes for marginal gains. It's about recognizing that when implementation speed approaches zero, every role needs to operate at a fundamentally higher strategic level.
Traditional Role: Write detailed user stories, manage backlogs, prioritize features
New Role: Define strategic boundaries and success criteria, enable autonomous team decision-making
Instead of specifying exactly what to build, the Playbook Owner creates the framework within which rapid experimentation can happen safely. Think of it like establishing "trading limits" that allow teams to move fast without requiring approval for every decision.
Traditional Role: Create mockups and specifications for future development
New Role: Guide users through amazing experiences in real-time, adapting the journey based on immediate feedback
When features can be built in minutes, designers can't work weeks ahead creating static mockups. Instead, like an expedition guide, they need to be embedded with the team in the rapid build-test-learn cycle, making design decisions in real-time as actual user feedback comes in. Concepts like Lean UX and Continuous Discovery have been moving in this direction for a while now, but this new level of speed will require even faster feedback loops.
Traditional Role: Implement predetermined specifications
New Role: Rapidly explore solution possibilities within defined constraints
With AI handling much of the implementation details, developers become more like researchers, quickly testing different approaches to find optimal solutions rather than meticulously coding predetermined features.
Traditional Role: Test completed features against specifications
New Role: Design rapid validation systems that can keep up with AI-speed development
When features deploy in minutes, traditional testing cycles are too slow. The focus shifts to creating automated validation systems and real-time user feedback mechanisms. Some human validation systems will likely still be needed, but we’ll need to find ways to make the feedback loops much faster.
Traditional Role: Approve budgets, review progress in monthly/quarterly cycles, make go/no-go decisions
New Role: Establish strategic constraints and success thresholds, enable rapid decision-making within bounds
When teams can pivot directions in hours instead of quarters, traditional approval cycles become bottlenecks. Business sponsors, in close collaboration with the new Playbook Owner, will need to set clear strategic boundaries upfront and trust teams to operate autonomously within those guardrails. Regular communication channels will need to be lean, focused, and low-overhead.
Traditional Role: Design system architecture, define technical standards, review implementation details
New Role: Create architectural frameworks that guide AI-generated code, establish patterns that enable rapid experimentation, create systems to validate architectural compliance
As AI generates more code, the focus shifts from designing specific implementations to creating the foundational patterns and constraints that ensure system coherence regardless of how features are built. Working with developers and testers, the Architect will need to create systems to rapidly validate that newly generated solutions are adhering to defined platform guidelines.
Traditional Role: Facilitate sprint ceremonies, remove impediments, ensure process adherence
New Role: Design and optimize rapid feedback loops, facilitate real-time decision-making, manage cognitive load
When work completes in minutes instead of weeks, traditional dev processes, like daily standup and weekly planning, become irrelevant in their current form. Principles of continuous flow and lean processes become paramount. Ease of communication, rapid testing, and an attitude of experimentation are essential. The rapid pace and amount of information flow will create new challenges with cognitive load. The focus shifts to ensuring teams can process continuous feedback and make rapid decisions without burning out.
Moving up a level of abstraction isn't simply about doing the same work faster. It's about fundamentally different types of work that become possible when implementation speed approaches zero.
Consider how this changes product development:
AI-Speed Approach: High-level Strategic framework → rapid experimentation → real-time validation → continuous evolution → multiple loops within hours
You may recognize the Build-Measure-Learn structure, popularized by Eric Ries in his book The Lean Startup, but this happens at light speed. The new model enables true validated learning at speeds that were previously impossible. Instead of making big bets based on assumptions, teams can rapidly test multiple approaches with real users and real working software. This dramatically reduces risk, but requires a new level of focus and experimental mindset.
This transformation requires genuinely different cognitive abilities. Strategic thinking, rapid pattern recognition, comfort with ambiguity, and real-time decision-making become more valuable than detailed execution skills.
However, detailed human input and feedback are still necessary. The humans are still leading the dance, but this guidance will need to be applied strategically with speed and utmost efficiency. The need to guide, orchestrate, and control the AI dance partners will be essential.
The challenge isn't just retraining existing teams - it's recognizing that some people naturally operate better at higher abstraction levels while others excel at detailed implementation. AI-augmented development will require fundamentally different hiring, training, and personal development programs.
If you're a product development leader, start asking these questions:
The teams that master operating at higher abstraction levels will have exponential advantages over those trying to optimize traditional processes for marginal speed gains.
From a business perspective, remember that the newfound ability to rapidly test real ideas in a fraction of the time is a much lower cost, lower risk approach. However, it will require a complete mindset shift and a commitment to embracing ambiguity through an experimentation focused approach.
Moving up a level of abstraction solves the coordination problem, but it creates new challenges. If everyone is operating at a higher level, what happens to validation and feedback? How do you maintain quality and strategic coherence when everything moves so fast?
These questions point to the next fundamental shifts that AI-speed development requires - shifts in how we think about validation, decision-making, and team coordination itself.
The future belongs to teams that can think and operate at the speed their tools enable. One path to this newfound way of working is recognizing that everyone needs to move up a level.
This is the first post in a series exploring how AI-augmented development changes everything about building software. Next up: applying financial trading floor principles to coordinate teams at AI speeds.