What You'll Learn
The Three Levels of AI Maturity
Ashraf Murtada, former HMRC Deputy IT Director and founder of Evidentia AI, has identified a critical framework that separates companies experimenting with AI from those building their entire operations around it. His three-level model provides a practical roadmap for understanding where you are in your AI journey—and more importantly, where you need to go.
Level 1: AI Assistance — Beyond Basic Q&A
Most people start with AI by asking simple questions and getting simple answers. But Level 1 maturity means demanding specific, structured deliverables. Murtada demonstrated this with a NASA report breakdown example: instead of asking "What's in this report?", an AI-mature approach demands "Extract all technical specifications into a structured table with columns for component, specification, and performance metrics."
This shift from conversation to specification fundamentally changes how you interact with AI tools. You're no longer having a dialogue—you're delegating concrete tasks with defined outputs.
Level 2: AI Agents — Context-Aware Automation
The second level introduces autonomous agents that understand context and take action without constant prompting. Murtada demonstrated Team Brain, an AI agent that joins meetings, listens to discussions, and automatically generates Product Requirements Documents (PRDs) and user stories based on what it hears.
This represents a fundamental shift from tool to teammate. The AI isn't waiting for instructions—it's participating in your workflow, understanding the broader context of your project, and producing documentation that would traditionally require hours of post-meeting work.
The key insight here is chaining: one AI output becomes the input for another process. Meeting notes become PRDs. PRDs become user stories. User stories become code specifications. Each handoff multiplies your productivity.
Level 3: AI Coding — The Superpower for Non-Technical Builders
The third level is where the paradigm truly breaks. Murtada presented multiple live demonstrations:
Space Invaders in 2 Minutes: A complete, playable game built from a simple prompt, demonstrating that even complex interactive applications are now trivial to create.
SpecSnap SaaS in 6 Weeks: A full commercial product built by a single person in six weeks—a timeline that would have required a team of developers and six months just two years ago.
CRM System in 48 Hours: A complete customer relationship management system built over a weekend, customized to specific business requirements.
Internal Tools in 6 Hours: Complex business tools that would have been multi-month development projects compressed into a single workday.
These aren't prototypes or proofs-of-concept. They're production-ready applications being deployed and used by real customers.
Software Development as Undifferentiated Work
Murtada's most provocative claim is that software development is becoming "undifferentiated work"—a commodity that no longer provides competitive advantage on its own. When anyone can build a CRM in 48 hours, having a CRM isn't special. When you can spin up a custom tool in 6 hours, buying off-the-shelf SaaS often doesn't make sense.
This has profound implications for the traditional SaaS business model. Why pay monthly subscriptions for software that only meets 80% of your needs when you can build something perfectly tailored to your workflow in less time than it takes to onboard your team to a new platform?
The Era of Disposable Software
We're entering an era of hyper-customized, disposable software. Need a tool for a three-month project? Build it in an afternoon, use it, discard it. Requirements changed? Rebuild it in a few hours rather than wrestling with configuration settings or waiting for feature requests.
This disposability isn't wasteful—it's efficient. The cost of building has dropped so dramatically that the overhead of adapting, maintaining, or working around limitations in existing tools often exceeds the cost of just building exactly what you need.
The Widening Gap
The uncomfortable truth Murtada presents is that the gap between AI Aware and AI Native is widening rapidly. Companies that understand this shift are moving with unprecedented speed. Those still thinking in traditional software development timelines are being left behind by competitors who can validate ideas, build products, and iterate in a fraction of the time.
Moving Through the Levels
Progression through these levels isn't automatic. It requires deliberate practice and mindset shifts:
From Assistance to Agents: Start demanding structured outputs. Stop asking questions, start specifying deliverables. Learn to chain outputs into inputs.
From Agents to Coding: Overcome the psychological barrier that you need to "know how to code." Modern AI coding tools like Claude, Cursor, and Windsurf provide the superpowers Murtada describes—the ability to describe what you want and have it built.
From Coding to Native: Build AI into your core workflows, not as an add-on. Make it your default approach to solving problems, not something you turn to occasionally.
The Practical Reality
This isn't theoretical. The demos shown at the Create With Manchester meetup were real applications, built by real people, solving real problems. SpecSnap is a commercial product with paying customers. Team Brain is being used in actual business meetings. The internal tools are running production workflows.
The technology is here. The question is whether you're willing to make the mindset shift required to use it effectively. Are you AI Aware—understanding that AI exists and occasionally using it—or are you becoming AI Native, building your entire approach around what's now possible?
The gap is widening. Which side are you on?
