Zaq Mughal

Brighton (UK) based software engineer with over a decade of industry experience focussing on building robust and scaleable applications for mission driven organisations.

Head of Engineering at Genius Within CIC.

Can AI Code Generators Really 3x Dev Team Productivity?

Published on:

We’re firmly in the era of vibe coding and with AI code generation tools growing in popularity, development teams are all asking the same question: can these tools really deliver the 2x, 3x, even 4x productivity gains we keep hearing about?


Now let’s be honest — AI is everywhere, especially in tech. It’s evolving fast and organisations are faced with a choice: either embrace it or risk falling behind.


At Genius Within, our digital team has been trialling GitHub Copilot and other AI code assistants like Cursor. There’s no doubt these tools are changing how we write code. They can autocomplete boilerplate, suggest entire functions and translate vague instructions into structured logic. It’s bloody impressive.


But here’s the thing: being a developer isn’t just about writing code. In fact, coding itself often makes up a surprisingly small portion of our time.


For example, a typical project in our department involves:

  1. 1. Discovery & requirements gathering
  2. 2. Design
  3. 3. Build
  4. 4. Testing & QA
  5. 5. Deployment
  6. 6. Post-deployment support

And when you zoom into the developer’s day-to-day, it looks more like this:

  1. 1. Team stand-ups and planning
  2. 2. Code reviews
  3. 3. Stakeholder conversations
  4. 4. Onboarding new tickets (and context switching)
  5. 5. Investigating logs or production issues
  6. 6. Actual “hands-on-keyboard” coding

When you lay it out like that, it’s clear: the time we spend actually writing code is just a fraction of the job.


So while AI-powered tools do help us code faster, they only support one part of a much bigger process. That’s why expecting game-changing productivity gains from AI just by using code generators is likely to fall short.


And this isn’t just true for developers. Across most roles, the doing part is only one piece. The real time goes into thinking, testing, aligning, reviewing, and communicating.


If we want to truly unlock the full potential of AI (those genuine 2x, 3x, 4x gains), we need to stop focusing solely on code generation. Instead, we should be exploring how AI can streamline our entire workflow: discovery, designing, testing, reviewing, planning, and above all, communication.


#VibeCoding #AIDeveloper #DeveloperProductivity #FutureOfWork