For 25 years, I've been building web applications. I've worked inside agencies, consultancies, and directly with product teams - everything from marketing sites to complex SaaS platforms. It's been a good career. But over the past couple of years, I started noticing something that made me rethink the entire model.
The agency model is broken
When a business hires an agency to build software, the process is predictable. They meet with a smooth account manager who promises the world. Behind the scenes, the work gets handed to a team that might include a junior developer or two, a mid-level lead, and maybe a senior architect who reviews code once a week. The client pays premium rates but gets mid-level output, delivered slowly, with most of the budget absorbed by overhead - the office, the project managers, the layers of people between the client and the code.
I've seen it from the inside. I've been the contractor parachuted in when timelines slip. I've inherited codebases where it's obvious that the person who built them was learning on the job. And I've watched clients pay £50,000 for work that should have cost £15,000.
AI changed the economics
When AI coding assistants started being useful enough to matter - not the "autocomplete on steroids" phase, but the phase where they could reason about architecture and write production-quality code with proper guidance - I realised the game had changed.
A single experienced developer, working with AI tooling, could now match the output of a small team. Not by cutting corners, but by eliminating the busy-work: the boilerplate, the repetitive patterns, the routine code that takes time but not thought. The parts that agencies staff with juniors.
This isn't about AI replacing developers. It's about AI making experienced developers far more productive. The key word is "experienced" - you need to know what good architecture looks like to direct AI towards it. You need to recognise when generated code is subtly wrong. You need the judgment that comes from decades of shipping real software.
The thesis behind PNK Digital
So I asked myself: what if you combined senior-level expertise with AI-accelerated workflows, stripped out all the agency overhead, and passed the savings on to clients?
That's PNK Digital.
In practice, it looks like this:
- Faster delivery - I can ship in weeks what used to take months, because AI handles the repetitive work and I focus on the decisions that matter.
- Lower cost - No office, no middle management, no account managers. You pay for development, not infrastructure.
- Higher quality - every line of code is written or reviewed by someone with 25 years of experience. You won't get passed to a junior or lose context in handoffs.
- Direct communication - You talk to the person writing your code. Questions get answered in hours, not days.
It's a bet, and I'm confident
Starting PNK Digital is a bet that the market is ready for this model. That businesses are tired of paying agency rates for mediocre output. That founders and product managers would rather work directly with a senior developer than navigate layers of project management.
I could be wrong. But every conversation I've had so far suggests I'm not.
If any of this sounds familiar - if you've been burned by agencies, or you're sitting on a project idea but can't justify the traditional cost - I'd love to talk. No pitch, no hard sell. Just a conversation about what you're building and whether I can help.