AI can build a lot. Here's what it can't.

We use AI tools ourselves. They’ve made our team faster and our clients’ budgets go further. But there are problems AI genuinely can’t solve, and knowing the difference is what 25+ years of experience looks like.

What AI does well

We’ll be straight with you. 

If your problem is a brochure website, a simple booking form, or an off-the-shelf ecommerce store, an AI tool or a low-code platform might genuinely be the right answer. We’ll tell you that before you spend money with us. 

Speed to a first version : AI tools can produce a working prototype in hours. For validation, early-stage testing or simple content sites, that speed is real and valuable. 

Low upfront cost : For businesses that aren’t yet sure what they need, low-code tools reduce the cost of getting started. Sometimes that’s the right call. 

Routine content and copy : Drafting, summarising, templating, AI handles these tasks well. We use it in our own workflows to save time on the repeatable stuff. 

Standard feature builds : Contact forms, standard auth flows, basic dashboards. If the requirement is clearly defined and widely solved, AI scaffolds it quickly.

Where it breaks down

The ceiling is lower than the demos suggest.

Systems that need to talk to each other 
When your platform needs to integrate with legacy software, third-party payment processors, compliance systems, or proprietary APIs, AI tools hit a wall fast. The complexity isn’t in generating code. It’s in understanding the architecture of what already exists and what can’t break. That requires human expertise and institutional knowledge. 

Compliance and regulated environments 
Reuben LIMS - our laboratory management platform, operates in environments where data integrity, audit trails, and regulatory compliance are non-negotiable. No AI tool understands the stakes the way a specialist does. When something goes wrong in a regulated environment, “the AI built it” is not an acceptable answer. 

Platforms built to last a decade 
AI-generated code optimises for now. It doesn’t architect for scale, for the team that will maintain it in three years, or for the business requirements that don’t exist yet. Theatre Tokens has been our partner for years. That relationship, and the platform we built together, would not exist if we’d optimised for speed over architecture. 

When ownership and accountability matter 
If something breaks at 11pm on a Friday before your biggest sales day, who do you call? AI tools have no accountability. Low-code platforms have support queues. We have a phone number, and clients who’ve been calling it for over twenty years. 

Novel problems without existing patterns 
AI is exceptionally good at pattern-matching against what already exists. When your business needs something genuinely new - a gifting platform built around a specific licensing model, a LIMS configured for an unusual lab workflow - you need people who can think from first principles, not interpolate from training data.

25 years in practice

The proof is in the relationships. 25 years building platforms. 22+ years with our longest client.

“Reuben have been our digital partners here at Theatre Tokens for many years. They provide fantastic support, innovative solutions, and contribute positively to our brand reputation.” 

Jamie Snelgrove - Society of London Theatre 

The honest framework

How to decide.

Before you talk to us - or anyone - here’s a simple test. If most of your answers land in the left column, a DIY tool might genuinely serve you (although we don't mind taking the strain if you still want help). If they land in the right, it’s definitely worth a conversation.

Still unsure which side you're on?

A 30-minute call costs nothing. We'll tell you honestly whether we're the right fit - or point you somewhere that is.