The Iceberg

AI is the tip. Our expertise is the 90% underneath.

An AI demo can now be put together quickly. That's the part above the waterline, the part everyone sees, and it has become cheap to produce. Whether it works in your business, keeps your data safe and is still running in five years depends on what's underneath.

We call our approach the Iceberg. AI workflow is the tip, about 10% of the work. The rest is below the surface, and it's where our experience counts.

AI workflow

What people see

The waterline

AI development

Models, search over your documents, agents and integrations

Best practice

Architecture, code review, automated tests and documentation

Security

Data boundaries, access control and threat modelling

Knowledge

Data, integrations, legacy systems and high availability

Experience

How your business really works, and what goes wrong

About 90% below the waterline

Below the waterline

Experience

We've spent decades building systems that businesses depend on. That experience helps us spot where a project is likely to go wrong, often before anyone writes code.

We start by learning how your business works: the people, the processes and the exceptions nobody wrote down. Sometimes AI isn't the right tool yet, and we'll tell you.

Knowledge

AI has to work with your existing systems and data. We've built big-data platforms, multi-user applications handling millions of records, and integrations that can't be allowed to fail. We know how to modernise legacy systems without breaking what already works.

Security

We decide what AI is allowed to see and do before we build anything. That means private, ring-fenced AI instances with strict data boundaries, where only public or pre-approved data reaches the model. We also set access controls, decide where data is stored and assess how the system could be misused.

Glusoft is Cyber Essentials certified. Read more about security and privacy.

Best practice

AI can write code quickly. Experienced engineers still need to review it and take responsibility for the architecture. Every project gets code review, automated tests, documentation and a record of the architecture decisions, so your team can take it on without depending on us.

At the waterline: AI development

This is where AI gets built into your systems: choosing models, searching your own documents, agents that carry out tasks, and integrations with the tools you already run. Because it sits on the layers below, it fits your processes and stays inside your data boundaries. See our AI solutions.

The tip: AI workflow

The last 10% is how AI is used day to day, by your team and by ours: assistants, automation and AI-assisted development. It's quick to add and looks good in a demo, but it depends on the systems, data controls and engineering underneath.

Start below the waterline

Every engagement starts with a review of your business, your systems, your data and your security, before we quote for any build. You get a written report with our recommendations, which is yours to use whether or not we build anything. See consultation and analysis.

Start below the waterline

Tell us what you want AI to do for your business, and we'll start with what's underneath.

Get in touch