Our AI Lab
The Lab is our research and development. We build products, frameworks, methods and ways of working with AI, and we run them in our own business before we offer them to anyone else. What survives that goes into the work we do for you.
AI changes every few months. New models, new tools and new ways of working arrive faster than anyone can read about them. Most of what is said about them has never been tried on real work.
So we try it on ours. Before we recommend a model, a tool or a way of building, we have used it on our own products, our own helpdesk and our own engineering. We measure what it does. If it doesn't earn its place, we don't recommend it.
Built and run by us
Sign2Approve, our e-signature and approvals product, is where we test AI in a product people depend on. AI is used only where it earns its place, and every model call is logged and costed.
Proven on our own work
Reusable architectures for AI agents, with the hard parts solved once: retrieval over your documents and code, confidence scoring, passing to a person when unsure, and testing on every change. Our support agent ran on the group's own helpdesk before anyone else saw it.
Published in full
The Converge Methodology, and three books on how AI is adopted, built and governed: CONVERGE, ENGINEER and GOVERN. Each is tested on real engagements and written up so you can inspect it before you buy.
How we build
How software gets built when a coding assistant writes most of it and an engineer decides, reviews and tests. Sign2Approve was built this way, and so was this website. The practices we learned are the ones we set up for engineering teams.
For everyday work
Assistants configured for the recurring work in a function, with its templates, tone and rules built in. We work out which tasks pay back, build the assistant, and refine it until its output needs finishing rather than rewriting.
Beyond language models
Not every problem needs a chatbot. We research forecasting and optimisation on real business data, such as adaptive demand forecasting for replenishment and AI-optimised purchasing between suppliers and buyers.
Most experiments never leave the Lab, and that is the point. Every idea is tried on a real problem of ours, measured against a test set agreed before the work starts, and run with real users. Only what proves itself reaches a client.
When we recommend a model, a tool or a product, we have used it, or something close to it, on our own work.
Our frameworks start with the hard parts already solved, so your build starts further on and is tested from the first day.
The workflows we set up for your teams are the ones that raised quality or productivity in ours.
Write-ups from the work, including what we would do differently.
Running an AI agent on our own agency's helpdesk before offering it to anyone else taught us things a client project would not have. The decisions we would repeat, and the ones we would make earlier.
Every AI consultancy names a framework. Almost none shows one. We publish the Converge Methodology in full, with its artefacts, because a method you can inspect is one you can trust, and because the value is in the running of it, not the secret.
Agentic coding tools change how software is built, but only when the project instructions, standards, review gates and budgets are right. The setup we use ourselves, and recommend to clients.
The difference between a licence and a tool people use is configuration: assistants set up for the recurring work in a function, with its templates, its tone and its rules built in. Six we configure for sales, and the pattern behind them.
An agent that cannot be measured is not ready for production. What an evaluation set is, how to build one with your own experts in a few hours, what to measure, and why it has to run on every change.
The engineering story behind our own helpdesk agent: why a generic chatbot could not do the job, how code and documentation are indexed so retrieval is grounded in the actual build, and the decisions that made it safe to put in front of customers.
It is the research and development side of Enceladus AI. It is where we build and test products, frameworks, methods and ways of working with AI on our own business first.
No. The same people who run your engagement do the Lab's work, so what we learn there reaches your project directly.
No. The Lab works on our own products, systems and data. Client work is done under the terms agreed with each client.
Yes. The frameworks are the starting point for the AI systems we build, the methods run every engagement, and the workflows are part of our training and tool rollouts.
Tell us where AI could help, and we'll tell you what we have seen work.