Buying AI took an afternoon. Building the capability is the work.
Workshops and training for organisations that already have the licences and want their teams doing something substantial with them. Run on your own workflows, by engineers who build production AI systems for a living.
Most teams are still using AI like a chatbot
Rollout is the easy half. A licence lands on everyone’s desk, and then three things tend to happen at once. None of them are anybody’s fault, and all three have the same cause.
Things get built that nobody owns
Someone automates a job that genuinely needed automating. It works, a few colleagues start relying on it, and it is not in anyone's architecture diagram. That is not a discipline problem. It usually means the sanctioned route was slower than going around it.
The output comes back needing rework
Work arrives that takes longer to fix than it would have taken to write in the first place. Almost always the workflow was scoped wrong rather than the person being careless. A general-purpose tool was handed over without a definition of what good output looks like for that specific task.
People use the tool you did not approve
When the approved path is missing or slow, people find another one. Every instance is a fairly precise specification for something worth building properly. Most organisations never get that information this clearly.
Three symptoms, one cause. The company bought AI access and has not yet built AI capability. Access is a purchasing decision that takes an afternoon. Capability is engineering work, and it is the part that is usually still outstanding. Closing that gap is what these sessions are for.
Three ways to start
In rough order of commitment. Plenty of people come to a breakfast session first and decide from there, which is a sensible way to do it.
A day on your own work, not a generic curriculum
The reason most AI training does not stick is that it is taught on someone else’s examples. People leave impressed and then cannot map any of it onto Monday morning. So we build the session on your workflows before we turn up.
Thirty minutes to find two or three real workflows from your team that are worth the room's time. If we do not find any, we will say so.
Built beforehand on your processes so people can see the finished thing and how it was assembled, rather than a toy example.
The part that changes how people think about the tools. Hands on keyboards, on their own work, with us in the room when it does not behave.
What got built, what is safe to keep using, what needs engineering before anyone relies on it, and what we would do next. Useful whether or not you work with us again.
Teams that already have AI tools available to them. It works equally well for a function that has barely started and one where people are already building things quietly.
No engineering background needed. We have run these for operations, finance, legal, marketing and HR teams as well as technical ones.
Not sure which of the three you need?
That is a reasonable place to start. A short call is usually enough to work out whether a session with your team is the right next step, or whether something else is.
Read what we think first
A workshop is a big ask on trust. These are free, and they will tell you fairly quickly whether we are worth a conversation.
AI at scale, from the inside
Conversations with the people running AI programmes in large organisations. What actually worked, and what they would do differently.
A practical guide to AI agents in the enterprise
Where agents genuinely help, where they are the wrong tool, and what has to be true before one runs unsupervised.
Building AI that ships
The distance between a promising prototype and something a business can depend on, and why most of the work sits in that gap.
More in resources, including past sessions and case studies.
People who build this software, not people who teach it
Applied Edge is the AI practice of Applied Blockchain, which has been engineering privacy-first software for regulated industries since 2015. Members of the team are Anthropic Claude Certified Architects, and the workflows we demo are ones we run inside our own company across project management, finance, marketing, sales and HR.
The Claude & Croissants series has brought together senior people from the Bank of England, Citi, Shell, bp, Lloyds Banking Group, NatWest, Deutsche Bank and others.
Archax
Production work in a regulated environment, where the compliance requirements were part of the design rather than a review at the end.
CryoPDP
Operational software delivered end to end, from the first scoping conversation to something the business runs on.
The team behind it
Who we are, how the practice is structured and the standards we hold ourselves to, including ISO 27001.
