Custom AI implementation
We connect your company’s knowledge, tools and people to build AI around the work they do.
We help your people find the company knowledge they need and prepare for what comes next.
We build AI agents around agreed tasks, with clear permissions and your team in control.
We help your people see work, commitments and decisions that need their attention.
More time to listen, solve problems and improve what you do.
Our process connects perspectives across business, design and AI. Technical and learning and development specialists bring the depth the project needs.
We follow the whole job with your people, then build, test and practice the changes that can give them more time for customers.
Make work more human.
Our specialists follow real jobs through your department, connect the information and tools they rely on, and help your team use the change confidently.
Before we begin
Start with recurring work that needs information or action across people and tools. Someone needs to own the change, and the people doing the work need time to help shape and test it. We start with your existing systems and check whether configuration, a process change or custom engineering makes sense.
See what we buildWe bring technical implementation and an understanding of people’s work into the same engagement. Michael connects perspectives across roles and departments; the relevant specialists help build, teach and support the change. We test it with the people doing the job and agree what a useful result looks like, who reviews it and who keeps it working.
See how an Install worksWe discuss the work, the people and a sensible next step. If the opportunity needs investigation, we agree a paid Audit before starting. It can recommend configuration, a process change or no build. An Install is a separate decision; we can discuss it directly when the scope is clear.
Explore the AuditWays to work with br8n
Useful AI starts with the people, knowledge and processes behind the work.
Understand how your company really works.
We sit with the people doing the work, trace processes and examine the knowledge held across your systems. Together, we identify what is missing, where work gets stuck and what needs to be understood before choosing an AI implementation.
Build the context. Then put AI to work.
We turn that understanding into working capability. Where knowledge is missing or scattered, we capture and organise it with your team. We connect suitable tools, build agreed workflows and test them alongside your people, with training and practice built in from the start.
Keep the knowledge and the work up to date.
We agree ongoing help with you for the systems we have built or configured: updating knowledge and connections, reviewing failed or unclear work, and helping your team adapt as processes change. Support is optional, with responsibilities and scope agreed together.
Pick a useful piece of work to start with.
Use the practical guide to identify recurring work, check the information it needs and decide where a person should review the result. Define a first project you can test before committing to a wider implementation.
Free field guide · 2027
Choose a first workflow, rehearse it with your team, and plan the review, costs and handover.
Get the free guideTell us what gets in the way—and what you want to make possible.
Field Notes and practical resources for more useful AI at work.
If I were planning training for a new AI workflow, I would start with the part of the demonstration most likely to be skipped: something going wrong.
Read the note
I want an AI implementation to give people room for useful work. To judge that, I would follow the task beyond the moment an answer appears.
Read the note
Follow one workflow from request to accepted result. Count the work people still do and choose a next test.
Read the guide
A first project should give your company a useful result and a clear way to judge the work. A broad ambition such as “use AI in operations” gives you neither.
Read the guidePractical ideas for making AI useful at work. About twice a month.
