Intelligence lives at the edge
Analysis runs on the machine and on the site. The cloud is a library and a back-up, not a vault every decision must pass through.
Our approach to edge intelligence
Local knowledge, shared wisdom.
The long-term belief behind ATF, and what it means for the organisations we work with today.
01 · What we believe
AI should work where the work happens.
AI is becoming more powerful, and more centralised. Much of it runs in a handful of large data centres, where information is gathered and controlled far from the people who created it.
We believe there is a better path for industry. AI should be specific to the machine and the process it serves, and it should work alongside the people who use it, strengthening their judgement rather than replacing it.
An old hand on every machine.
Think of the experienced engineer who knows every machine by its sound. When something new comes up, they ask a colleague who has seen it before. Nobody waits for head office to decide what is wrong. Knowledge stays local; wisdom is shared.
02 · What we stand for
Six principles that shape everything we build.
Raw data stays home
Raw operational data stays with the organisation that generates it. What leaves site, if anything, is your decision.
Wisdom is shared, not surrendered
In time, lessons learned on one site could help others, as insight rather than data, and only with each organisation’s permission.
The right knowledge at the right moment
Relevance, not volume. Insight should surface when it is needed and stay quiet when it is not, and you control the filter.
People first
AI should amplify the expertise of the people who run and maintain your assets, not reduce them to interchangeable parts of a process.
Open to inspection
Anyone relying on the system should be able to see how it reached a conclusion. Trust is earned through transparency.
03 · Where we start
Industrial maintenance: a clear test of value.
Every long journey needs a first step that can be proven. Manufacturing is data-rich and demanding, and it has a clear measure of success: a machine either fails unexpectedly or it does not.
That is why our vision is to make unplanned downtime a thing of the past, for every organisation that depends on physical assets, not just the giants.
Our platform puts these principles to work. It is designed to learn each machine’s normal behaviour, keep raw data on site and build a body of verified insight. Over time, that includes the hard-won knowledge of experienced maintenance teams, captured with their involvement and consent.
Sensing, connectivity, local monitoring, dashboards and threshold alerts are available now. Anomaly detection and predictive insight are in development and validation.
04 · How our analytics are designed
Principles for analytics people can trust.
Anomaly detection and predictive insight are in development and validation. These principles describe how they are being designed.
Learns each machine’s normal
Models are designed to be calibrated on each machine’s own healthy operation, not generic assumptions.
Reads the whole picture
Signals will be interpreted together and in operating context, so routine changes aren’t mistaken for faults.
Understands the machine’s structure
The aim is for alerts to relate to specific components, not just individual sensors.
Shows its reasoning
Established engineering checks are designed to sit alongside the learned model, so alerts can be explained.
Learns from outcomes
What maintenance teams find will be fed back, so performance is measured honestly and can improve over time.
05 · Our commitments
What this means if you work with us.
- Your data belongs to you. It is processed on site by default and is never shared without your explicit agreement.
- We will be honest about what our AI can and cannot do, and if ATF isn’t the right fit, we’ll say so.
- Insight comes with the reasoning behind it, so your team can check it and decide.
- We work alongside your people, not over their heads.
06 · The road ahead
Manufacturing is where we start, not where we stop.
The same principles apply wherever organisations depend on physical assets, from facilities management to infrastructure. Where we go next will be chosen by evidence, and by where our customers pull us.
We are stronger together: not because information is gathered and sold back to us, but because knowledge can flow between people who keep control of what is theirs.