Accelerator 06 · Manufacturing
Maintenance copilot
Gives technicians a ranked list of likely causes and step-by-step checks from machine logs, manuals and past work orders.
The problem
- Fault-finding depends on a few experienced technicians, and night shifts often do not have them.
- Manuals, past work orders and sensor history live in different systems.
- Unplanned downtime is extended by trial-and-error part swaps.
What it does
- Reads the alarm, recent sensor readings and the matching manual section.
- Ranks likely causes with the evidence for each.
- Lists checks in a safe order, starting with lock-out/tag-out, and the parts to keep ready.
- Drafts the work order entry so the fix is recorded for next time.
Pick a machine alarm with sensor readings and get likely causes, checks in order and parts to keep ready.
Live demo: Maintenance copilot
Synthetic data only · nothing you type is stored
Output
Likely causes ranked, safe checks in order, parts to keep ready and a draft work order.
The demo first calls the Surge in-house AI server and falls back to a hosted open-source model if it is busy. The badge on the output shows which one answered. Limited to a few runs per visitor per day.
How it integrates
In your deployment the accelerator runs inside your environment, behind your single sign-on, and is connected to the systems below. You choose the hosting and the model.
Systems and APIs
- CMMS / EAM (work orders, asset history)
- Historian or SCADA via OPC UA / MQTT export
- Manuals and SOPs as PDFs
- Tablet app on the shop floor, works on the plant LAN
Data we need for a pilot
- Equipment manuals
- 12 months of work orders
- Alarm code list
- Sensor tag list for target machines
Where it runs
- On-prem in the plant (no internet dependency)
- Edge GPU box per plant for larger sites
Open-source models we start with
- Qwen3-class 8B–14B instruct model (runs on a single 16 GB GPU when quantised)
- BGE-M3 embeddings over manuals and work orders
Walkthrough
A short recorded walkthrough of the maintenance copilot: the input, the output and how a reviewer uses it.
Next accelerator: Shipment exception agent →Walkthrough video coming soon
We are recording a short walkthrough of this accelerator. Meanwhile, run the live demo above or book a working session.
Run the manufacturing accelerator on your data
In a working session we map it to your process, systems and rules, and agree what a pilot must prove before it goes live.