AndonX is the on-site WatchTower that enables your operators and technicians to reduce downtime and increase their productivity.
The challenge
Blind spots between the PLC, the WMS and the binder on the shelf.
When a line stops, the answer is spread across four places — and the clock is running.
Source 01
PLCs & SCADA
Live tag values and alarms, visible only from the control room.
Source 02
WMS & ERP
Orders, inventory and work orders — blind to what the equipment is doing.
Source 03
Shared drives
Manuals, schematics and SOP PDFs nobody can find at 2 a.m.
Source 04
Senior technicians
Recovery steps that live in people’s heads — until they retire.
AndonX puts all four on one screen — and answers in plain English.
The platform
One on-site AI layer for your equipment, systems and documents.
Your site’s WatchTower: ask a question in plain English and get the answer from your equipment, your systems and your own documents — on one screen, without leaving the floor.
Asset intelligence
Know the state of every asset.
Asset data
Machine data
Alarms & notifications
Electrical & PLC schematics
Work intelligence
Turn that state into action.
Conversational agent
Incident management + RCA
Wiki — SOPs & manuals
Forecasting agent
Reporting & analytics
IoT
Edge data sources
Sensors, gateways and MQTT.
OT
Control systems
PLCs, SCADA and historians.
IT
Business systems
ERP, WMS, WES and MES.
OEM
Vendor equipment
AGVs, AMRs, AS/RS and robots.
Context
Context from across your site, so answers fit how you run it.
AndonX connects data, systems, knowledge and context so AI raises your team’s productivity instead of guessing.
Knowledge
Electrical schematics, manuals, SOPs and RCAs — a digital record of your physical site.
Assets
Specifications, documentation and lifecycle information for every physical asset.
Systems
Integrations with business systems including WMS, ERP and MES.
Operations
Site-specific context on people, processes and tools.
Telemetry
Continuous time-series data and locations from sensors, machines and control systems.
Events
Alerts routed to the on-call engineer to troubleshoot and restore service.
Controls
Automation rules and AI/ML models that guide system behavior and decisions.
Processes
Flows, activities, parameters and operator instructions for operational tasks.
Modules
Four modules. One data layer.
Start with the conversational agent and live PLC data, then add incident response and the wiki as you grow.
Module 01
Andy, the conversational agent
Trained on your equipment, knowledge and systems to answer questions in natural language.
Key benefits
Answers in seconds, not screens — one question replaces a hunt across WMS, SCADA, historian and docs.
Zero training overhead — no query syntax or dashboards to learn.
Institutional knowledge on tap — SOPs and runbooks answerable by anyone on shift.
Guardrails block unsafe or non-compliant content.
Use cases
“What’s the temperature on the picking line, and has it trended up this week?” — the latest tag value, the trend and the threshold.
“@Docs what’s the lockout procedure for the dock leveler?” — answered strictly from your manuals, with cited spans.
Module 02
Real-time PLC tags & alerting
Visualize and monitor PLC tags and sensors in real time — on the floor or on your phone.
Key benefits
Real-time telemetry — hundreds of tags streaming from every connected PLC.
Find tags by meaning — semantic search instead of memorizing tag names.
Semantic context — units, descriptions, alarms and status on every raw value.
Alarm notifications — PagerDuty pages the on-call team by SMS, phone or email.
Use cases
Connect Allen-Bradley, Siemens, CODESYS, Beckhoff or any OPC UA source.
Stream OT data to Databricks or Snowflake (coming soon).
Module 03
Wiki
A knowledge base for SOPs, manuals and schematics that live beside the equipment they describe.
Key benefits
One source of truth — structured Spaces instead of shared drives.
Written together — real-time co-editing with full page history and rollback.
Your data never leaves site — self-hosted and air-gap capable.
Role-based access — control who can author and publish.
Use cases
Search IO points and wiring references on PLC and electrical schematics.
Step-by-step recovery runbooks for VFD failures or cold-chain excursions.
Module 04
Incident management & RCA
A digital Andon cord that streamlines response and turns every incident into post-incident learning.
Key benefits
Less load on the incident commander — channels, docs and meeting links provisioned automatically.
Pages the right SME based on incident type, priority and description.
From outage to operational in minutes, not hours.
Low reporting friction — small blips get the same structured treatment as major outages.
Use cases
An automatic timeline from chats, decisions and system events, ready for an objective post-mortem.
A searchable catalog of past incidents, services and key contacts.
On-prem at the edge
Your data never leaves your building.
AndonX runs on one industrial server with a GPU inside your own network — built for IT and OT teams who have been burned by cloud tools touching operational data.
Data provenance
Every PLC tag, transaction and document stays on the on-prem server. No production, IP or safety data leaves the plant network.
Low latency
Inference runs on a local GPU, so answers and alerts return in seconds — and keep working if the internet link drops.
No per-user or per-question billing
Models run on hardware inside your building, so usage stays unmetered as adoption grows.
Deployment
How long does it take to deploy AndonX?
Live on your floor in about 12 weeks — on the PLCs, WMS and ERP you already run. No rip-and-replace.
01
Weeks 1–8
Design & engineering
Site survey and data-source inventory
Integration design for PLCs, WMS and ERP
Server hardware specified and ordered
02
Weeks 9–12
Install & acceptance
Server installed on your network
FAT, SIT and UAT sign-off
Operator and technician training
03
Week 12
Go-live
Agents live on your floor
Pilot success metric baselined
First runbooks captured in the Wiki
04
Ongoing
Operate & expand
Remote monitoring and support
Results measured against the baseline
More sites at the same per-site rate
Why AndonX
Why not the alternatives?
The options a mid-market site has today — and what only AndonX brings together.
AndonX
DIY ChatGPT or Claude
Enterprise platforms
CMMS
Plain-English answers
Yes — grounded in live PLC data, SOPs and past incidents.
Fluent, but not grounded in your SOPs — it can guess at procedures.
Limited — built for data teams, not the floor.
Basic assistants, without live machine context.
Live PLC & SCADA data
Yes — live tags across PLC vendors.
No — technicians paste in screenshots and exports by hand.
Yes, after a long integration — often tied to one vendor.
No — work orders, not live PLC data.
Where your data lives
On your server, inside your network.
Pasted into an outside AI service — sensitive data leaves your network.
Often vendor cloud or hybrid.
Vendor cloud.
OT-aware knowledge base
Schematics, SOPs and manuals linked to assets and past incidents.
None — context is re-pasted into every chat.
A separate document system.
Documents stored, but not linked to live machine data.
Incident management & RCA
Pages the right SME, logs the timeline and builds the RCA.
None — no timeline, audit trail or runbook.
An add-on module.
Strong work orders, without machine-data context.
Built for mid-market
Per site — every operator included.
Billed per seat or per token — costs climb with every question.
Enterprise contracts, out of reach for most sites.
Mid-market pricing, but a narrower scope.
Pricing
Per site. Unlimited users.
Annual pricing is configured to your site. Every tier includes on-prem deployment, unmetered usage and implementation.
Essentials
One site
The core platform for a single site.
Andy conversational agent
Wiki for SOPs, manuals and schematics
Real-time PLC tags and alerting
Standard · Recommended
Standard
Everything in Essentials, plus incident response and forecasting.
Incident management
Root-cause analysis
Forecast agent
Enterprise
Multi-site
For networks of sites under one account.
Multi-site roll-up
Portfolio dashboards
Central knowledge management and dedicated support
Included at every tier
On-prem deployment · No usage metering · Implementation, FAT/SAT and training · Unlimited users on site
01 · Pilot
One site, one success metric, fixed scope and fixed fee.
02 · Prove
Measure downtime and time-to-repair against the pre-pilot baseline.
03 · Expand
Add sites at the same per-site rate once the metric holds.
04 · Standardize
Move to Enterprise for portfolio-level reporting across your network.
FAQ
Frequently asked questions
What controls engineers, IT and operations leaders ask before a pilot.
What does AndonX integrate with?
The systems you already run — no rip-and-replace. PLCs from Allen-Bradley (ControlLogix and CompactLogix), Siemens, CODESYS and Beckhoff plus any OPC UA source, usually via an edge gateway publishing over MQTT or Ignition Edge. SCADA and historians; WMS, ERP and MES systems such as SAP EWM, Manhattan and Oracle WMS; AGVs, AMRs and AS/RS; PagerDuty for alerting; and REST and MCP for everything else. PLC tags are read, never rewritten.
Where does our data live?
On one industrial server with a GPU inside your own network. PLC tags, transactions and documents never leave the site, and AndonX keeps working if the internet link drops.
What does adding users cost?
Nothing. AndonX is priced per site, not per seat, so every operator, technician and supervisor on site is included.
How long until we see value?
About 12 weeks from kickoff to go-live at one site, covering design, installation, acceptance testing and training.
We just implemented a WMS. Why add AndonX?
AndonX doesn’t replace your WMS. It sits beside it and connects it to the PLC telemetry, documentation and incident workflows the WMS can’t see.
Can Andy make up a procedure?
No. Knowledge answers come only from your own manuals, SOPs and wiki, with the source cited. Guardrails block unsafe or non-compliant responses.
Who maintains it after go-live?
We do. Remote monitoring and support are part of the subscription, and your team controls content, permissions and integrations through role-based access.
Does it replace our technicians?
No. Agents take the lookups, reports and paging off your team’s plate, so experienced people spend their time fixing equipment instead of hunting for information.
What does a pilot look like?
One site, one agreed success metric, fixed scope and fixed fee. We measure downtime and time-to-repair against your pre-pilot baseline before you expand.
Which operations is it built for?
Warehouse and distribution first — 3PLs, e-commerce fulfillment, food and beverage and automotive or pharma distribution — plus manufacturing sites running the same PLC-driven equipment.
Next step
Book a 90-day pilot at one site.
One site, one agreed success metric, fixed scope and fixed fee. We measure downtime and time-to-repair against your baseline — then you decide.