Output recorded on paper
Keyed into Excel in the evening, so numbers reach management a day late.
MES — Manufacturing Execution System
An MES wired to real machines,with Local AI built in
We connect the plan in your ERP to the machines on the floor and turn production data into numbers management can act on immediately — built from real projects across many plants and industries.
Local AI works across every level
Level 4
ERP / finance / sales
Business planning, monthly to weekly
Level 3
MES — manufacturing execution
Dispatch, track and measure by shift and minute
Level 2
SCADA / HMI
Machine monitoring and control
Level 0–1
PLC · sensors · machines
Real data from the floor
Where factories lose money
Six leaks we find again and again on site surveys
Machines stop, nobody knows why
Downtime goes unrecorded — you see the target missed, not where the time went.
No OEE figure
Without one number for real machine effectiveness, bottlenecks are found by gut feel.
The plan lives in one person’s head
A huge spreadsheet that takes hours per cycle — when that person is away, planning stalls.
POs and BOMs keyed again and again
The same data is re-typed several times, and one mistake ripples through the line.
Old, mixed-brand machines
SCADA licences are expensive and nobody wants to connect the older equipment.
None of this is fixed by hiring more people — you need one system that pulls data from the machines and gives every department the same numbers.
TW MES stack
Eight modules — start where it hurts most
All modules share one database. Roll them out one at a time and extend later, with data flowing to and from ERP / WMS through APIs.
+ AI 01
Planning & scheduling
Capacity-based plans
Plans built on machine capacity, with an AI co-pilot that explains the reasoning.
+ AI 02
Orders, BOMs & work orders
Order-to-production
AI reads POs, explodes the BOM and dispatches jobs to each department.
03
Production execution
At the machine
HMI at the machine: choose the lot, start and stop, confirm the job.
04
Monitoring & OEE
Real-time
Output per hour, machine status and downtime by cause.
+ AI 05
Quality & vision
Defects
Defect logging plus AI vision that inspects parts at line speed.
06
Inventory & intralogistics
WMS · AMR
WMS and AMRs feeding the line, with poka-yoke against wrong picks.
07
Energy & utilities
Energy monitoring
Meters from several brands and energy cost per unit produced.
08
Reports & traceability
Audit trail
Management dashboards, LINE alerts and an audit trail of every action.
OEE & real-time monitoring
Real-time OEE — see exactly where machine time goes
Overall Equipment Effectiveness = Availability × Performance × Quality
A Availability
Run time ÷ planned time
90%
P Performance
Actual speed ÷ standard speed
95%
Q Quality
Good parts ÷ total parts
98%
OEE for this machine
83.8%
Each factor looks high, yet together they give 83.8% — calculated automatically for every machine and shift.
At the machine (HMI / tablet)
- Choose the job, lot and lot size before starting
- Start–Stop buttons; run and stop time counted automatically
- Machine stops → pick the cause from a button, no writing
- Record defects by type on the spot
- Input–output counted from sensors / PLC, not by hand
In the office (web / TV / LINE)
- Output per hour by machine and product
- Machine status: Run / Standby / Breakdown / Stop
- Downtime tracking — time lost by each cause
- Quality report — defects by cause, machine and day
- LINE alerts when a machine stops, wherever you are
Real screens
Real screens from MES systems we have delivered
Examples from several plants — from line-side OEE to production planning with an AI that explains itself.
OEE Dashboard
Real-time machine effectivenessPlant-wide OEE
How it helps: Management sees the whole plant’s performance on one page.
How it works: Data from PLCs and sensors feeds OEE, Availability, Performance and Quality automatically, with a 24-hour trend.
Status of every machine
How it helps: Know instantly which machine has stopped or lost connection.
How it works: Green = running, grey = stopped, with a Connection Loss alert when the signal drops. Works on a TV at the line.
OEE per machine + top 5 causes
How it helps: Shows which machine and which problem to fix first.
How it works: Compares OEE for every machine in one table and ranks the most frequent defect and downtime causes.
MES Real-time
Lines, raw-material silos and downtimeLine performance + live silo status
How it helps: Packing lines and raw-material silos on one screen.
How it works: OEE per machine coloured by status, plus the fill level of each silo and when it is expected to run out.
Real-time line layout
How it helps: See machine status on the actual plant layout.
How it works: Each machine is placed on the layout, with the line’s OEE, Availability, Performance and Quality alongside.
Downtime report
How it helps: Shows where the time goes so downtime can be cut where it matters.
How it works: Splits stoppages by cause and ranks the machines that stop most often and longest — daily, weekly, monthly or yearly.
Production planning + AI co-pilot
An engine schedules by the rules, AI explainsThe schedule the system builds
How it helps: Plans days ahead in seconds instead of a giant spreadsheet.
How it works: The engine sequences filling, blending, packing and cleaning for every silo within all the rules — the result must have zero hard violations.
Ask the AI why the plan looks this way
How it helps: Planners don’t have to guess the reasoning — they just ask in Thai.
How it works: The AI reads the collected rules and know-how and answers citing the rules it used. A person makes the final call.
The scheduler’s decision trace
How it helps: Check what a plan is based on before it goes live.
How it works: Summarises batches scheduled, silos in use, rules checked and warnings, with the utilisation of every silo.
Screens from delivered systems. Customer names, product names, document numbers and other confidential data are hidden.
Connect real machines
Old or new, any brand — we connect it
No ripping out machines and no SCADA licences — we pull data from what the plant already has.
Machines / lines
- • Packers · palletizers
- • PLCs · counters · sensors
- • Power meters
Shop-floor gateway
- • WiFi access points
- • RS485 → Ethernet
- • Touch-screen HMI
MES server on site
- • Central database
- • Permissions + audit trail
- • Local AI
Users
- • Web / tablets
- • TV screens at the line
- • LINE alerts
Protocols and devices our team has connected on real projects
Siemens S7-300Siemens S7-1200Mitsubishi FX3UMitsubishi Q-seriesModbus RTU / TCPRS485 / RS232MQTTOPC UASECS/GEMDry contact / DISchneider PM5350 · PM2x00ABBCircutorMitsubishi ME96Weight checker
From a real project: 40+ PLCs and devices across a warehouse and packing lines, plus 7 power-meter models from 4 brands, over plant-wide WiFi.
MES + Local AI
Explore our Local AI What Local AI adds to an MES — our core strength
The AI runs on servers in your plant, learns from your data, and production data never leaves the company.
AI turns POs into work orders
Reads POs from images or PDFs, explodes the BOM and dispatches jobs by department — no re-keying.
Planning co-pilot
An engine schedules within every rule, then AI explains the reasoning and helps with exceptions.
Vision quality checks
Checks shape, characters and surface defects on the line in real time.
Ask the MES in plain language
“Why did line 3 stop yesterday?” — answered with numbers and sources, within the asker’s permissions.
Experience map
Hands-on experience in every module, across industries
Each row is a project our team has delivered or is delivering (shown by industry, customer names withheld).
| Project | Planning | BOM | Dispatch | OEE | Quality | Intralogistics | Energy | Local AI |
|---|---|---|---|---|---|---|---|---|
| Petrochemical logistics — MES + IIoT | · | · | · | · | ||||
| Petrochemical — AI planning co-pilot (2026) | · | · | · | · | · | · | ||
| Chemicals — OEE on 20+ machines | · | · | · | · | · | · | ||
| Medical devices — order-to-production + WMS | · | · | · | · | ||||
| Medical devices — enterprise knowledge AI (2026) | · | · | · | · | · | · | · | |
| Electronics — MES + AMR | · | · | · | · | · | |||
| Machinery — AMR intralogistics | · | · | · | · | · | · | · | |
| Consumer goods — WMS + AMR | · | · | · | · | · | · | · | |
| Vision — food · medical · agriculture | · | · | · | · | · | · | ||
| Energy · aviation · automotive — monitoring | · | · | · | · | · | · |
Module delivered or being delivered on a real project Uses AI running on the customer’s own servers
What changes
How a plant changes with MES + Local AI
Task
Before
With MES + Local AI
Recording output
Paper, keyed into Excel in the evening, a day late
Machines count automatically, on screen every minute, on your phone
Machine stops
Found when someone walks past, cause never recorded
Cause chosen at the HMI, downtime report produced automatically
Machine effectiveness
No shared number, bottlenecks by gut feel
OEE for every machine and shift, bottlenecks in plain sight
Customer POs
Keyed line by line, BOMs exploded by hand
AI reads the PO, explodes the BOM and dispatches jobs
Production planning
A huge spreadsheet in one person’s head
Rule-based engine + AI that explains every decision
Feeding the line
Staff walk parts over dozens of times a day
The MES sends an AMR to refill before stock runs out
Shop-floor knowledge
Buried in email and people’s heads
Ask the AI in Thai, get answers with sources, within permissions
Traceability
Searching through stacks of paper
Every job has a time and an owner, found in seconds
How we deliver
Start with one pilot line, then scale
Every phase has clear deliverables and acceptance criteria, with plan-versus-actual tracked weekly.
-
Step 1
Site survey
Confirm installation points, test signals and machine protocols, and assess safety risks.
-
Step 2
Blueprint
Design the process, production rules and screens together with the real users.
-
Step 3
Connect machines
Install gateways, HMIs and sensors; wire PLCs and meters into the MES.
-
Step 4
Build + test
Backtest against real historical data before going live.
-
Step 5
UAT + training
Users test it themselves, with Thai-language manuals for every role.
-
Step 6
Go-live + support
Go live with close support, then roll out to the next line.
Why start with one line
- Real results fast, with a clear before-and-after
- The floor team gets comfortable before the rollout
- Real numbers from line one make the case for the next
Ready to upgrade your factory?
Book a free consultation with our engineering team.