Any model can read a document. The work is landing what's in it in the systems you already run: matched, validated and correct every time. Built for manufacturers and distributors.
30 minutes, no pitch deck. You leave knowing which workflow is worth automating first.
9 months
in daily production, every business day
25,000+
customer orders read, matched and posted to the ERP end to end
~130
on an average business day, with peaks above 300
Measured on the order-processing system in daily production use at a US manufacturer and distributor. Figures to the end of July 2026.
No two operations run the same way. Your software shouldn't either.
Orders arrive as email attachments and someone retypes them into the ERP. Inventory is counted on paper, then reconciled against a spreadsheet. Dealers call because there's nothing to log into. Off-the-shelf software assumes a workflow you don't have, so your team fills the gap by hand.
Not the right fit for marketing websites or mobile apps. The right fit if your team is retyping documents into your systems.
Services
Usually in this order: get the data out of the documents, move it into the systems you already run, then build what doesn't exist yet.
Document AI
Reading the document is no longer the hard part. The system matches every line against your catalogue, flags what it is not certain about, and puts a person in front of the exceptions rather than the typing.
e.g. order emails → reviewed ERP entries
Integration
Custom middleware connecting your ERP, CRM and project tools, so orders, inventory and pricing move between them automatically.
e.g. dealer orders → posted to your ERP automatically
Custom Tools
Dealer portals, warehouse queue systems, inventory workflows: production-grade applications built around how your operation actually works.
e.g. a self-serve ordering portal replacing phone and fax orders
Work
Each system below was built for a client's own stack, and is stated at the stage it has actually reached. The figures are measured, not estimated.
AI + Automation
In production since Nov 2025Result
~130
customer orders every business day, from inbox to reviewed ERP entry. Peaks above 300. Each one used to take 15 to 30 minutes of typing.
AI + Operations
In productionResult
~10 min
from scanned sheets to approved inventory, down from hours. Transcription errors gone.
ERP Integration
In rollout · Aug 2026Result
~980
dealer accounts onboarded, each with their own pricing and live stock. In rollout: the first dealers go live in August 2026.
Client names are withheld by agreement; operational systems are competitively sensitive. Architecture and results can be walked through in detail on a call.
Process
01
A 30-minute call and a folder of your real documents. No workshop, no discovery phase.
02
Something working on your own data, in weeks. You react to a system, not a spec.
03
One workflow, running in production on your systems. Monitored from day one.
04
Something wrong? Your team flags it and it gets fixed, not queued behind a support tier.
M Solutions takes on a small number of clients at a time. Depth over volume.
About
I am an industrial engineer who writes production code. The person who works out what your operation needs is the person who builds it.
I am Dr. Thomas Möllers, founder of M Solutions. PhD in business model innovation from the University of St. Gallen, visiting scholar at Stanford, industrial engineering from RWTH Aachen. Before going independent I built enterprise software for leading APAC banks and led a multi-year research collaboration with BMW. I write the code, and I am the one in the room.
Education
Experience