Simon G. by Groma

For companies that run on orders, stock, machines and cash

Your managers spend the morning collecting numbers.
The decision still comes too late.

We do the work. You take the decision.

Simon G. prepares your operating decisions every night, from the data you already have, and brings each one to the person who signs.

On your own serverNo internet neededYour data shared with no one

Prepared 05:40 · for purchasing Needs a signature

Order 12,400 kg of PE granulate from Nordfilm by Thursday.

Extrusion line 2 runs out on 8 November at the current plan. Lead time 12 days.

TODAY 8 NOV WITH ORDER
Stock today
18.2 t
Needed, next 30 days
26.9 t
Expected effect
+ 4,100 EUR avoided
Inputs missing
none

Why now: three orders for the same film confirmed this week; the supplier's last two deliveries were four days late.

Limit · class B · ≤ 15,000 EUROwner · purchasing manager

Fictional plant. An example of a prepared proposal, not a promise.

Does this sound like your week?

Six mornings you will recognise.

The people are composites. The mornings are not.

  1. Monday, 08:10. Purchasing opens the ERP, two spreadsheets and Friday's e-mail to find out what to order. By eleven the plan has changed.
  2. The plan moves three times a day. Nobody saw the material shortage until the line stopped.
  3. Stock is too high and too low at once. Forty days of what nobody ordered; two days of what everybody did.
  4. The credit limit is what the salesman said. The customer who pays at ninety days keeps getting deliveries.
  5. Three departments, three versions of the truth. They meet on Thursday to argue about which one is real.
  6. The cash forecast lives in one person's spreadsheet. When she is away, the company flies blind.

None of these is a people problem. They are what happens when decisions are prepared by hand, in the gaps between other work.

So what does that cost you?

The expensive part is not the software you lack. It is the decisions you prepare by hand.

Nobody books these hours to a cost centre, so nobody sees them. Set the figures to your company. Nothing you type leaves this page.

Management time, per year
1,725 hours
About one full-time senior person who never produces, sells or plans anything.
Cash in excess stock
548,000
Days of stock you could carry fewer, times one day of your cost of goods.
One late credit decision
12,000 a year
One customer invoiced 50,000 a month who pays at ninety days: 150,000 always outstanding, financed at 8%, before any write-off.
One lost shift a week
230,000 a year
Eight hours of one line, once a week, forty-eight weeks.
5
1.5
20,000,000
10
600

Orders of magnitude, not promises. In a pilot the first thing we do is measure yours, with a baseline fixed before anything changes.

What would Tuesday look like instead?

Same company. Same people. A different seven o'clock.

Mara runs purchasing, Tomas the plan, Ilse the money, at a plant that could be yours.

Before
Mara · 08:10

Three windows open. Builds the order list by hand. Misses the supplier who was late twice this month.

Tomas · 09:30

Reworks the week because printing has no rolls for order 4458. Nobody told him extrusion was behind.

Ilse · 11:00

Asks for the cash position. Gets it Thursday. Decides Friday. The overdraft was needed on Tuesday.

After
Mara · 08:10

One list, prepared at 05:00: twelve orders with stock, need, lead time and the supplier's record. Nine are inside her limit. Three need her signature. She enters all twelve in the ERP. Done by 08:40.

Tomas · 09:30

Reads the plan per machine and shift, built from the orders that exist. The missing rolls were flagged yesterday. The order went out.

Ilse · 11:00

Sees thirteen weeks of cash with the two largest customers' real payment behaviour. The decision on terms for customer A is prepared. She signs it.

What changed between the two columns?

Nobody worked harder. Someone prepared the decision.

Your systems record. Your spreadsheets reconcile. Your managers decide, in the time left over. That last step is where the money leaks, and it is the only step nobody has automated.

Today
  1. 01Systems recordwhat happened
  2. 02Spreadsheets reconcileby hand, between other work
  3. 03People decidelate · unrecorded · when present
With Simon G.
  1. 01Systems recordas they are, untouched
  2. 02Simon G. preparesevery night · with evidence
  3. 03A named person signs08:10 · recorded · measured

The systems do not change. The spreadsheets go. The decision arrives prepared, and the person who signs it is still the person who signs it.

How it gets there

Three steps. Nothing to prepare on your side.

No implementation project. No budget line for one. You say what you want decided better; we read your data and set it up.

1

You tell us what you want

Which decisions hurt or come too late: orders, production, credit, cash, the plan. One conversation, with the person who owns them.

Thirty minutes · no slides
2

We read your data and set it up

We take what your systems already record, ourselves, with your IT beside us for a few hours: the first export, then the same export scheduled to run each night. Your products, your limits, your people.

On your server · set up for you
3

You take the decisions

Every night Simon G. prepares them with their evidence. Inside a signed limit they need no signature. Outside it, the person you named approves or rejects, with a reason. Simon G. changes nothing in your systems: your people enter the decisions in the ERP, as they do today.

Named owner · limit · record

What lands on your desk

Six decisions, prepared by eight in the morning.

Each one with the person who signs it. Your field decides which come first.

01

Know tomorrow's production plan before the first shift starts.

Per machine and shift, from the orders that exist, changeovers counted.

The planner signs
02

See the material shortage a week before the line stops.

Needs by week against stock and open orders, with the supplier to call today.

Purchasing signs
03

Order what will sell, not what sold last year.

Proposals per supplier every night, forecast error shown honestly.

Head of supply chain signs
04

Spot the customer who will be overdue before the invoice is.

A limit from each customer's own payment record, and this week's action.

The credit controller signs
05

Know the week the cash runs short, thirteen weeks ahead.

Receivables by the date each customer really pays. Three scenarios.

The finance director signs
06

See which product, customer or plant moved the result this month.

One management result that ties to the books, variance explained.

The board reads

The question you are about to ask

"Great. Another IT project."
It isn't. Here is why.

Five questions a managing director asks before the second meeting.

How long does this take?

Four days to a written verdict on your data. Three weeks to the findings and one decision prepared on a real case of yours. Eight to twelve weeks for a pilot with your people signing real decisions. Installation takes minutes.

Will we need new systems?

No. Simon reads exports from the systems you have, spreadsheets included. It never writes back to your ERP and never asks for passwords.

Is this expensive?

A fixed-fee diagnostic, credited to the pilot. A monthly pilot you can cancel. A yearly licence per kind of decision, given in the first conversation. No implementation budget, because there is no implementation project.

Do we need consultants for months?

No. A few hours of your IT's time on one day, to run the first export with us and schedule it for each night. Then a named owner who signs, and thirty minutes a week with us during the pilot.

Will people actually use it?

Nobody learns a system. They arrive to a list of prepared decisions, a reason beside each, and an Approve button. A rejection is written down with its reason. If the reason is a rule or a limit, a person changes it, and the next proposal follows.

Your data stays yours.On your server. Not shared with other companies, not with us, not even through a chatbot.
Deterministic. No internet needed.The same inputs give the same answer, traceable step by step. It runs with the cable unplugged.
Hosted by us only if you ask.In a region you choose, under the same rules.

And then?

We measure it. Against a baseline fixed before.

Before a decision is applied, we write down the figure it should change and its value today. Afterwards we measure the same figure the same way. A result without a baseline proves nothing.

Forecast against what was sold, one product family, 26 weeks

Weekly error 14%. The gap in week 38 is a missing export, drawn as a gap, never as zero. Fictional data.

24 t16 t8 tW20W45 W38 · NO DATA DECISION
ForecastSoldBaseline fixed before the decision
Three figures we measureday 1 · week 12
Hours spent preparing decisionsCounted on day one and at week twelve, for the people who sign.
Stock days and stockout daysBoth, because one without the other is a trick.
Days to a decisionFrom the moment it could be taken to the moment it was.
If the pilot does not pay for itself in the measured effect, we say so first.

Don't take our word for it

No logos. Three things you can check.

Ask it

One decision, thirty minutes

A founder goes through one decision of yours, on your questions, and you get a written answer on whether it fits.

Talk about one decision
Read it

Seven principles, one page

The rules behind every decision Simon G. prepares, written down so your auditor and your IT can hold us to them. The full method comes in the first conversation.

The seven principles
Test it

Four free tools, in your browser

Stock, working capital, credit, decision charter. On your own file, never uploaded.

Open the free tools

Five things Simon G. does not do.

  1. It does not execute anything.It proposes. A named person decides, and your people carry it out. Nothing in your systems changes.
  2. It does not replace your ERP.It runs above it and never writes back.
  3. It is not a chatbot.An assistant may explain a proposal. It never decides or invents one.
  4. It does not share your data with anyone.Not other companies, not us, not a cloud model.
  5. It does not make you compliant by itself.It produces the evidence: who decided, on what, with what effect.

Start with one decision

Bring one decision your company takes badly or too late.

Thirty minutes with a founder. We go through how Simon G. would prepare it and say in writing whether it fits.

We do the work. You take the decision.

Or write: office@gromasystems.com
Groma S.R.L. · Timișoara, Romania · English, Romanian, French, German.

We reply within one working day, from office@gromasystems.com. No newsletter, no sequence.

Questions we are asked first

Short answers.

Where does my data go?

Nowhere. Simon G. installs on your own server, needs no internet connection to run, and shares your data with no one: not with other companies, not with us, not with a cloud model or a chatbot.

Does it replace our ERP or our planning spreadsheets?

The spreadsheets, not the ERP. It reads what the ERP records and prepares the decisions the spreadsheets were trying to prepare, with evidence and a record. It never writes back. If you plan production in Excel today, read about an Excel alternative for production planning.

Is this artificial intelligence?

The engine is deterministic: the same inputs give the same proposal, and every step can be traced. An assistant on your own server may explain a proposal in plain words; it does not decide, and no company data goes to a cloud model.

What do we need to prepare?

Nothing. You tell us what you want decided better. We take the exports from your systems ourselves, with your IT beside us for a few hours, and set Simon G. up for your activity.

What does it cost?

A fixed-fee diagnostic credited to the pilot, a monthly cancellable pilot, then a yearly licence per kind of decision and company size, given in the first conversation. How the pilot works.

Can I try something before talking to anyone?

Yes: four free browser tools on groma.ro that never upload your file.