Today
Cash forecasts live in spreadsheets and go stale fast. Large projects with milestone payments make them even less reliable. Shortfalls often show up only when cash is already tight.
An example
At Talwerk Antriebstechnik the 13-week forecast shows a gap in calendar week 9. The milestone payment for order 4711-A, 24 geared motors, slips because acceptance moves back, while a large steel order falls due. The agent shows two options: split the steel order into two deliveries or draw on the credit line for three weeks.
Your data, wherever it lives
Today open orders, invoices, purchase orders and payment terms sit in the ERP, payment milestones of large projects in contracts as PDFs, payroll dates in the HR system, account movements in online banking and the plan in spreadsheets. demi reads, understands and links them into payment flows per customer and supplier, with your rights. demi recalculates the forecast continuously instead of by hand in spreadsheets, the AI explains every change with the documents behind it, and your finance lead decides.
Where the data lives today
- ERP
- PDFs
- HR system
- Web & portals
- Office files
How demi solves it
Three layers in this order: connect the data first, then the rights, then the AI.
- 1 · ConnectData Layer
Data sources in this case: the ERP, bank accounts and planning spreadsheets. demi replaces nothing and links open orders, invoices, purchase orders, payment terms, payroll dates and bank transactions. The history shows how each customer and supplier actually pays.
- 2 · PermitPermission Layer
Bank balances, open items and payment behavior per customer are visible to finance and management; sales sees only the aggregated forecast. Finance decides on measures, and every step is recorded.
- 3 · ActIntelligence Layer
demi computes cash flows week by week deterministically from payment terms, dates and recorded payment behavior, not the language model. The agent explains every change with the records behind it and runs scenarios.
How it runs
- 1
Clarify the data sources: ERP, bank accounts and planning spreadsheets.
- 2
The agent computes the forecast week by week.
- 3
Every new order, payment or delayed delivery flows into the forecast at once.
- 4
It flags shortfalls and explains every change with the documents behind it.
- 5
Finance runs scenarios and decides.
What changes
Shortfalls visible weeks ahead
Less spreadsheet upkeep
Financing decisions on solid ground
What does it bring you?
Check it with your own numbers
Every starting value is an assumption you can change. The calculation sits right below, step by step.
Your numbers
Result
+77 hfreed per year
Assumptions and calculation
demi projects cash in and out from orders, purchases and payment habits; you manage liquidity.
4 h → 2.4 h per case
How we calculate
4 cash forecasts × 4 h × 40% × 1.0 (manual) × 12 months = 77 h
Value per year, before demi costs
Freed per year+77 h
Assumption: labor cost of $48.62 per hour in US manufacturing in June 2026 (BLS), plus workplace, IT and overhead.
= Value per year, before demi costs
+$5,005
gained per year
An estimate, not a promise. The starting values are assumptions for an industrial company with 150 to 500 people.
Combine with more cases and get the packageQuestions about this use case
Which data feed the forecast?
Open orders with payment schedules, sales and purchase invoices, purchase orders, payroll and tax dates, and bank transactions. The history shows how punctually each customer really pays. Anything no system holds, such as a planned investment, you add as a planning item.
Can I run scenarios?
Yes. Ask, for example, what happens if a key customer pays a month later or an investment is brought forward. The agent recalculates and shows both forecasts side by side. The base forecast stays untouched.
How often is the forecast updated?
Whenever something changes in your data: a new order, a payment, a delayed delivery. No more manual upkeep. For every change you see the document that caused it.
Does the AI calculate the figures or estimate them?
Cash flows are calculated, not estimated: from amounts, payment terms and dates in your systems. Payment behavior per customer comes from recorded payments. The AI explains and compares; the figures themselves come from the documents.
