Today
Anyone who bids for contract logistics gets a mountain of data from the customer: an item master with dimensions and weights, a year of order lines, receipts and shipments, plus a price sheet in the customer's own format. Before anyone knows whether the business pays, a logistics planner spends weeks cleaning spreadsheets, finding peak weeks, counting pallet locations and estimating headcount. When time runs short, peaks and special cases get missed, and the price carries the risk for the whole contract term.
An example
An importer of garden furniture and barbecues invites Nordhanse to bid for its warehousing in Germany: 3,400 items, 14 months of order lines and around 520 containers a year through the port of Hamburg. The agent finds 410 items without dimensions or weight and fills them in from the containers' packing lists, flagged for review. It shows the stock peak in February and the shipping peak in April at three times the annual average and calculates 6,400 pallet locations, about 7,500 m² of warehouse and 22 staff on two shifts for the season. Because only 5,200 locations are free in Nordhanse's own warehouse in February, it adds an overflow warehouse and fills in the importer's price sheet with Nordhanse's rates. Costing checks the assumptions, and management releases the bid.
Your data, wherever it lives
Today the data for a contract logistics bid is scattered: the customer's item master, order lines and shipments as spreadsheets and exports in a mailbox, your performance figures in the warehouse management system, your cost rates in the ERP and earlier bids in folders. demi reads, understands and links them per item and process, with your access rights. The AI calculates storage locations, staff, space and the price sheet in days instead of weeks, and costing and management check and decide.
Where the data lives today
- E-mails
- Office files
- Databases
- ERP
- Folders & drives
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 tender in the mailbox, the customer's data extracts as spreadsheets, your warehouse management system, the ERP and earlier bids on the file share. demi replaces nothing and links the customer's item master, order lines, receipts and shipments with the performance figures from your warehouses and your cost rates. Each item then sits in one record with dimensions, weight, throughput and storage type.
- 2 · PermitPermission Layer
Only the bid team sees the customer's data, because it reveals their volumes, suppliers and buyers. Only costing sees cost rates and margins. The bid goes out only after management releases it; the release is recorded.
- 3 · ActIntelligence Layer
The agent cleans the customer's data, shows gaps such as missing dimensions or weights and derives stock levels, peak weeks and orders per day. It sizes storage locations, staff and floor space with your performance figures, such as lines per hour or unloading time per container; prices come from your cost rates, not from the language model. The language model only explains assumptions and risks in the bid text.
How it runs
- 1
The tender arrives with the customer's item master, order history and price sheet.
- 2
The agent checks the data, flags gaps and shows volumes, peak weeks and the seasonal curve.
- 3
From stock and throughput it calculates storage locations per storage type, floor space and staff per shift.
- 4
With your cost rates it fills in the price sheet in the customer's format and writes down the assumptions.
- 5
Costing checks assumptions and prices, and management releases the bid.
What changes
You know in days, not weeks, whether the business pays
Peaks and data gaps show up in the bid, not in operation
Every bid uses the same performance figures and rates
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
+259 hfreed per year
Assumptions and calculation
demi sizes storage, staff and price from the customer's data with your rates; you check and release.
24 h → 16.8 h per case
How we calculate
3 contract logistics bids × 24 h × 30% × 1.0 (manual) × 12 months = 259 h
Value per year, before demi costs
Freed per year+259 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
+$16,835
gained per year
259 h ÷ 1,760 h per role = 0.1 full-time roles
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 does the agent need from the customer?
At least the item master with dimensions, weights and packaging units, a year of order lines and the goods receipts; stock levels and shipments make the calculation more precise. Format and layout do not matter: demi reads spreadsheets, exports and PDF lists. If something is missing or does not add up, the agent shows the gap and suggests the question for the customer.
Where do the performance figures for staff and space come from?
From your own operation: lines and pallets per hour, unloading times per container and locations per rack type, taken from your warehouse management system and your planning spreadsheets. Where you have no figure yet for an activity, costing sets it. Every figure in the bid shows where it comes from.
Can the agent calculate variants, such as a different rack or a third shift?
Yes. You change an assumption, such as narrow aisle instead of wide aisle, a third shift or an overflow warehouse, and the agent recalculates locations, staff, space and price. The variants sit side by side, with the differences in cost and risk.
Who decides on price and assumptions?
Costing and management. The agent calculates and writes down the assumptions, such as peak factor, performance figures and filled-in data, so that they appear in the bid and later in the contract. The bid goes out only after release by an authorized person.
