All people, companies and projects shown are synthetic Rillsoft sample data – from the shipped industry packages or, where stated, from a test project.
30 maintenance orders result in 117 operations. RES-0100 needs only 98.6 of 765 available hours overall. Nevertheless, two hourly buckets have a combined staff-capacity shortfall of 24.3 hours; daily totals hide this gap.
Source data and preparation#
Prerequisites are Connect SAP and Rillsoft in one AI client and the preparation in Reproduce the Rillsoft SAP demo step by step.
AI and you together: choose production or maintenance. Confirm the plant, work centres, source period and any date shift; for production, also establish whether lines are interchangeable. Show the mapping before creating anything. Use a separate Rillsoft instance with a new, empty project and a copy of the resource pool. Obtain specific confirmation before discarding an open project or saving a file.
Maintenance: people per operation and role#
The six tools already include MaintenanceOrder and MaintenanceOrderOperation, work centres and capacities. No second gateway setup is needed. The tool screenshots in Set up SAP MCP Gateway for Rillsoft also show these two GET endpoints.
| SAP data | Rillsoft mapping |
|---|---|
| Maintenance order / operation | Subproject / task with original SAP dates in its notes |
Work centre with staff capacity 002 | Role and team; headcount from CapacityNumberOfCapacities |
| Headcount without employee names | Placeholder people such as “RES-0100 Technician 01”; explicitly not employees read from SAP |
People per operation NumberOfCapacities | Role demand with this quantity and 100% utilisation |
| Net duration and people | Occupied duration = SAP duration × factor; if duration is zero but work is positive, first use duration = work ÷ people |
| Shift and breaks | Calendar on team, people and task; factor per work centre from the median of comparable SAP dates |
| Operation sequence | Do not invent dependencies: MaintOrderOpRelationship is absent from the six tools; preserve SAP start dates |
The reference factors are 1.0625 for RES-0100 and AMS-100, and 1.0 for AMS-200. SAP schedules AMS-200 without a break allowance even though its capacity includes a break. Do not apply the battery formula to every case. The run creates ten placeholder people for RES-0100 and five each for AMS-100/AMS-200. RES works Monday–Friday 06:00–14:30; AMS works 08:00–16:30, without an additional break window.
Read employee_capacity, grouping ["role"], display: "absolute", capacity: true, with the time filter and hourly scale. Supply follows the people’s calendars. Unlike machines, staff do not need a report per project calendar. Demand includes the break allowance here; reverse it per task to reconcile net work.
Subproject dates from their tasks: a subproject created over MCP is fixed and keeps its finish – without finish, the finish of the project, even when its tasks end earlier. After creating all tasks, therefore read rillsoft_project_tree_get and set each subproject with rillsoft_subproject_update: start to the earliest and finish to the latest date of its tasks. This matches the subproject context command “Take over”, which takes the dates from the activities; there is currently no MCP tool for that command. The task dates must not change.
Rillsoft Project screenshots#
Screenshots taken on 23 Oct 2026 from Rillsoft Project 10.0.626 (the test build), client area 1920 × 1080, effort in hours, time filter as stated; the numbers in the image belong to that view and may differ from the report totals in the table – each caption states where its number comes from.



Shortfall column and hourly report: the program’s column calculates at minute resolution and shows 31.92 h for RES-0100 on 13 Nov 2026; the report on the hourly scale sums the same peak to 24.3 h because it averages demand and supply per whole hour (recalculated from the operation dates on 23 Oct 2026: 31.92 h). Both figures describe the same bottleneck; the report states the scale it uses.
Reference results and timescale#
These figures refer to the 20 October 2026 retrieval. SAP may change its sandbox. Reference values are a comparison, never numbers for the AI to substitute into a different run.
| Check | Maintenance: staff |
|---|---|
| Plant / period | 1010; SAP August 2022, shifted +222 weeks → 11–23 Nov 2026 |
| Orders / operations | 30 / 117 |
| Net reconciliation SAP / Rillsoft | 173.5 / 173.8 h work; tolerance 1.2 h |
| Dependencies | None; relationships absent from gateway |
| Occupied time or demand / supply | RES-0100 98.6 / 765 h |
| Shortfall: hour / day / week | RES-0100 24.3 / 0.0 / 0.0 h |
| Peak concurrent demand | RES-0100 up to 48 of 10 people; SAP dates: 39 |
| Other roles | AMS-200 37.0 h shortfall; AMS-100 0.0 h |
| Date comparison | 77 of 85 comparable operations finish as in SAP (±1 min) |
For maintenance, the capacity filter runs from 11 November 2026 00:00 to 24 November 2026 00:00. SAP gives 28 operations zero duration despite positive work; Rillsoft receives the calculated duration. This explains part of the increased simultaneous headcount and is not an unchanged transfer of dates.
Shortfall = sum of positive differences between demand and supply in each time bucket. Day and week allow spare time within their bucket to offset demand. Thus the 24.3 h staff shortage at RES-0100 disappears entirely in daily totals. Start with hours for detailed scheduling, then add day/week for an overview. Spare capacity in a weekly total does not establish a conflict-free schedule.
Reconcile before reporting and restore the starting state#
Compare task, order and dependency counts and net time or net work against SAP. Also check that the start and finish of every subproject equal the earliest start and latest finish of its tasks. Where Rillsoft duration includes a break allowance, reverse that factor per task before comparing net values. Document rounding and tolerance before judging the result: less than 0.05 h difference for batteries, at most 1.2 h for the maintenance reference. An excess or rejected dependency requires a discrepancy message, not a completed success report. Restore the view and original pool and verify restoration. Save project/pool files only to the confirmed destination.
Frequently asked questions#
Can Rillsoft plan staff from SAP maintenance orders?#
Yes. The reference run transferred 117 operations from plant 1010 and assigned the people required per operation as role quantities. Despite only 13% overall utilisation, RES-0100 showed 24.3 hours of shortfall in two hourly buckets. Staff supply follows the people’s calendars.
Where do the people come from if SAP provides only a headcount?#
The 20 people in the reference pool are labelled placeholders: ten for RES-0100 and five each for AMS-100/AMS-200. Only headcounts and shifts come from SAP. Assigning actual employees requires a confirmed mapping to your own resource pool.
Reproduce and compare#
Reproduce the Rillsoft SAP demo step by step · Check SAP production order capacity with AI
Sources and trademarks#
Demo evidence: Rillsoft runs of 20 October 2026 for battery manufacturing, maintenance in plant 1010, line 26XR1 and four lines, each using Rillsoft Project 10.0.626. This guide was compiled from that evidence on 22 October 2026; the figures were confirmed on 24 October 2026 in the free prompt run (German and English, through to the report).
This example uses the MCP Gateway in SAP Integration Suite, which SAP describes as its intended access path for external AI agents connecting to SAP systems. The SAP API Policy, particularly section 2.2.2, and the terms and access restrictions of the services used still apply. The demo only reads sandbox data and does not write changes back to SAP.
SAP, SAP S/4HANA, SAP BTP and SAP Integration Suite are trademarks or registered trademarks of SAP SE or its affiliates in Germany and in other countries. This guide is published by Rillsoft GmbH and does not represent SAP certification or partnership.
Back to Connect Rillsoft Project to SAP.
Reproduce it yourself
Get started- Product version
- Rillsoft Project 10.0.626; SAP Integration Suite Trial