Manufacturing SystemsERP Platform

Asprova + private AI

AI for Asprova APS: Explaining the Schedule and Connecting It to Your ERP

Short answer

AI for Asprova adds a private assistant that reads the scheduling project and the ERP data around it, so planners can ask why an order is late, what limited a resource, and how the plan differs from ERP records. It does not replace Asprova's scheduling logic, and any proposed change remains a draft that a planner applies.

ERP: Asprova
Industries: Manufacturing, Electronics, Automotive, Process Manufacturing
Written for: Production Planner

Discuss your Asprova environment See use cases

Asprova is an advanced planning and scheduling product with a long record in Japanese and wider Asian manufacturing and with users in Europe and North America, especially in high-mix electronics, automotive supply and process industries. Its model is built around orders, items, resources, operations, operation modes and calendars, with scheduling behavior shaped by rules and, in many sites, by scripting and custom extensions.

That flexibility is the reason Asprova fits complicated plants and also the reason each installation is unique. A model may include material allocation logic, setup and changeover rules, multiple-resource operations and lot-sizing choices that were tuned over years. Explaining a result to a manager means unpacking all of that.

Many planners are comfortable inside the Gantt and the product's own reports. A private AI layer is useful for the questions that follow the schedule: why did this order slip, which constraint was binding, how do the ERP's order quantities and dates compare with the plan, and what changed since the last run. It also helps teams whose language, ERP data and documents are mixed across sites.

This page covers how such a layer connects to Asprova and the ERP, what it adds, and what to ask any vendor. It is an explanation and reconciliation tool, not an alternative optimizer.

What usually gets in the way

The problems we hear most from production planner teams running Asprova.

Heavily customized models

Scripts, custom properties and tuned rules mean no generic documentation describes your Asprova project. The knowledge sits with a few people.

High-mix complexity

With many items, modes and changeovers, understanding why a particular sequence resulted is difficult without tracing it by hand.

ERP integration drift

Imports of orders, items and stock from the ERP can silently differ from the source because of mapping or timing issues.

Multi-language and multi-site documentation

Rules and notes are written in different languages and formats across sites, which makes them hard to search or compare.

Confidential demand data

Customer forecasts and delivery dates are sensitive and may be restricted by contracts from use in cloud AI tools.

Where AI earns its place in Asprova

Each use case names the ERP objects it reads or writes, so your ERP team can judge the integration effort before anyone commits budget.

Late order explanation

Describe, for a late order, the operations, resource loads, material availability and calendar effects that caused the delay.

Touches: Orders, Operations, Resources, Items, Calendars

Outcome: Gives a consistent explanation for sales and planning meetings, with the supporting numbers.

Plan versus ERP reconciliation

Compare ERP production and purchase orders with those loaded into Asprova and list differences in quantity, date or routing.

Touches: ERP orders, Asprova orders, items, stock

Outcome: Finds stale or incorrect imports before they distort the plan.

Model and script documentation

Read project settings, operation modes and scripts to produce a searchable description of how the plant is represented.

Touches: Operation modes, resource definitions, scripts, custom properties

Outcome: Lowers key-person risk and supports migrations or rule reviews.

Changeover and setup analysis

Summarize which sequences cause the most changeover time on a resource and suggest groupings for a planner to test.

Touches: Operations, Setup rules, Resources, Items

Outcome: Points planners to realistic sequencing improvements worth trying in the scheduler.

Capacity what-if summaries

Explain the effect of calendar changes, added shifts or outages on late orders and resource load across copies of the project.

Touches: Calendars, Resources, Orders, project versions

Outcome: Speeds up capacity decisions with side-by-side evidence.

Plan versus actual

Compare scheduled times with ERP or MES confirmations and show where durations or yields differ most.

Touches: Scheduled operations, shop-floor confirmations, MES data

Outcome: Supports better standard data in the model.

Supplier and shortage visibility

List scheduled orders waiting on materials and tie them to open purchase orders and promise dates.

Touches: Purchase orders, stock, BoM, scheduled orders

Outcome: Lets buyers act on the items that actually gate the schedule.

Reference architecture

The assistant reads Asprova project data and ERP sources via read-only interfaces and returns explanations and drafts. Asprova remains the system that schedules.

1
Asprova and ERP connectors
Reads schedule data through Asprova's data export or database interfaces and ERP data through APIs or SQL views, with scoped read-only accounts.
2
Semantic layer
Maps orders, items, resources, operations and modes, plus custom properties, to the terms in use at each site and language.
3
Model serving
Open-weight multilingual models served with vLLM or Ollama on customer hardware or a private cloud.
4
Retrieval and agents
Retrieval over live data, scripts and notes; proposals are drafts for planners.
5
Governance and audit
Per-user logging of questions, queries and answers inside your boundary.

Integration notes for your ERP team

Asprova exchanges data with ERPs by file, database or interface programs; connect the assistant to the same staging data.
Use a read-only path to project data, not a session that can reschedule or save over the live project.
Inventory scripts and custom properties early, since they carry plant-specific logic that answers depend on.
Choose an open-weight model with good coverage of the languages your planners and documents use.
ERP reads use documented APIs or read replicas, keeping the ERP's role rules in force.
Where MES or shop-floor confirmations exist, read them to compare plan with actuals.
Test the connector against your Asprova version and re-test at upgrade.

Deployment options

Air-gapped on-prem

Electronics, automotive and defense suppliers with strict customer data rules.

Everything runs on site with no outbound connection.

Private or sovereign cloud

Companies with regional data rules for ERP and scheduling data.

Deployed in a tenant in the same jurisdiction as the source systems.

Hybrid

Groups with plants in several countries.

Central governance, local connectors, data kept in each region.

Compliance and data control

How the architecture supports your obligations. Certification and accountability stay with your organisation; the design keeps the evidence straightforward.

Customer confidentiality and NDAs

Forecast and order data stays in infrastructure you control rather than going to a public AI service.

Regional data protection (GDPR, APPI and similar)

Personal data stays in the region where it is held, which simplifies transfer analysis.

IATF 16949 and ISO 9001 records

Logged, reproducible explanations of planning outcomes support control of planning information.

Export control (ITAR / EAR where applicable)

You keep controlled data inside your own access-controlled environment and apply your own rules.

Where Netray fits

ERPray

Planners get grounded answers about ERP orders, inventory and purchasing next to schedule snapshots.

DataRay

Suited to mixed sources: project exports, documents, spreadsheets and databases queried together on-prem.

Custom build

Model documentation and plan-versus-actual agents are specific to each Asprova project and usually custom.

How an engagement runs

Phase 1 . 2-4 weeks

Discovery

-Inventory of Asprova version, data interfaces and scripts
-Prioritized planner questions
-Language and documentation scope
-Deployment recommendation

Phase 2 . 6-8 weeks

Pilot

-Read-only connectors on a non-production project
-Late-order explanation and reconciliation live
-Accuracy review with planners
-Logging validated

Phase 3 . 4-6 weeks

Production

-Production cutover with scoped accounts
-Runbook for new questions
-Training
-Audit log review

Phase 4 . Ongoing

Scale

-More plants and languages
-Draft-action workflows
-Model refresh
-Quarterly review

Questions to ask any vendor, including us

A short list that separates real Asprova AI work from a chatbot demo.

How does the assistant read Asprova data, and can it ever alter the project?
How are our scripts, operation modes and custom properties represented?
Which languages are supported for questions and documents?
Where does the model run, and what leaves our network?
How are answers validated against a planner's manual tracing?
How does it reconcile with our ERP and what accounts does it use?
What is the plan for Asprova version upgrades?
What do we retain if we end the engagement?

Frequently asked questions

Does AI replace Asprova's scheduling?

No. Asprova remains the scheduler, using its own rules and scripts. The AI layer explains results, compares plan and ERP data, and drafts proposals that a planner applies, so the scheduling logic and its audit trail stay unchanged.

Can it work with a heavily scripted project?

Yes, but scripts and custom properties must be inventoried and mapped, because they carry the plant-specific logic. That work in discovery determines whether answers are trustworthy.

Does it handle Japanese or other languages?

Open-weight models vary in language quality, so the model is chosen and tested against your actual documents and questions. Multilingual coverage is validated in the pilot rather than assumed.

Is it worth it if our planners already know Asprova well?

For routine work, perhaps not. The value is in explanation for managers, plan versus ERP reconciliation, and capturing knowledge so a new planner does not depend on one expert.

How does it connect to the ERP?

Through documented APIs, SQL views or a read replica with a scoped read-only account, respecting the ERP's own roles. The same staging data Asprova uses is a natural source.

Is our forecast data sent outside?

No. Model, retrieval and connectors run on infrastructure you control, so forecasts and delivery dates are not sent to a public AI service.

How long does a pilot take?

Usually two to four weeks of discovery and six to eight weeks of pilot. Heavy scripting or several languages and plants lengthen it.

Related guides

Forecasting plus a reason why

AI Demand Forecasting for Manufacturers: Better Numbers and an Explanation Planners Can Trust

AI demand forecasting explains ERP forecast changes in plain language, flags exceptions for planners, and handles short-history SKUs without Excel overrides.

ERP AI for Japan's manufacturing sector

AI for ERP in Japanese Manufacturing: On-Prem, APPI-Aligned, and Monozukuri-Ready

On-prem AI for SAP, Oracle, and domestic ERPs in Japanese manufacturing: APPI-aligned, built for the METI 2025 digital cliff and monozukuri quality.

EMS / PCBA + on-prem AI

AI for EMS and PCBA Manufacturers Running SyteLine, Epicor, or NetSuite

AI for EMS and PCBA manufacturers on SyteLine, Epicor, or NetSuite: automate BOM scrubbing, AVL checks, and obsolescence alerts, grounded in your own ERP data on-prem.

Shop floor + on-prem AI

AI shop floor assistant for operators working inside your ERP

An on-prem AI copilot answers operator questions against your ERP work instructions, travelers, and routings in plain language, at the machine, without a screen full of menus.

RAG + SQL + permissions

A Private LLM Grounded on Your ERP Data

How a private LLM answers questions on your ERP data: RAG plus text-to-SQL, role-based permissions inherited from the ERP, and where each fits.

Ask your ERP anything

Natural Language Query for ERP Data: Ask SAP, Infor, or Oracle a Question in Plain English

See how natural language query over SAP, Infor, Oracle, and NetSuite data works: grounded text-to-SQL, role-based permissions, and a visible audit trail.

Plan it with numbers

Free Tool

Talk it through with an engineer who knows Asprova

Bring one real question your team cannot answer from the ERP today. We will map the data path, the model, and where it runs, and tell you honestly if AI is the wrong tool for it.

Start the conversation See ERPray

All AI for ERP guides