Specialist ERPsUse Case

Sage X3 purchasing + stock AI

AI for Sage X3 Purchasing and Stock: Orders, Receipts and Lots

Short answer

AI for Sage X3 purchasing and stock reads purchase orders, receipts, supplier data, stock balances and lots through the X3 GraphQL or web service APIs. It reviews orders, tracks supplier delivery performance and answers stock questions, while approvals and postings stay in X3.

ERP: Sage X3, Sage X3 Purchasing, Sage X3 Stock Management
Industries: Food and beverage, Chemicals, Distribution, Light manufacturing
Written for: Purchasing Manager or Supply Chain Manager on Sage X3

Discuss your Sage X3 environment See use cases

Sage X3 is widely used by mid-size food, chemical and distribution companies that need multi-site, multi-company purchasing and stock control with lot tracking. Purchasing runs through requisitions, orders, receipts and invoice matching, and stock is managed by site, location, lot and quality status.

In daily use, the friction is information access. A buyer wants to know which orders are overdue by supplier, a stock controller wants all stock in quality hold for a product family, and a finance user wants receipts not yet invoiced. X3 can answer each of these, but through different screens, inquiries or custom reports.

AI can sit above those screens. It takes a question in plain language, translates it into an API query, returns the records and shows how it got there. It can also scan recent orders for oddities and watch delivery history for patterns.

It does not replace X3 workflow or its approval rules. Sage also has its own assistant features, and for simple inquiry needs they may be enough. This page describes where a separate, private layer is worth the effort.

What usually gets in the way

The problems we hear most from purchasing manager or supply chain manager on sage x3 teams running Sage X3.

Overdue orders tracked by hand

Buyers export open purchase lines and chase manually. Supplier patterns, such as always late on one product family, are not obvious.

Stock in the wrong status

Lots sit in quality control or rejected status longer than intended, blocking production or sales without anyone noticing.

Receipts not matched to invoices

Goods received but not invoiced accumulate, creating accrual noise at period end.

Customized inquiries and reports

X3 sites often rely on custom screens and Crystal-style reports that few people can modify.

Multi-site visibility gaps

Stock and demand across sites and legal entities is hard to see together when purchasing decisions need it.

Where AI earns its place in Sage X3

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

Overdue and at-risk order tracking

Lists purchase order lines past or near their expected date, grouped by supplier, with delivery history and suggested follow-up.

Touches: PORDER, PORDERQ, supplier records, PRECEIPT

Outcome: A prioritized chase list that replaces a weekly export-and-sort routine.

Supplier delivery performance

Computes on-time and in-full delivery per supplier and product using order and receipt dates and quantities.

Touches: PORDERQ, receipt lines, BPSUPPLIER

Outcome: Evidence for supplier reviews and for adjusting planned lead times.

Stock status monitoring

Finds lots in quality control, rejected or blocked status beyond a set age and summarizes value and affected products.

Touches: STOCK, STOLOT, ITMMASTER, stock status A, Q and R

Outcome: Earlier release or disposition of held stock; less hidden working capital.

Lot trace questions

Answers where a given supplier lot went, which orders and customers received it and what remains on hand.

Touches: STOLOT, STOJOU, sales deliveries, receipts

Outcome: Faster response to a recall inquiry, with the final decision made by quality staff.

Purchase order anomaly review

Flags lines whose price, quantity or unit differ from history or contract terms before approval.

Touches: PORDER, PORDERP, price lists, ITMMASTER

Outcome: Fewer errors reaching the supplier and the invoice matching step.

Goods received not invoiced analysis

Lists receipts without matching invoices by supplier and age, and suggests likely reasons such as price mismatch or missing invoice.

Touches: PRECEIPT, PINVOICE, accruals

Outcome: Cleaner period-end accruals and less manual reconciliation.

Cross-site stock visibility

Lets a planner ask for availability of an item across sites before ordering more.

Touches: STOCK, sites, ITMFACILIT, transfers

Outcome: Reduces duplicate purchasing when stock exists elsewhere in the group.

Reference architecture

A service reads X3 through GraphQL or web services using a dedicated, scoped user, stores what it needs for analysis in your environment and serves a local model that explains results. Anything beyond reading is proposed and approved.

1
API access
X3 exposes GraphQL through the Syracuse server, along with SOAP web services published for the folder; both can be used for read access.
2
Folder and site mapping
X3 data is organized by folder and site; the mapping carries legal entity, site and language into every query.
3
Analytics store
Order, receipt and stock history held in a store you control for delivery performance and trend calculations.
4
Model serving
Open-weight model hosted on vLLM or Ollama on your hardware or sovereign cloud.
5
Approval gate
Any write, such as a changed expected date, is created as a proposal and applied by an authorized user through X3.

Integration notes for your ERP team

Check your X3 version: GraphQL availability and web service options depend on the release and the Syracuse setup.
Create a dedicated user with limited function profiles; queries should inherit the asking user's rights wherever possible.
Specify the folder in every call; multi-folder environments will otherwise return the wrong data.
Table names such as PORDER and STOCK are stable, but customer-added fields (such as those with the X prefix) need to be mapped.
Use incremental extracts on update dates for history; avoid heavy reads in the posting windows used by finance.
Where X3 is integrated with a WMS or EDI, include those event sources so receipts and shipments look complete.

Deployment options

On-prem X3 with on-prem AI

Companies hosting X3 themselves

Store and model sit in the same network as the Syracuse server.

Sage-hosted or partner-hosted X3 with private AI

Cloud X3 customers

API access is arranged with the host, and processing stays in your tenancy.

Regional deployments

Groups with EU or other residency rules

Regional nodes keep local entity data local; a restricted summary can feed group reporting.

Compliance and data control

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

FSMA and EU food traceability

Lot trace queries read X3 data to support your trace procedures; the official records remain in X3.

REACH and chemical regulations

Product and lot attributes in X3 can be queried; regulatory determinations stay with your compliance team.

Financial controls

Approval limits and segregation of duties in X3 remain unchanged; AI does not approve or post.

GDPR

Supplier contacts are excluded from prompts unless a use case requires them, and data stays in your chosen region.

Where Netray fits

ERPray

Plain-language purchasing and stock questions with the query shown and X3 access respected.

Custom build

Supplier delivery models, stock status monitors and anomaly rules tuned to your items and contracts.

How an engagement runs

Phase 1 . 1 week

Discovery

-X3 version, folder and hosting review
-Top buyer and stock controller questions
-Access and security plan

Phase 2 . 2 weeks

Connect and map

-Scoped API user
-Mapping for purchasing and stock objects
-Custom field register

Phase 3 . 3 to 5 weeks

Pilot

-Overdue order tracker
-Stock status monitor
-Accuracy check against X3 inquiries

Phase 4 . Ongoing

Extend

-Supplier delivery scorecards
-PO anomaly review
-Monitoring and audit log review

Questions to ask any vendor, including us

A short list that separates real Sage X3 AI work from a chatbot demo.

Which X3 version and API will you use, and is it supported on our hosting?
Do queries run with the asking user's X3 rights?
How do you handle multiple folders, sites and legal entities?
How are custom fields and screens included?
Can I see the query behind every figure?
Is any data sent to a public AI service?
How are changes proposed, approved and logged?

Frequently asked questions

Does Sage X3 have built-in AI?

Sage has introduced AI assistant features across its products. Check what your X3 version and subscription include. A separate private layer is useful when you want models on your own infrastructure, cross-site analytics or custom checks such as purchase order anomaly review.

Which X3 interfaces can AI use?

Typically GraphQL through Syracuse and published SOAP web services for reading, and database replicas in some self-hosted cases. Availability depends on your version and hosting, so we confirm during discovery before promising any approach.

Will AI approve or post purchase documents?

No. Approval and posting stay with authorized users in X3. AI can flag issues and draft suggestions, but the transaction is made by a person, preserving your control and audit evidence.

Can it handle several legal entities?

Yes, if the mapping carries folder, company and site into every query and respects each user's access. Cross-entity views need explicit authorization, and we test for data leakage between entities.

How does it help with lot tracking?

It answers questions about where a lot went and what remains, using stock journal and receipt data. It speeds up the lookup. The formal recall decision and records stay with your quality team.

What is the effort for a small team?

A read-only pilot for purchasing and stock typically takes 6 to 8 weeks. Heavier customization, several folders or restricted hosting access extends that.

Is a GPU needed?

For a handful of users, one GPU server with a mid-size open-weight model is usually enough. Sizing follows the number of concurrent users and the amount of document retrieval you want.

Related guides

Sage X3 + private AI

AI for Sage X3, Grounded in Your Own Ledgers and Stock Data

Add AI to Sage X3 finance and stock data without exporting it to a public model. On-prem and private-cloud LLMs for Sage X3 in the UK and France.

Sage X3 + food traceability AI

AI for Sage X3 in Food and Beverage: Lot Genealogy and Shelf-Life Answers From Your Own Data

Private AI on Sage X3 for food and beverage: lot genealogy, shelf-life, allergen and batch yield questions answered from live data, on-prem or private cloud.

Sage X3 + chemical process AI

AI for Sage X3 in Chemicals: Formula, Batch and Regulatory Questions Answered Privately

Private AI on Sage X3 for chemical makers: formulation, batch genealogy, SDS and REACH questions answered from your data, hosted on-prem or in your cloud.

Acumatica + private AI

AI for Acumatica Manufacturing Edition, Built on Your Own Data

Add AI to Acumatica Manufacturing Edition without giving up control of your shop and cost data. Private LLMs and agents grounded in your Acumatica instance.

SYSPRO + private AI

AI for SYSPRO, Grounded in Your Own Manufacturing Data

Add AI to SYSPRO without moving your manufacturing data to a public model. On-premise and private-cloud LLMs for SYSPRO in Australia, NA, and beyond.

PO follow-up that does not wait on a person

AI Purchase Order Automation: Supplier Follow-Up and Expedite Without a Buyer Chasing Email

AI drafts routine PO confirmation and expedite messages from open ERP order data, with a buyer approving every message and change before it goes out.

Plan it with numbers

Free Tool

Talk it through with an engineer who knows Sage X3

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