Agentic AI Capability

What the ontology holds, and how you add your own.

The supply chain ontology is Juno’s built-in map of supply chain things and how they relate.

ONE NEIGHBOURHOOD OF THE SUPPLY CHAIN ONTOLOGYbelongs toscheduled asrefers toplanned atsupplied byanswered byheld atPurchase orderbuyer, terms,supplierPurchase orderlinematerial, quantity,priceSchedule linerequested date,quantityMaterial7741-X controllerboardMaterial atplantplant 1210 · safetystockSupplierDelta Fasteners LtdSupplierconfirmationconfirmed ship dateseven of the objects that ship with the product · showcase estate values
Seven objects around the purchase order line. Each line is a named relationship the ontology holds. Showcase estate values.
4
parts to every object: attributes, relationships, measures, synonyms
product
9
groups of objects ship with the product, from partners to finance links
product
2
routes to your own model: extend the ontology, or load yours as the target
product
100%
of ontology changes keep a version, an author and a time
product
Where this sits

Stage 2 of six: map

Step by step

How your data joins the map

The ontology exists before your first source; your team reviews proposals against it.

FROM A REGISTERED TABLE TO A PUBLISHED ONTOLOGY VERSION01Register thetablesthe description and asample reach Juno02Match to theontologyeach column meets anobject and an attribute03Review and confirmyour owners agree, adjustand add04Extend where youneed ityour own objects,attributes andrelationships05Publish a versionquestions, agents andreports read itnew sources and new questions send you round again
The ontology exists before your first source arrives, so step two has a target from the first minute.

The ontology is the target of every mapping. Your team reviews proposals against a model that is already there, in place of defining one from a blank page. The objects are the ones a supply chain conversation uses: material, plant, supplier, purchase order and its lines.

Each object carries four parts. Attributes are its properties. Relationships say how it connects to other objects. Measures are the numbers calculated from it. Synonyms are the words your people use for it, which is how a question in plain language finds it.

Questions and agents read the same ontology, so a term means one thing everywhere. Days of supply is calculated one way for ChatLima, for the safety stock agent and for the weekly digest. A change to that calculation reaches all three at once.

Your own objects join the same map. Juno keeps your additions beside the ones that ship with the product and marks which is which, so a later release of the built-in ontology arrives alongside your work.

One object in full

Everything the ontology holds about a purchase order line

Attributes, relationships, measures and synonyms: one object, fully described.

AttributesLine number, material, plant, ordered quantity, unit of measure, net price, currency, requested delivery date, cancellation window, deletion flag, delivery completed flag.
RelationshipsBelongs to a purchase order · refers to a material at a plant · supplied by a supplier · scheduled as one or more schedule lines · answered by supplier confirmations.
MeasuresOpen value · projected excess in weeks of demand · days left before the cancellation date · the supplier on-time rate behind the line.
Synonyms“PO line”, “order line”, “item”, “PO item”; your own words join this list through training.
IdentityThe purchase order number with the line number; the same line in two systems is recognised through the linking rules.
Source fieldsSAP EKPO and EKET · Oracle PO_LINES_ALL and PO_LINE_LOCATIONS_ALL · the supplier portal’s confirmation feed.

A question in plain language lands on this description. L. Tan’s Wednesday question about the open lines she can still cancel needs the object, the open state, the cancellation window, the schedule line date and the open value. Juno reads them here, then finds the mapped columns that carry them.

What ships built in

Nine groups of objects arrive with the product

Objects, attributes and relationships for the whole chain, from partners to finance links.

Partners

Supplier, supplier site, customer, carrier, contact.

Products and structures

Material, material at plant, bill of materials, where-used, product group.

Locations

Plant, storage location, distribution centre, port, customer site.

Orders and schedules

Purchase order, line, schedule line, sales order, scheduling agreement, delivery schedule.

Stock and movements

Stock on hand, stock in transit, batch, goods movement, reservation.

Plans and forecasts

Planned order, production order, forecast, exception message, safety stock.

Logistics

Delivery, shipment, freight order, container, tracking event.

Quality

Inspection lot, quality notification, defect, usage decision.

Finance links

Invoice, payment term, price condition, cost centre, currency.

Measures ship with the objects they belong to. On-time in-full, days of supply, forecast error, open order value and supplier late rate each carry one calculation, written once and read everywhere. Each measure names the attributes it uses, so a reader checks the arithmetic on the screen.

Adding your own

A custom object: the cancellation window

One object, defined by the people who own the data, published as a version.

Part of the objectWhat the team definedWhere the values come from
Object nameCancellation windowadded by P. Sharma on Tuesday 11:20
AttributesSupplier · material group · window in days · valid from · source documentthe supplier agreement register and SAP EKPO.ZZ_CXL_WDW
RelationshipsApplies to a purchase order line · agreed with a supplierchosen from the objects already in the ontology
MeasuresDays left in the window, counted from todaythe window with the requested delivery date
Synonyms“cancel window”, “NCNR window”, “free cancellation period”typed by the team and taught again as vocabulary
StatusPublished as ontology version 4reviewed by D. Okafor and J. Ridley

A custom object takes minutes to define and stays in your estate. It behaves like any other object: questions find it, mappings point at it, agents read it, and the review screen confirms the columns that fill it.

Custom attributes and custom relationships follow the same route where a whole object would be too much. The cancellation window began as one attribute on the purchase order line and became an object once the team needed windows by supplier and material group.

Every change writes a version

The ontology keeps an append-only history. A new object, a changed measure or a renamed synonym each write a version with an author and a time. Questions and agents read the published version, and restoring an older one writes a new version, so the history stays complete.

Your own model

Two routes, and what each one suits

Two routes, one published version at the end of each.

ROUTE 1 · EXTEND THE LIMA ONTOLOGYStart from the built-in objectsmaterials, plants, suppliers, orders, stockAdd what your business needsobjects, attributes, relationships, measuresKeep the strong proposalsJuno already knows the built-in attributesROUTE 2 · LOAD YOUR OWN MODELUpload your model documentobjects, attributes, relationships, descriptionsJuno maps your columns to itthe same five steps, your model as the targetYour names stay your namesanswers use the words your company publishesboth routes endin one publishedversion
The target of the mapping is yours to choose. The review screen, the versions and the audit record work the same way on either route.

Most teams take the first route. The built-in ontology carries the supply chain objects and the dictionaries behind the mapping proposals, so the first proposals arrive strong. Extending it costs an afternoon for each subject area.

The second route suits a company that already runs a published data model and wants answers in its own terms. The model document describes each object, its attributes and its relationships; Juno takes it as the target and your team confirms the result on the same screen.

The running example

Tuesday 11:20: the cancellation window becomes an object

Showcase estate.

The running example

The team had mapped ZZ_CXL_WDW to the cancellation window attribute of the purchase order line. On Tuesday D. Okafor pointed out that three suppliers agree their window in the contract, by material group, while the ERP field carries the days a buyer typed on the line.

P. Sharma defined a custom object at 11:20: cancellation window, with supplier, material group, days and a valid-from date. Two relationships tied it to the purchase order line and to the supplier. The ontology went to version 4, and the purchase order cancellation policy later read the contract window where one was agreed.

Next capability

The knowledge index

How Juno indexes every table, every column, the values inside them and the words your people use.

How this is usually done

A model built from a blank page

The supply chain arrives described; your judgement goes to the parts that are yours.

The usual route to business meaning is a modelling project. A team defines entities, metrics and joins for its own estate, agrees them across departments and maintains them as sources change. Practitioners describe that work in quarters for a large estate.

Juno arrives with the supply chain already described. Your team reads the objects, keeps what fits, and spends its judgement on the parts that are yours: your custom fields, your agreements, your words. The published result is a model your own people signed.

Terms on this page
ObjectA thing the supply chain talks about, such as a supplier, a material or a purchase order line.
AttributeOne described property of an object, such as the ordered quantity of a purchase order line.
RelationshipA named link between two objects, such as a purchase order line supplied by a supplier.
MeasureA number calculated from objects and attributes, written once and read everywhere.
SynonymA word your people use for an object or an attribute, so a question in plain language finds it.
Client modelYour own published data model, loaded as the target of the mapping in place of the Lima ontology.
Questions
Do we have to use the Lima ontology?

Two routes work. Extend the built-in ontology with your own objects and attributes, or load your own model as the target of the mapping. The review screen confirms either one.

What happens to our additions when Lima updates the ontology?

Your objects and attributes sit in their own space with your name on them. A later release of the built-in ontology arrives beside them, and the mapping keeps pointing where it pointed.

Which standards does the ontology line up with?

It uses its own business vocabulary and keeps documented alignments: SCOR for processes and metrics, GS1 keys for identity, ISA-95 for the plant, and each ERP’s own objects for write-back.

Where do measures live?

With the objects they are calculated from. One calculation serves ChatLima, the agents and the weekly digest, and the screen names the attributes it used.

In the platform

Where Juno sits among the four capabilities

Each one builds on the one before it.

One shared foundationCatalogLineageAccess policyWrite path

See Juno answer and act on your own data

Bring two systems and one decision your planners make every week. In the first week we register the data, confirm the mapping with your team, answer your questions in ChatLima and run the first agent at the level you choose.