Agentic AI Use cases

Every readiness item checked daily from twelve weeks out, and every gap owned.

Before the first production order of a new product, a launch checklist of 100 to 300 items has to be true: material and long-lead orders, supplier and OSAT qualification, test program release, part numbers and master data.

WHAT JUNO’S ROLES WOULD DO BEFORE THE FIRST PRODUCTION ORDER01Twelve weeks outthe daily scan starts; agate, a change or a newwork order also triggers…02Check every itemmaterial, long-leadorders, OSAT and supplierqualification, test…03Fix within rulespart numbers and SKUscreated by the namingrule; the ERP confirms04Route the resteach gap to its owner witha service-level clock05The gate decidesthe scorecard at the rampgate; go or hold stayswith peoplegaps that recur across launches become new checklist items and rules
From twelve weeks out to the gate: every item checked against the record that proves it, the routine ones closed within rules, the rest owned. Designed, yet to be built.
10%
of companies launch every product on time; 90% of engineering leaders report delays (CoLab, 26 July 2024, 250 respondents)
dated source
75%
of chip projects ran behind schedule in the 2024 Wilson Research Group study (Semiconductor Engineering, 27 March 2025)
dated source
75%+
of engineering and R&D executives expect AI to deliver 20% to 50% gains in productivity, time to market and cost (Capgemini Research Institute, 13 November 2025)
dated source
212 of 240
readiness items complete twelve weeks before one showcase launch, with three exceptions at the gate
showcase estate
A moment in the day

Where this use case starts

JR
Monday 08:15 · controller board 7741-Y · twelve weeks before launch

Twelve weeks before the launch of the new controller board 7741-Y, the daily readiness scan ran at 06:00. The scorecard shows 212 of 240 items complete. Overnight Juno created the 14 sales part numbers within the naming rule and the ERP returned each item number; it found that the work definition at plant 1210 lacks a resource for the new test step, and that the long-lead flash memory order covers seven of the first twelve weeks of the ramp plan.

P. Sharma approves the master-data fix from her queue in one click, and the ERP confirms the record. D. Okafor approves the top-up order with Acme Components GmbH, drafted with the coverage gap as its basis. The test program release candidate waits for the quality engineer’s signature, with a clock of five days.

At the ramp gate review J. Ridley sees the three open exceptions with their owners and clocks, and the 25 routine items inside their clocks, on one page read from eight systems. The decision to go or hold at the gate is his and the programme manager’s. showcase estate

Why this is a top use case

What the market is doing, with the evidence

The ERP vendors ship readiness agents, and the descriptions are theirs. Oracle’s Production Readiness Workspace, an agentic application, checks material availability, master-data gaps, work definitions and resources, generates prioritised corrective actions, and its Product Updates Advisor watches engineering changes that hit manufacturing; a manufacturing engineer approves each correction (Oracle, 29 June 2026; Oracle documentation, 28 August 2026) 7, 8.

SAP’s Joule Production Planning and Operations Agent automates prerequisite checks on material, capacity and scheduling before releasing production orders, recommends workarounds or releases orders at once, planned for the first quarter of 2026 (SAP, 7 October 2025) 3. Oracle’s Supplier Qualification Workspace turns qualification into a guided, risk-based process, and SAP’s Supplier Onboarding Agent validates data and handles escalations (Oracle, 29 June 2026; SAP, 7 October 2025) 7, 3. Planning suites offer like-item forecasting and launch execution for products that lack sales history (o9 Solutions, checked 23 September 2026) 54. Customer results for these agents have yet to be published.

The delay pain is measured. 90% of engineering leaders delay product launches and 10% launch every product on time (CoLab, 26 July 2024, 250 respondents) 30; 75% of chip projects run behind schedule (Semiconductor Engineering, 27 March 2025) 31. Over 75% of engineering and R&D executives expect AI to deliver 20% to 50% gains in productivity, time to market and cost (Capgemini Research Institute, 13 November 2025) 47. 33% of manufacturers cite faster time to market as the challenge driving AI investment (Aras, 20 March 2025, 656 executives) 17. Gartner predicts 40% of enterprise applications will carry task-specific agents by the end of 2026 (Gartner, 26 August 2025) 51.

The readiness list itself spans eight systems: items, bills of materials and work definitions in two ERPs, the product lifecycle management (PLM) project, the planning suite, the OSAT portal, the quality system, test data management and the supplier portal. Each vendor’s agent checks its own system. The platform’s part would be one scorecard across all of them, with the routine items closed within rules and every gap owned.

By hand today

How it is done today

A programme manager for new product introduction (NPI) keeps a launch checklist of 100 to 300 items in a spreadsheet or a PLM project and chases owners at a weekly readiness meeting. A supply planner checks material availability and long-lead orders line by line. A buyer confirms supplier qualification and OSAT capacity by e-mail. A quality engineer confirms the qualification report, the test program release and first-article approvals.

A master-data steward creates part numbers, bills of materials, work definitions, prices and classifications in each ERP by hand. Documentation releases the datasheet. Gaps surface at the first production order and cost days each. Seven roles, eight systems and readiness cycles of several weeks; in the market’s figures, 10% of companies launch every product on time (CoLab, 26 July 2024) 30.

Today

A spreadsheet, a weekly meeting and seven inboxes

  • A checklist of 100 to 300 items kept by hand and chased at a weekly meeting
  • Material and long-lead orders checked line by line in the planning suite
  • Supplier qualification and OSAT capacity confirmed by e-mail
  • Part numbers, bills of materials and work definitions keyed into each ERP
  • Gaps found at the first production order, costing days each
weeksreadiness cycles by hand; 90% of leaders report delayed launches (CoLab, 26 July 2024)
With Juno

One scorecard, checked daily against eight systems

  • Every item read from its source each day from twelve weeks out
  • Routine part numbers and SKUs created within the naming rule, confirmed by the ERP
  • Long-lead orders drafted where coverage falls under the ramp plan
  • Every other gap routed to its owner with a service-level clock
  • The scorecard at the ramp gate; go or hold stays with people
dailyone page for the gate review (showcase estate)
Stage by stage

How Juno enables it

Item, bill of materials, purchase order and supplier are objects the ontology holds already; checklist item, milestone, work definition, qualification report and test program release would extend it. The six stages below say what would be registered from eight systems, which items the rules may close on their own, and which people sign before the gate.

01
Register
Item, bill of materials (BOM), work definition, purchase and production order tables in SAP S/4HANA and Oracle Fusion; the PLM project and its checklist; the quality system’s records; the OSAT portal; the test data management release table; the supplier portal; the programme tool. All read in place. Objects: checklist item, milestone and gate, item master, work definition, material availability, long-lead order, supplier qualification, OSAT commitment, qualification report, test program release, part number and SKU, datasheet.
02
Map
Checklist item, milestone, work definition, qualification report and test program release would extend the ontology; item, bill of materials, purchase order and supplier exist already. Each checklist item is mapped to the source record that proves it, so a tick on the scorecard means a record was read, and the programme manager confirms the mapping once per checklist template.
03
Ask
“Which readiness items for 7741-Y are open, who owns them, and which block the first work order?” and “Does the long-lead flash memory order cover the ramp plan at plant 1210 week by week?” The answer names the eight systems it read and shows the query.
04
Teach
The gate names (qualification complete, risk start, production release), the readiness codes on the checklist and the company’s item-creation rules: the naming pattern for sales part numbers and SKUs, a SKU being the stock-keeping unit the ERP sells and ships. “Risk start” and “ramp plan” would be taught as the programme office uses them.
05
Act
Trigger: a gate milestone, the daily scan from twelve weeks out, a change on a product in ramp, or a work order for a new item. The policy: “Check every checklist item daily against its source; create sales part numbers within the naming rule; draft long-lead orders where coverage falls under the ramp plan; route every other gap to its owner with a clock.” It would be released at A3, routine creations automatic, drafts approved by the steward and the buyer. Write-back to the ERP, which returns the item and order numbers.
06
Learn
Gaps that recur across launches become new checklist items and rules, proposed by Juno from the record and approved as a new version. The measure that improves is checklist items closed before the gate, with master-data defects found before the first work order beside it; expedite spend during ramp is the cost of what the checklist missed.
Readiness scorecard · 7741-Y · Mon 08:15 · twelve weeks to launch · 212 of 240 items complete · showcase estate
ItemSource checkedStatusOwnerClock
Sales part numbers (14)SAP S/4HANA item masterCreated within rules · 06:05P. Sharma reviewsclosed
Work definition · plant 1210 · new test stepOracle Fusion work definitionsResource missing · fix draftedManufacturing engineer2 days
Flash memory · long-lead orderPurchase orders · supplier portalCovers 7 of 12 weeks · top-up drafted, $118,000D. Okafor3 days
OSAT capacity commitmentOSAT portalConfirmed for weeks 1 to 12Buyerclosed
Test program releaseTest data managementRelease candidate 3 · signature pendingQuality engineer5 days
Qualification reportQuality management systemSigned 12 SeptemberQuality engineerclosed
234 more itemseight systems209 complete · 25 open inside their clocksvarious

The scorecard J. Ridley would see at the ramp gate: every item with the source that proves it, the three exceptions with owners and clocks. Showcase estate; the roles are designed, yet to be built.

Who does what

The person and the agent

The agent A3 Delegated
  • Checks every readiness item daily against the record that proves it, across eight systems
  • Creates routine sales part numbers and SKUs within the naming rule; the ERP returns each number
  • Drafts long-lead orders where coverage falls under the ramp plan, and work-definition corrections for approval
  • Routes every other gap to its owner with a service-level clock and publishes the scorecard to the gate
the scorecard and the exception queue: gaps with owners and clocks; then the signatures at the gate
The person
  • The manufacturing engineer approves corrections to work definitions and resources
  • P. Sharma, steward, approves master-data changes outside the rule’s scope from her daily queue
  • D. Okafor, buyer, approves long-lead commitments above the value threshold; the quality engineer signs qualification and test program readiness
  • J. Ridley with the programme manager decides go or hold at each ramp gate; an auditor samples the automatic creations

The market runs routine corrections within rules with exceptions surfaced to people: one ERP vendor’s agentic applications progress routine work within guardrails and the other’s agent releases orders when checks pass (2025 to 2026) 7, 3. The blueprint designs the write-back roles here at A3 with the steward and the buyer taking exceptions, and spreads the fourteen roles across the gamut: part number and SKU creation at A4, the checklist and the scorecard at A3, ramp planning, long-lead commitment and OSAT readiness at A2, the gate review at A1, and the production release decision at A0, because the release is a signature.

What you would configure

The settings on this agent

Agent settings · Launch readiness checks across supply, quality and data
ElementSettingSet byChanges
ObjectiveEvery readiness item for a product in ramp is proven from its source each day, routine gaps closed within rules and every other gap owned before the ramp gateThe NPI programme manager, in a sentenceNew version
Data scopeSAP S/4HANA and Oracle Fusion items, BOM, work definitions, purchase and production orders; the PLM checklist; quality records; the OSAT and supplier portals; test data managementA. Novak, data leadNew version
RulesThe readiness policy and its planning table: check daily from twelve weeks out; create part numbers within the naming rule; draft long-lead orders under the ramp plan; route the restThe programme manager with the Planner AgentNew version, simulated first
Trigger06:00 daily from twelve weeks before launch; each gate milestone; an engineering change on a product in ramp; a work order for a new itemThe programme managerAt once
Actions permittedUpdate a checklist item with evidence · create a part number or SKU · draft a work-definition correction · draft a long-lead order · route a gap with a clock · publish the scorecardThe programme manager, ownerNew version
LimitsLong-lead orders above $25,000 go to D. Okafor · part numbers only from the naming rule, 50 a day · a work-definition change always waits for the manufacturing engineer · confidence floor 0.90 on a coverage checkThe programme manager, J. Harrison and P. SharmaNew version
EscalationAn item that blocks the first work order, to the programme manager at once · a clock that runs out, to the owner’s manager · a failed write-back, to P. SharmaThe programme managerAt once
LevelA3 Delegated with P. Sharma and D. Okafor on exceptions; go or hold at the gate stays with J. Ridley and the programme manager at A0J. Ridley, supply chain directorUp on evidence, down in one step
FeedbackEvery approved correction, edited order and gap that recurred is recorded and usedAlways onFixed

Every row is a setting or a versioned rule; the level and the limits are the ones a manager signs.

Built from the blueprint

The roles this use case would draw on

Lima’s agent blueprint is a design of 960 agent roles across five end-to-end chains. It is the map of what Juno can be configured to do, and the first agents are built from it one at a time, the way the PO cancellation agent was.

Blueprint · designed, yet to be builtThese roles are designed on the map. Each is configured from the same parts.
Launch 9 roles
Ramp plan & capacity readiness A2
Long-lead material commitment A2
Pilot line & risk-start coordination A2
Supply readiness scorecard A3
OSAT & assembly readiness A2
Ramp exception escalation A3
Ramp gate readiness review A1
Launch readiness checklist A3
Part number & SKU creation A4
Conceive 1 role
Milestone readiness assessment A3
Qualify 4 roles
Test hardware & load-board ready A3
Test program development & release A2
Customer qual evidence delivery A3
Production release decision A0

Additional capability in Idea-to-Market lists the rest of the map for this chain.

How it is measured

The measures that show it working

Measure
Launches on time, baseline: 10% of companies launch every product on time (CoLab, 2024)
Projects behind schedule, baseline 75% (Wilson Research Group 2024 study)
Checklist items closed before the gate
Master-data defects found before the first work order
First-pass production order release rate
Time from qualification complete to production release
Expedite spend during ramp

Baseline checklist items closed before the gate first: the last launch’s checklist gives the number in an afternoon, and it is the one the gate review already argues about. Master-data defects found before the first work order follows, because it is what the A4 part-number role is judged on. Launches on time and expedite spend during ramp read over a year of launches.

Terms on this page
OSATOutsourced assembly and test: the partner that packages the die and runs the production test.
RampThe weeks between the first production order and full volume, run against a plan of weekly quantities.
GateA milestone review, such as qualification complete or production release, where named people decide go or hold.
Work definitionThe ERP record of the steps, resources and materials that make an item at a plant; a routing.
Test programThe software the tester runs to sort good die from bad; its release is a readiness item.
Questions
What would run on its own, and what would wait?

Part numbers and SKUs that follow the naming rule, and checklist updates with their evidence, would run within the rule at A3. Work-definition changes, long-lead orders above $25,000 and every other gap wait for a named person.

What is the go or hold decision built on?

The scorecard at the gate: every item with the record that proves it, the open exceptions with owners and clocks, and the coverage of the ramp plan week by week. The decision stays with J. Ridley and the programme manager.

How does the checklist get better?

Gaps that surface after the gate, at the first production order, are recorded against the launch. Juno proposes them as new checklist items and rules, and the programme manager approves them as a new version for the next product.

Next chain
The question this leaves

The product is released and its first orders arrive from the customers who evaluated samples. How does the same machinery carry an order from intake to cash?

Answered next

Order-to-Cash

Order intake, cash application, collections and disputes: the chain where the market’s evidence is deepest.

Sources

Dated sources used on this page

  1. [3] SAP News Center, Innovative Updates for Supply Chain Management (SAP Connect), 7 October 2025 (checked 23 September 2026) · https://news.sap.com/2025/10/sap-connect-innovative-updates-supply-chain-management/
  2. [7] Oracle, Oracle Adds New Fusion Agentic Applications to Help Customers Improve Supply Chain Performance, 29 June 2026 (checked 23 September 2026) · https://www.oracle.com/news/announcement/oracle-adds-new-fusion-agentic-applications-to-help-customers-improve-supply-chain-performance-2026-06-29/
  3. [8] Oracle Help Center, Oracle Fusion Cloud Manufacturing 26C What's New: AI Agentic App, Production Readiness Workspace, 28 August 2026 (checked 23 September 2026) · https://docs.oracle.com/en/cloud/saas/readiness/scm/26c/mfg26c/26C-mfg-wn-f44539.htm
  4. [17] Aras (Business Wire), 91% of Manufacturers Plan to Increase Investment in AI for Product Development, 20 March 2025 (checked 23 September 2026) · https://www.businesswire.com/news/home/20250320558337/en/91-of-Manufacturers-Plan-to-Increase-Investment-in-AI-for-Product-Development
  5. [30] CoLab Software (survey of 250 engineering leaders by Global Surveyz), 90% of companies delay product launches, 26 July 2024 (checked 23 September 2026) · https://www.colabsoftware.com/research/90-of-companies-delay-product-launches
  6. [31] Semiconductor Engineering (reporting the 2024 Siemens EDA and Wilson Research Group Functional Verification Study), First-Time Silicon Success Plummets, 27 March 2025 (checked 23 September 2026) · https://semiengineering.com/first-time-silicon-success-plummets/
  7. [47] Capgemini Research Institute, Engineering and R&D Pulse 2026, 13 November 2025 (checked 23 September 2026) · https://www.capgemini.com/insights/research-library/engineering-research-development-pulse-2026/
  8. [51] Gartner, Gartner Predicts 40% of Enterprise Apps Will Feature Task-Specific AI Agents by 2026, Up from Less Than 5% in 2025, 26 August 2025 (checked 23 September 2026) · https://www.gartner.com/en/newsroom/press-releases/2025-08-26-gartner-predicts-40-percent-of-enterprise-apps-will-feature-task-specific-ai-agents-by-2026-up-from-less-than-5-percent-in-2025
  9. [54] o9 Solutions, New Product Introduction (NPI) Planning Software Solution Powered by AI (solution page), undated, checked 23 September 2026 (checked 23 September 2026) · https://o9solutions.com/solutions/demand-planning/npi-planning

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.