Manufacturing AI agent · Obsolescence

Component Obsolescence, EOL & Counterfeit-Control AI Agent

Watch the parts on your bills of material rather than the notices you were sent, size the last-time-buy exposure behind each one, and hold the disposition for your engineering authority to sign.

Typical delivery
4–6 weeks
Deployment
Your stack
Authority signs
Watch-first
After launch
Agent Care

What this agent does

Sizes the exposure, not the disposition

In

  1. 01

    A notice arrives under J-STD-046A or J-STD-048A, both revised December 2025, and is matched to your BOM lines.

  2. 02

    A notice never arrives — the usual case — and the part is inferred obsolete from lifecycle and stock signals.

Reason

  1. 03

    A discontinuance names a last-order and a last-ship date, and the buy is sized to the shorter window.

  2. 04

    A buy quantity is set against production, service and spares demand over the remaining product life.

  3. 05

    A source is offered, and it is placed in the DFARS 252.246-7008 tiers rather than called authorised.

Decide

  1. 06

    A GIDEP entry lands, and it is screened by part number, date-code range, lot code and manufacturing site.

  2. 07

    A clause set is read per contract: the DFARS system test and the FAR 52.246-26 (AUG 2024) test differ.

Out

  1. 08

    A disposition pack is drafted with its evidence and routed to the named engineering authority.

  2. 09

    Execute write actions only inside the approval boundaries agreed during implementation.

Product statement

The agent sizes and drafts; an engineering authority approves the disposition, and a design authority approves any substitution.

Example workflow

One obsolete part, end to end

  1. 1Obsolescence signal receivedProduct-change notice, GIDEP entry, distributor stock signal or a silent lifecycle change
  2. 2Exposure sizedWhere the part sits on the BOMs, demand to end of life, stock on hand and the buy window
  3. 3Disposition draftedLast-time buy, alternate, redesign or quarantine, with the evidence and the sixty-day clocks
  4. 4Controls appliedSourcing-tier checks, GIDEP screening, last-order and last-ship date checks and confidence threshold

No human action required

Stages 1 to 4 run unaided, and nothing is dispositioned at any of them — the agent is sizing, and the authority's lane opens at the confidence gate.

5DecisionBranches at the confidence threshold
High confidence

Goes to the engineering authority to approve.

Low confidence

Adds a component-engineering read first.

Authority approval

The disposition is held with its evidence, its open questions and the confidence.

Approve · Amend · Send to component review

Approved — disposition recorded

  1. 6PLM and ERP updatedOnly where write access and approval policy allow it
  2. 7Outcome evaluatedAmendments, notices missed, buy shortfalls and parts that went obsolete unseen

Amendments

Each disposition the authority amends is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required

8 items

  • Approving a disposition on an obsolete part.
    • Approving an alternate, substitute or re-source.
    • Notifying the Contracting Officer or filing to GIDEP.
    • Committing the last-time-buy quantity and the cash.
  • Automation boundary

    Agent acts unaided

    • ✓Watch the parts on the BOMs and match notices to them.
    • ✓Size the buy exposure against demand and remaining life for the named owner.
    • ✓Place each offered source in the DFARS sourcing tiers.
    • ✓Mark what the authority should weigh, and hold the pack into the review queue.

    Any write happens inside the boundaries agreed at implementation, never ahead of approval.

    • Determining that a part is counterfeit.
    • Returning, scrapping or releasing a quarantined part.
    • Declaring which contracts carry which clause.
  • Changes to sourcing, buy-window or screening rules.

Example output

One obsolete part, annotated

Everything the agent sizes is attached to the record it was drawn from.

Obsolescence output · single part

Illustrative example
Part
Linear regulator
Signal found
No notice issued; distributor stock falling
BOM lines hit
14 assemblies
Source of signal
Authorised distributor
Confidence
84%
Disposition drafted
Last-time buy, sized to end of life

As received

Taken from the part master and the distributor feed — nothing on this side is inferred by the agent.

Evidence used

  • Part-master lifecycle
  • Distributor stock feed
  • Demand to end of life

Why this disposition

No notice reached us and stock is falling and a person still decides.

What the score decides

Below the configured threshold the pack picks up a component-engineering read first.

Action

ApproveAmendSend to component review

Value

Where AI adds value

The same four claims, placed at the point in the workflow where each one applies.

Where the value lands

Value 01 – 04

Parts on your BOMs

From the part master

03Watching

Watch from the BOM

Draw on the part master, the distributor feeds and AS5553E (13 November 2025) disposition rules.

01Approved path

No notice is the usual case

Parts drifting to end of life surface from the BOM, not from a notice that may never come.

02Human review

Send review to the exposed parts

Unmatched notices and low-confidence packs are marked, so the authority's read starts where exposure sits.

04

Build an evidence trail

The part, the notice it was matched to and the authority who dispositioned stay on the item.

Integrations

Typical integrations

Five system groups connect to the same agent. Which of them are in scope is decided in discovery.

  • PLM and part master

    • Teamcenter
    • Windchill
    • Arena
    • Duro
  • Component intelligence

    • SiliconExpert
    • Z2Data
    • IHS Markit BOM data
  • Distributors and sources

    • Arrow
    • Avnet
    • Mouser
    • Digi-Key
    • Rochester

Agent

Obsolescence & counterfeit control

  • Watches the BOM
  • Sizes the exposure
  • Holds for approval
  • ERP and quality

    • SAP
    • Oracle
    • Infor
    • GIDEP
    • nonconformance
  • Observability & evaluation

    • OpenTelemetry
    • Langfuse
    • Supported monitoring/evaluation sources

Integration availability depends on the client's existing systems and API access.

Agent controls

Six layers between the model and the disposition

Six layers, folded one into another. What reaches the model uncaught is named in the map below.

L6 · Outermost — last line of defence

Inward → L1 · closest to the model

L6Rollback / safe modePull sizing back to watch-list-only when evaluation or production signals degrade.Roll back
L5Version monitoringTrack model, prompt, sourcing-rule and screening-rule changes.Track
L4TraceabilityRecord the notice, the match, the evidence and the approval time.Record
L3Authority approvalHold packs for the named authority; it governs release, not whether the disposition is right.Gate
L2Policy guardrailsTest packs against configured sourcing rules; a failure returns the pack, and no screen proves a part authentic.Restrict
L1Confidence thresholdsRoute low-confidence packs to a component-engineering read before the authority sees them.Require review
Model corePack produced — exposure, buy window, sourcing tier, open questions and confidence
L1 – L2
Test whether a pack may stand
L3
Puts the disposition in an authority's hands
L4 – L5
Keep the part and the notice behind it
L6
Narrows to watch-list reporting when signals degrade

How Nestack evaluates it

Evaluate the whole watch workflow — not only the finished pack.

Coverage runs the whole depth of the workflow, and every layer is cut by slice.

Surface — the pack the authority signs

  1. E1Final-output evaluationDid the exposure trace back to a BOM line and a demand figure?
  2. E2Step-level evaluationDid the agent use the right part master, notice and sourcing rules?
  3. E3Tool evaluationDid it read and write the correct part and the correct BOM?
  4. E4Confidence calibrationDo low-confidence packs actually attract more amendments?
  5. E5Slice evaluationHow does performance change across specific commodities?
  6. E6Business outcomeHow many events met all three DFARS 231.205-71 conditions together?

Floor — the exposure the programme answers for

Failure modes

Where each failure originates in the agent

Seven failure modes, set at the stage each one starts in.

Agent lifecycle

Direction of processing →

  1. 01 · Retrieval

    1 mode

    • ZO-03

      Suffix match missed

      A packaging suffix hides the BOM hit for the notice.

    Stage gathers

    Part masters, notices, stock feeds and demand

  2. 02 · Reasoning

    2 modes

    • ZO-04

      Spares tail omitted

      The buy covers production and not field support.

    • ZO-06

      Sub-tier part unseen

      Obsolescence sits inside a purchased assembly.

    Stage proposes

    Exposure, buy window, sourcing tier and confidence

  3. 03 · Tool / write

    2 modes

    • ZO-02

      Last-ship date used

      The buy is placed after the order book closed.

    • ZO-05

      Line card out of date

      A terminated franchise still reads as tier one.

    Stage writes

    Only where write access and approval policy allow it

  4. 04 · Output

    1 mode

    • ZO-01

      Approval treated as final

      Stock is consumed and nothing re-opens the case.

    Stage returns

    The disposition the authority approves and signs

  5. 05 · Change / Version

    1 mode

    • ZO-07

      Silent rule regression

      A rule change narrows what the watch list flags.

    Stage tracks

    Model, prompt, sourcing rules and screening rules

  • Sev-1 · pack closed outside the boundary
  • Sev-2 · a part goes obsolete unwatched
  • Sev-3 · feed degrades, pack goes to review
See the complete Manufacturing failure-mode catalogue

Affected slices

One commodity can carry the exposure

A BOM-level exposure figure can read as manageable while a few commodities carry nearly all of the unwatched risk. Nestack reports the amendment rate by slice, not only in total.

Slice performance — reported separately, not only in aggregate

Illustrative example
SliceFailure rateLiftLift vs. thresholdStatus
Parts with no PCN issued8.8%3.5×Review
Sub-tier parts in assemblies6.1%2.4×Review
Suffix and package variants3.9%1.6×Watch
Directly-bought active parts1.2%0.5×Normal

Bar: amendment-rate lift vs. active-part baseline · scale 0–4.0× · tick marks the 2.0× review threshold

2 of 4 slices over threshold

Evidence-linked improvement

A cycle ends with one more case

A cycle shuts when the silent discontinuation is a case the next release has to catch. That suite is what the next watch list run is measured against.

Improvement cycle · five stages

Switchback — the path turns at Improve and returns at Learn

  1. 01

    Detect

    Amendment rate rises in a commodity.

  2. 02

    Diagnose

    The part that went end-of-life without a notice anyone received is read back to one cause.

  3. 03

    Improve

    The change is numbered on the way out, with the notices that revealed it attached.

  4. 04

    Verify

    A disposition case that has not cleared keeps the release parked until it does.

  5. 05

    Learn

    It joins the suite for good, and the watch list is edited to match.

  6. Learn → Detect

    The return edge. The next watch list runs against a suite one case longer.

Typical build scope

Twelve workstreams across six weeks

The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, watch and sizing, evaluation, integration, then production validation and handover.

Workstream

Week 1Week 2Week 3Week 4Week 5Week 6

  1. 01Programme-workflow and disposition boundary work.Week 1
  2. 02BOM, part-master and feed assessment.Week 1, Week 2
  3. 03Sourcing-tier and screening-criteria mapping work.Week 1, Week 2
  4. 04Part-master ingest and notice matching.Week 2
  5. 05Exposure sizing and evidence binding.Week 2, Week 3
  6. 06Confidence scoring and exception routing.Week 3
  7. 07Authority approval workflow.Week 3, Week 4
  8. 08PLM and ERP-system integration.Week 5
  9. 09Notice and sourcing cases.Week 4, Week 5
  10. 10Guardrails and sourcing controls.Week 4
  11. 11Item-trail instrumentation.Week 5, Week 6
  12. 12Deployment, documentation and Agent Care handover.Week 6

12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel

Final scope and sequence confirmed in discovery

Engagement tiers

What each tier includes

Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.

Capability✓ in scope · — not at this tierPilotOne programme, one commodityProductionProduction part mastersAdvancedMultiple programmes / sites
Introduced at Pilot
Watching to your BOMs and rules✓✓✓
Authority approval✓✓✓
Watch-coverage baseline✓✓✓
Introduced at Production
Reporting by commodity—✓✓
Approval workflow in your systems—✓✓
Approved write-back—✓✓
BOM and distributor integration—✓✓
Introduced at Advanced
Multi-standard sourcing rules——✓
Multi-stage engineering approvals——✓
High part-count BOMs——✓
Multi-programme sourcing controls——✓
Build priceFrom $5,000From $8,000Custom quote
Final build price
Confirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
Separate from build
Build pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.

What we need from you

What you bring, and what we build with it

Each input maps to a piece of build scope and a week in the delivery timeline.

  1. 01Your bills of material and part master

    Part-master ingestion and notice matching

    Week 1
  2. 02Representative notices and BOM revisions

    Watch baseline, notice matching and BOM binding

    Week 2
  3. 03Your sourcing tiers and approved suppliers

    Sourcing-tier and screening-rule mapping

    Week 1
  4. 04Access to relevant APIs, feeds or exports

    BOM, part-master and feed assessment, then integration setup

    Week 2
  5. 05Parts you would not want bought

    Counterfeit cases and the evaluation suite

    Week 4
  6. 06What no source may be trusted on

    Confidence scoring, exception routing, guardrails and sourcing controls

    Week 3
  7. 07A named engineering authority to approve

    Authority approval workflow, then pilot and production validation

    Weeks 5–6

Nothing else is required

Deployment, documentation and Agent Care handover are ours.

Delivery timeline

Four phases across six weeks

Bands sit where the notices put them, which is why week 5 holds evaluation and pilot at once.

Phase

W1

W2

W3

W4

W5

W6

  1. DiscoveryW1
  2. BuildW2 – W3
  3. EvaluateW4 – W5
  4. Pilot & LaunchW5 – W6

Week focus

W1Obsolescence workflow discovery, rule mapping and the automation boundary

W2Source integration and the watch-list baseline

W3Exposure sizing, confidence logic and approval controls

W4Evaluation suite, sourcing and screening checks and failure-mode testing

W5PLM and ERP integration, pilot parts and targeted corrections

W6One programme review run under the component engineer, then Agent Care handover

Reading the band
Each bar covers only the weeks its work is named in. The week 5 overlap is work, not padding.
At the end of W6
Once the review validates, Agent Care owns the running agent.
Duration
Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Manufacturing AI agent

Build an obsolescence agent around your engineering-authority disposition.

Show us your BOMs and who signs a disposition. What you get, plainly: a watch list against your own parts, the exposure sized with its evidence, and a pack your authority can approve.

Typical delivery
4–6 weeks
Deployment
Your PLM and ERP
Authority signs
Watch-first
After launch
Agent Care available

Nestack Agents · Obsolescence & counterfeit control

AGT-MFG-20 · Agent Care available after launch

Build an obsolescence agent

Four questions so the call starts from your actual setup — then pick a time that suits you.

About you
Your parts

No obligation. We’ll map the workflow, name the automation boundary and tell you if an agent is the wrong answer.