Electronics AI agent · Conflict minerals

Conflict Minerals Form SD AI Agent

Chase the declaration down the tiers it has to cross — the smelter behind each part, who answered and who went quiet — and leave the Form SD to the officer who signs it.

Typical delivery
4–6 weeks
Deployment
Your stack
Officer signs
Smelter level
After launch
Agent Care

What this agent does

Chases the response, never the conclusion

In

  1. 01

    A supplier answers for its own tier and goes quiet on the one below, and the chase carries on down.

  2. 02

    A template returns, and it is reconciled against the bill of materials it was meant to cover, part by part.

Reason

  1. 03

    A smelter is named, and it is checked against the conformant-smelter list, not against the supplier naming it.

  2. 04

    A filing date approaches, and it is re-derived from the calendar year, not from the fiscal year.

  3. 05

    A deadline falls on a Sunday, and Exchange Act Rule 0-3(a) moved it to Monday 1 June in 2026.

Decide

  1. 06

    A product is described as DRC conflict free, and that one sentence is what puts an audit back in scope.

  2. 07

    An Annex I volume threshold is crossed, and Regulation (EU) 2017/821 reaches worldwide, not the DRC.

Out

  1. 08

    A record is kept whole — the responses, their dates, the non-responders and the officer who signed.

  2. 09

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

Product statement

The agent chases the responses and keeps the inquiry record; a named executive officer signs the Form SD on behalf of the registrant, and the issuer answers for it.

Example workflow

One part, template to signature

  1. 1Declaration requestedA CMRT or EMRT issued to the direct supplier of record
  2. 2Smelter list assembledSmelter and refiner names, the tier reached, response dates and the parts each answer covers
  3. 3Inquiry record builtCountry-of-origin reasoning per part, against the limbs Item 1.01 sets out
  4. 4Controls appliedCoverage checks, response-date checks, conformant-list checks and completeness confidence

No human action required

Stages 1 to 4 run unaided, and nothing is determined at any of them — the agent is chasing, and the counsel lane opens at the completeness gate.

5DecisionBranches at the completeness gate
Coverage sufficient

Goes to the supply chain counsel to release.

Anything thin

Adds a securities counsel read first.

Supply chain counsel review

The record is held with its responses, its coverage and its gaps.

Release the record · Append evidence · Send to counsel

Released — signed by a named officer

  1. 6Supplier and part records updatedOnly where write access and disclosure policy allow it
  2. 7Outcome evaluatedCoverage margin, response completeness, counsel corrections and what a later review found

Corrections

Each counsel correction is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required

8 items

  • Determining where a mineral originated.
    • Concluding the country-of-origin inquiry.
    • Signing or filing the Form SD.
    • Calling any product DRC conflict free.
  • Automation boundary

    Agent acts unaided

    • ✓Chase the template down each tier and log who answered.
    • ✓Reconcile each response against the bill of materials it covers.
    • ✓Check each named smelter against the conformant-smelter list.
    • ✓Hold the reasoning per part with the person who made it.

    Nothing is filed except by a named executive officer, inside the agreed boundaries.

    • Judging a supplier response adequate.
    • Deciding an audit is or is not required.
    • Writing or approving the website disclosure.
  • Changes to templates, tiers or chase rules.

Example output

One part, annotated

Form SD is filed and not furnished, so Section 18 liability attaches and the record is read back years later.

Inquiry record · single part

Illustrative example
Part
Tantalum capacitor, power rail
Chased as
CMRT reissued, smelter names incomplete
Tier reached
Third tier of four
Response of record
Supplier return, 14 April 2026
Confidence
Held unfiled
Held for
The supply chain counsel, in person

As received

Taken from the supplier return and the bill of materials — the chase reaches as far as a tier will answer.

What the record holds

  • The returned template
  • Smelter and refiner list
  • Non-responder log

Why no inquiry call here

Item 1.01(a) asks for a reasonable inquiry, and the officer runs it.

What the score decides

Below the configured threshold the record picks up a counsel read before.

Action

Release the recordAppend evidenceSend to counsel

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

Every part

From the bill of materials

03Evidence

Below the distribution path

Our procurement agent names the distribution path behind a candidate part; this one goes on down to the smelter under it.

01Approved path

Smelters, not suppliers

The control point under 17 CFR 240.13p-1 and under Regulation (EU) 2017/821 is the smelter or refiner, never the supplier that invoiced you.

02Human review

Where the certification lands

Our mining agent closes on the certification sitting downstream. Downstream is the manufacturer who files, and that is this page.

04

Build an evidence trail

The part, the smelter list behind it and the officer who signed the filing stay on the record.

Integrations

Typical integrations

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

  • Declaration templates

    • CMRT
    • EMRT returns
    • Conformant-smelter list
  • Supplier records

    • Supplier master
    • contacts
    • Response and chase history
  • Product and part records

    • PLM
    • bill of materials
    • Part master
    • AVL

Agent

Conflict minerals inquiry

  • Reads the template
  • Chases the tier
  • Holds for the officer
  • Filing and disclosure

    • Form SD workpapers
    • Website disclosure posting
  • Observability & evaluation

    • OpenTelemetry
    • Langfuse
    • Supported monitoring/evaluation sources

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

Agent controls

Six passes between the model and the record

Six passes, each narrower than the one before. What is still there at the end is mapped below.

L6 · Outermost — last line of defence

Inward → L1 · closest to the model

L6Rollback / safe modePull the agent back to chasing responses only when evaluation or production signals degrade.Roll back
L5Version monitoringTrack model, prompt, chase-rule and list changes; 17 CFR 240.13p-1 was unamended on 23 August 2026.Track
L4TraceabilityRecord each response, the tier it came from, the reasoning applied and each read.Record
L3Officer releaseHold the record for a named human; it governs the filing, and not whether an origin is known.Gate
L2Policy guardrailsTest the record against Item 1.01(a) and (b), both still required, and the tiers reached.Restrict
L1Confidence thresholdsRoute a thin record to a counsel read before the supply chain counsel sees it.Require review
Model coreRecord built — the responses, the smelters, the tiers reached and completeness
L1 – L2
Test whether a record may stand
L3
Puts the filing in a person's hands
L4 – L5
Keep the part and the smelter behind it
L6
Holds the inquiry open when signals degrade

How Nestack evaluates it

Evaluate the whole assembly — not only the record that comes out.

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

Surface — the record a reviewer reads

  1. E1Final-output evaluationDid the record hold the response the inquiry actually rested on?
  2. E2Step-level evaluationDid the agent use the right template, the right tier and the current smelter list?
  3. E3Tool evaluationDid it read and write the correct part and the correct supplier record?
  4. E4Confidence calibrationDo low-confidence records actually attract more counsel corrections?
  5. E5Slice evaluationHow does performance change across specific product groups?
  6. E6Business outcomeHow many records needed a correction before counsel released them?

Floor — the outcome the issuer answers for

Failure modes

Where each failure originates in the agent

Seven failure modes, each one set at the stage where it actually begins.

Agent lifecycle

Direction of processing →

  1. 01 · Retrieval

    1 mode

    • IK-03

      Superseded list read

      The conformant list read is not the one current that day.

    Stage gathers

    The templates, the tiers, the smelters and the dates

  2. 02 · Reasoning

    2 modes

    • IK-04

      Relief read too widely

      Staff relief on one paragraph is read across the rule.

    • IK-06

      Tier below dropped

      The chase stops at the tier that answered.

    Stage proposes

    The responses, the tiers reached and completeness

  3. 03 · Tool / write

    2 modes

    • IK-02

      Coverage assumed, not shown

      Parts outside the return are counted as covered.

    • IK-05

      Reissue overwrites

      A resent template keeps only the later answer.

    Stage writes

    Only where write access and approval policy allow it

  4. 04 · Output

    1 mode

    • IK-01

      Filed, inquiry unrecorded

      The record shows a filing but not what produced it.

    Stage returns

    The record an officer signs and a reviewer reads

  5. 05 · Change / Version

    1 mode

    • IK-07

      Calendar carried forward

      A filing date is reused rather than re-derived each year.

    Stage tracks

    Model, prompt, chase rules and the filing calendar

  • Sev-1 · a filing leaves without an officer
  • Sev-2 · a wrong smelter reaches the record
  • Sev-3 · source degrades, record holds unfiled
See the complete Electronics failure-mode catalogue

Affected slices

One product group absorbs the corrections

A product-level smelter-coverage figure can read clean while one product group absorbs most of the corrections. Nestack reports the counsel-correction rate by product group, not only in total.

Slice performance — reported separately, not only in aggregate

Illustrative example
SliceFailure rateLiftLift vs. thresholdStatus
Tungsten vibration-motor parts6.2%3.6×Review
Tantalum capacitor assemblies4.4%2.6×Review
Gold-plated connector parts2.9%1.7×Watch
Tin solder on board assemblies1.6%0.9×Normal

Bar: counsel-correction-rate lift vs. tin-solder board-assembly baseline · scale 0–4.0× · tick marks the 2.0× review threshold

2 of 4 slices over threshold

Evidence-linked improvement

What an unreconstructable record costs

A cycle ends when the unanswered tier-three supplier is a standing case. That suite is what the next inquiry run is measured against.

Improvement cycle · five stages

Switchback — the path turns at Improve and returns at Learn

  1. 01

    Detect

    Counsel-correction rate rises in one product group.

  2. 02

    Diagnose

    The declaration that answered for one tier and went silent on the one below is read back until one cause remains.

  3. 03

    Improve

    Every change leaves numbered, with the parts that caused it filed beneath.

  4. 04

    Verify

    A single red part case is enough to keep the release back.

  5. 05

    Learn

    The case is permanent, and the inquiry rules are rewritten with it.

  6. Learn → Detect

    The return edge. The next inquiry is measured 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, record assembly, evaluation, integration, then production validation and handover.

Workstream

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

  1. 01Inquiry-workflow discovery and boundary definition.Week 1
  2. 02Supplier, PLM and template sources.Week 1, Week 2
  3. 03Smelter-list, tier-depth and chase-rule mapping.Week 1, Week 2
  4. 04Template ingestion and normalisation.Week 2
  5. 05Bill-of-materials and smelter binding.Week 2, Week 3
  6. 06Completeness scoring and counsel routing.Week 3
  7. 07Officer release workflow.Week 3, Week 4
  8. 08Supplier and PLM integration.Week 5
  9. 09Smelter and response cases.Week 4, Week 5
  10. 10Guardrails and inquiry controls.Week 4
  11. 11Inquiry-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 product group, one templateProductionProduction inquiry workflowAdvancedMultiple groups / regimes
Introduced at Pilot
Record assembly to your rules✓✓✓
Named-officer release✓✓✓
Smelter-list baseline✓✓✓
Introduced at Production
Reporting by product—✓✓
Release workflow in your systems—✓✓
Approved write-back—✓✓
Bill-of-materials integration—✓✓
Introduced at Advanced
Multi-regime inquiry duties——✓
Cross-tier chase packs——✓
High part volume——✓
Multi-tier chase 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 bill of materials and supplier master

    Part ingestion and supplier mapping

    Week 1
  2. 02Representative returned templates

    Record baseline, tier binding and response capture

    Week 2
  3. 03Your chase rules and filing calendar

    Smelter-list, tier-depth and chase-rule mapping and the automation boundary

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

    Supplier, PLM and template assessment, then integration setup

    Week 2
  5. 05Inquiries you would not want re-run

    Response cases and the evaluation suite

    Week 4
  6. 06What no country-of-origin inquiry may conclude

    Completeness scoring, counsel routing, guardrails and release controls

    Week 3
  7. 07A named executive officer to sign

    Officer release 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

The widths are the weeks the work actually takes, and the fifth week is shared for that reason.

Phase

W1

W2

W3

W4

W5

W6

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

Week focus

W1Inquiry workflow discovery, chase mapping and the automation boundary

W2Source integration and the smelter-list baseline

W3Record assembly, chase logic and release controls

W4Evaluation suite, response cases and failure-mode testing

W5Supplier integration, pilot parts and targeted corrections

W6One filing year run under the supply chain counsel, then Agent Care handover

Reading the band
A bar covers only the weeks its own work is named in. The doubled fifth week is real work, not padding.
At the end of W6
When the inquiry record validates, Agent Care takes the agent on.
Duration
Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Electronics AI agent

Build a conflict minerals agent around the officer who has to sign.

Show us how a declaration reaches you today and who signs before 31 May 2027. Form SD is filed and not furnished, so it has to survive being read as a statement the issuer made, by a reviewer looking years later for what the inquiry rested on.

Typical delivery
4–6 weeks
Deployment
Your bill of materials
Officer signs
Smelter level
After launch
Agent Care available

Nestack Agents · Conflict minerals inquiry

AGT-EL-19 · Agent Care available after launch

Build a conflict minerals agent

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

About you
Your products

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