AI for Strategic Sourcing and Supply Risk Capstone Exercise — AI for Strategic Sourcing and Supply Risk

What you'll learn

Apply the full course to a single sourcing decision carrying commercial, resilience, and due diligence exposure at once
Produce the four documents that would let the decision survive an audit, a regulator, and a supplier challenge
Distinguish, throughout, what your data establishes from what it merely suggests

Across this course you have separated recoverable from structural decisions, interrogated composite risk scores and their tier-one horizon, validated spend taxonomies against the decisions they drive, built should-cost models that survive contact with a supplier, distinguished leading operational distress signals from lagging credit data, separated observed from inferred supply chain relationships, worked through the CSDDD, Modern Slavery Act and UFLPA obligations, tuned sanctions screening under strict liability, tagged emissions data by provenance, and built decision records that answer a rejected supplier.

This capstone puts them together on one award, deliberately constructed so that the commercially attractive answer and the defensible answer are not obviously the same.

Capstone Exercise

Awarding a Consolidated Category Under Due Diligence and Resilience Constraints

Context

You are the category lead for electrical assemblies at a manufacturer with EU operations and significant US sales. Annual category spend is 22 million, currently split across nine suppliers. Your mandate is to consolidate to three and deliver a 6 percent saving. Four facts complicate it. First, your AI spend analysis shows the nine suppliers include two that appear as separate entities but share a registered address and a common director. Second, the highest-scoring candidate on your risk platform, at 88, offers the best price and is a trading company rather than a manufacturer, and your tier-two mapping for it is entirely inferred from customs data. Third, a second candidate manufactures in a region where your category is known to carry forced labour risk upstream in mineral processing, has provided a signed attestation and a passed social audit, and is the only one of the nine with supplier-specific product emissions data. Fourth, a third candidate is a small private supplier, 14 years trading, whose risk score is 61 largely because it files abbreviated accounts and has no press coverage — and which is the only currently qualified source for one safety-relevant connector.

Your Task

Produce four deliverables. First, a supply base analysis: resolve the entity question between the two suppliers sharing an address and state what it means for your apparent nine-way diversification, then identify for each of the nine what the consolidation model cannot see about why the relationship exists. Second, a due diligence plan for the two candidates carrying human rights exposure: state what the applicable regimes require given your EU operations and US sales, what the attestation and social audit do and do not establish, what tracing records you would need to secure contractually before the first order, and what you would do about the candidate whose tier-two map is inferred rather than observed. Third, a risk and resilience assessment: explain how you would treat the 88-scoring trading company and the 61-scoring small manufacturer differently from what their scores suggest, and state what monitoring you would put on the small supplier given that its score reflects data availability rather than evidence of risk. Fourth, a decision record for your recommended award: written so that a rejected supplier can be told why, an internal auditor can see what was assessed, and a regulator can see the due diligence steps were taken before the award.

Your notes (optional)

Deliverable

Four written deliverables: (1) a supply base analysis resolving the shared-entity question, restating actual diversification, and identifying what the consolidation model cannot see for each relationship; (2) a due diligence plan naming the applicable regimes and their differing asks, stating what the attestation and audit do and do not establish, specifying the tracing records to be secured contractually before first order, and addressing the inferred tier-two map; (3) a risk and resilience assessment treating the 88 and the 61 according to what the scores actually measure, with a monitoring approach for the single-source supplier; (4) a decision record for the recommended award that answers a rejected supplier, an internal auditor, and a regulator, capturing criteria, analytics and coverage, screening and dispositions, residual gaps, human judgment applied, and the data as it stood at the decision date.

Quick check

The capstone candidate with upstream forced labour exposure is also the only one holding supplier-specific product emissions data. How does the brief say those two facts bear on each other?

Select one answer.

Key takeaways

Entity resolution comes before consolidation. Two suppliers sharing an address and a director mean the diversification you think you have is not the diversification you have, and consolidating from a mis-stated baseline compounds the error.
A high composite score on a trading company is close to meaningless for supply chain due diligence, because the entity that matters is the manufacturer behind it — and an inferred tier-two map built from customs data establishes shipments between named parties, not who produced the goods.
The three due diligence regimes ask different things, and a signed attestation with a passed tier-one social audit satisfies none of what UFLPA requires. Tracing records must be secured contractually before the first order, because they cannot be assembled once goods are detained.
A low score driven by abbreviated accounts and absent press coverage measures how much is known about a supplier, not how risky it is. The correct response is proportionate targeted diligence plus operational leading indicators, not exclusion.
Good supplier-specific emissions data and good labour practices are unrelated datasets. Allowing strength in one to colour an assessment of the other is exactly the blending this course warns against, and it is easy to do when both sit under a single ESG heading.