DLDesheng LuContact
All work
MVP / UAT-readyFuture-state MVPAnonymized

Geospatial Risk Intelligence / Flagship product

Tier-N Risk Navigator

See how a disruption travels from place to supplier to material before it becomes a sourcing crisis.

My mandate

Define and lead the material-centric geospatial MVP, turn Phase 1 feasibility into a UAT-ready release contract, and align resilience, procurement, data, and engineering around the evidence required for enterprise handoff.

Simulated Tier-N network viewSimulated product scenario
Product viewSimulated Tier-N network view
Simulated Tier-N supply-chain risk graph connecting material and supplier relationships with an evidence review panel

Public-safe reconstruction using fictional or anonymized information. It demonstrates product mechanics, not an original product screen.

Portfolio product state
MVP target · UAT-ready geospatial risk system
Evidence basis
Phase 1 feasibility POC completed; Phase 2 user and business validation planned.
Primary users
Procurement leaders / Supply-chain resilience teams / Material risk owners
Leadership scope
Strategy / roadmap / validation / cross-functional delivery

Evidence first

What is proven, what I changed, and what remains open.

A 90-second evidence map for hiring review. Synthetic product views explain mechanics; they never count as outcome proof.

Verified now
A Phase 1 feasibility POC and business-facing product record, plus current POC-user feedback that changed the validation target.
Next evidence gate
One site-specific product-to-material-to-Tier-N slice accepted through manual audit and guided UAT.
01

User / workflow evidence

Verified record

Current POC users asked to review one real finished-good, site, and country workflow first. They made authoritative direct-supplier data, supplier location, material context, and material-level volume or share explicit product needs.

02

Product decision

Verified record

I rebased the gate from an impressive network view to an output-first operating slice: product → site → material → direct supplier → evidence-labelled deeper tiers → accountable exposure review.

03

AI / system boundary

Supported synthesis

Authoritative internal records anchor direct-supplier relationships; AI enrichment fills genuine gaps and proposes deeper-tier paths with visible sources and confidence; the business owner validates exposure and action.

04

Evaluation → change

Supported synthesis

Business review shifted the sprint toward site specificity, authoritative source labelling, material-level context, sampled audit, and explicit evidence confidence. The connected retest and acceptance receipt are still pending.

05

Outcome / boundary

Reported · receipt pending

The current sprint reports expanded data, graph, UI, and quality-control work. Those implementation claims and numeric run results remain excluded here until code, test, environment, and user-acceptance evidence are connected.

AI product decision record

The choices behind the product, architecture, and release gate.

Use AI to find and explain plausible hidden exposure; use graph constraints and business review to decide what is real and actionable.

01

AI fit

Supported synthesis

Use AI for event interpretation, entity resolution, evidence enrichment, and plausible deeper-tier path discovery.

Upstream relationships are incomplete, multilingual, and distributed across records that do not join cleanly.

Alternative consideredPure rules miss unknown relationships; a fully generative risk verdict would turn uncertainty into false confidence.

02

Knowledge & context

Supported synthesis

Combine authoritative direct-supplier records, an explicit supply graph, and cited retrieval for deeper-tier evidence.

The product must distinguish known relationships, inferred paths, missing coverage, and material relevance.

Alternative consideredGeneric RAG over public articles cannot prove that an event connects to a specific material or operating site.

03

System architecture

Designed · not yet measured

Use one event-to-owner workflow with graph and research tools rather than a swarm of autonomous agents.

The hard product problem is preserving provenance and confidence across one exposure chain, not maximizing delegation.

Alternative consideredMCP or additional agents become useful only when multiple governed data and action services need reusable discovery and permissioning.

04

Evaluation & release

Supported synthesis

Gate on relationship audit, evidence quality, coverage, false escalation, material relevance, and owner action.

A visually impressive graph has no value if it cannot survive manual audit or improve a real resilience decision.

Alternative consideredNode count, graph density, or one successful demo is not an acceptance metric.

Evidence I still need from the private project record3 open items
  1. 01

    The exact Phase 1 demonstration or acceptance artifact approved for public use.

  2. 02

    A dated before/change/retest record for the site-specific POC journey.

  3. 03

    Business acceptance results for usefulness, relationship quality, coverage, false escalation, and owner action.

Verified = direct approved recordSupported = defensible synthesisReported = source receipt pendingDesigned = future test or control

01 / Executive impact

Why this mattered

A disruption is not actionable until a place resolves to a material, an evidence trail, and an owner.

Direct-supplier views can miss the hidden upstream dependency that actually stops supply. A compelling map still leaves the business guessing unless it connects the event to Tier-N relationships, material relevance, evidence quality, and an accountable response.

My impact

I turned a feasibility demonstration into an end-to-end risk-decision product and a testable MVP contract.

Verified now: Phase 1 established a feasibility POC, and current POC-user review changed the validation target to one output-first, site-specific product-to-material workflow. The end-to-end MVP is a public-safe future-state reconstruction; connected build, retest, manual-audit, user-value, network-coverage, alert-precision, and UAT evidence remain promotion gates.

  1. 01
    Unit of value

    Move from location to exposure

    I reframed the product around the chain from global event to supplier relationship to material-level business relevance.

    Map to decision workflow
  2. 02
    Trust design

    Explain before alerting

    I made sources, relationship confidence, coverage gaps, and human validation part of the product rather than caveats behind a score.

    Evidence before escalation
  3. 03
    Release model

    Turn feasibility into UAT criteria

    I defined the MVP boundary, business scenarios, acceptance dimensions, and promotion gate needed to move beyond a successful POC.

    Feasibility to testable release
  4. 04
    Operating ownership

    Route risk to accountable action

    I aligned the flow around validation, prioritization, mitigation, and a named owner rather than autonomous sourcing action.

    Signal to owned response

02 / Future-state MVP system

The complete geospatial-risk MVP, designed around the decision and the UAT evidence required to trust it.

A map becomes intelligence only when it resolves to a decision. Connect event, place, Tier-N relationship, material, evidence, confidence, and accountable action in one reviewable flow.

North-star experience

When an external event changes the risk landscape, the product geolocates the disruption, traces potential exposure across suppliers and sub-tiers, resolves it to relevant materials, explains evidence and uncertainty, and routes a prioritized signal to the people responsible for mitigation.

Zero-context product contract

The user, the system, and the decision are explicit.

A cold reader should be able to describe the product without relying on internal context, AI terminology, or the rest of this portfolio.

Primary user
A resilience or procurement lead responsible for materials that may be exposed to an external disruption.
Trigger
A geopolitical, climate, logistics, or operational event changes the risk landscape.
Inputs
Event location and timing, public evidence, supplier and Tier-N relationships, material relevance, coverage gaps, and ownership context.
MVP product action
Resolve the event to geography, trace plausible network exposure, distinguish evidence from inference, and prioritize material-level review.
MVP output
A reviewable exposure path with sources, confidence, gaps, affected materials, and an accountable action queue.
MVP decision enabled
Which exposure requires validation or mitigation now, who owns it, and where evidence is still insufficient.

How the product works

From signal to accountable action.

  1. 01

    Detect and locate

    A geopolitical, climate, logistics, or operational event is anchored to an affected geography.

  2. 02

    Trace Tier-N

    The product follows potential exposure across locations, suppliers, sub-tiers, materials, and dependencies.

  3. 03

    Explain impact

    Users see affected relationships, sources, confidence, material relevance, and coverage gaps.

  4. 04

    Coordinate action

    An accountable owner validates the signal, prioritizes exposure, and coordinates mitigation.

Interactive product walkthrough

Follow one disruption from signal to accountable mitigation.

This walkthrough demonstrates the UAT-ready geospatial-risk MVP. Phase 1 feasibility is verified evidence; the end-to-end operating states remain a public-safe future-state reconstruction until promoted by approved UAT evidence.

Tier-N Risk NavigatorExposure intelligence workspace
MVP / UAT-ready walkthrough
Simulated product scenario
Public-safe simulation

Every organization, supplier, geography, value, date, confidence measure, and output shown here is simulated. This demonstrates product mechanics and decision design, not an employer interface or an actual project result.

Scenario

A fictional external event may affect a critical material beyond the buyer's direct supplier visibility.

Decision owner
Category and resilience leads
AI role
Trace, qualify, and structure action
Boundary
No unsupported risk verdict
01 / Detect

Translate an external event into a decision signal.

The product resolves the event, its location, and the decision horizon before searching for exposure.

Human gateSignal qualified
External eventArden Coast corridor disruptionFictional geography · simulated event
Geo resolution18 km operating radiusSimulated product scenario

Product principleA signal becomes useful only when evidence, uncertainty, and ownership travel with it.

Use arrow keys to move between workflow stages

AI operating and evaluation model

The model reasons. The product controls. The human remains accountable.

This case presents the complete MVP/UAT-ready operating flow as a future-state public-safe reconstruction. Phase 1 supports the verified technical-feasibility claim; business usefulness, network-scale coverage, evidence quality, false escalation, and owner action remain promotion gates.

01Deterministic product

Graph constraints and ownership

Relationship types, tier boundaries, material mappings, coverage status, alert rules, and action ownership stay explicit and testable.

02AI-assisted reasoning

Event interpretation and evidence linking

AI helps classify the event, retrieve supporting evidence, propose relevant network paths, and explain why a material may be exposed.

03Accountable human

Validate exposure and choose mitigation

Resilience and procurement owners confirm business relevance, reject weak paths, and decide whether to investigate, mitigate, or monitor.

Representative failure mode

A visually convincing path joins unrelated entities or elevates a low-confidence relationship into an urgent business alert.

Designed product control

Source-level evidence, relationship confidence, coverage warnings, business-relevance review, and no autonomous sourcing action.

Illustrative acceptance benchmark

an alert cannot enter the owner queue without a traceable event, location, relationship path, material relevance, confidence state, and visible gap.

Decision record

How I led it, decision by decision.

The executive impact summary is substantiated here through the choices I made, the boundaries I held, and the next investment gate I defined.

Reframe

From a map to an operating workflow

I connected location intelligence to Tier-N dependencies, affected materials, evidence quality, bulk use, and accountable mitigation.

De-risk

Turn feasibility into a release contract

I used Phase 1 evidence to define the MVP boundary, UAT scenarios, acceptance dimensions, human-review controls, and the promotion gate for business readiness.

Validate

Validate the decision, not the visualization

The VOC direction focuses on whether geospatial exposure, material relevance, explainability, batch workflows, and alert routing change a real resilience decision.

Measure

Define trust before automation

Precision, recall, coverage, response time, and human acceptance are product measures because a fast but noisy risk signal creates operational cost.

MVP readiness and evidence

Complete product story. Explicit evidence basis. One promotion gate.

The case shows the intended MVP/UAT-ready product in full. Verified delivery evidence, future-state mechanics, and post-UAT value remain separate so the page can be promoted later without being rebuilt.

01Verified today

Evidence basis

  • Phase 1 established a bounded feasibility POC and business-facing product record; independent acceptance evidence is not available in the public record.
  • Current POC users requested an output-first review of one site-specific product workflow and made authoritative direct-supplier data, supplier location, material context, and material-level volume or share explicit product requirements.
  • The validation model now anchors direct-supplier relationships in authoritative records and uses AI enrichment to fill genuine deeper-tier gaps with visible evidence and confidence; the connected retest and UAT acceptance remain open.
02Future-state MVP

MVP / UAT-ready product system

  • A continuous event-to-location-to-Tier-N-to-material decision chain for procurement and resilience users.
  • Bulk network analysis, geospatial relevance, coverage visibility, evidence qualification, and reviewable uncertainty.
  • Accountable alert routing, human mitigation workflow, decision record, and learning loop for future events.
03Post-UAT scale

Value and scale thesis

  • Value hypothesis: reduce the time from external signal to an owner-ready exposure review without increasing false supplier escalations.
  • Make hidden dependencies, affected materials, concentration risk, and evidence gaps easier to investigate.
  • Shift resilience work from reactive checks toward repeatable, proactive risk intelligence.

Investment-grade product judgment

The discipline required before the next enterprise AI investment.

A polished prototype is only the beginning. The product lead must make value, risk, ownership, evidence, and the next investment decision legible.

Technical gate

Can the product retrieve, relate, and explain the evidence with useful confidence?

Business gate

Does geospatial and Tier-N context improve a real resilience decision for the people accountable for it?

Trust gate

Are uncertainty, sources, and human review clear enough to prevent noisy automation?

Scale gate

Can the workflow move from one material to repeatable analysis without collapsing quality?

This is a capability signal, not a claim that this public case carried a disclosed budget or realized a specific dollar value.

Evidence promotion gate

What converts this future-state MVP into a verified public result.

Approved UAT evidence for usefulness, evidence quality, coverage, false escalation, and owner action.

  1. 01

    Validate geospatial and material-level usefulness with business users

  2. 02

    Test bulk network-analysis workflows

  3. 03

    Define event classification, entity matching, network coverage, false escalation, time-to-qualified-review, and owner-action metrics

Next flagshipNegotiation Pro: Zero‑to‑Close