A Specialized and Secure AI Orchestrator for Swiss Financial Compliance
View the Project on GitHub Digital-AI-Finance/wecan-innosuisse-ai-draft
Home > Risk & Contingency
The project maintains a risk register across all 24 monthly execution plans. Each month identifies 3 risks with impact assessment, probability rating, and mitigation strategy.
| Metric | Value |
|---|---|
| Total Risks Identified | 72 |
| Risk Categories | 4 |
| Risks per Month | 3 |
| Review Frequency | Monthly (steering committee) |
| Category | Count | Percentage |
|---|---|---|
| External | 28 | 39% |
| Process | 16 | 22% |
| Resource | 10 | 14% |
| Technical | 7 | 10% |
| Recurring (cross-category) | 11 | 15% |
Risks related to model performance, system integration, and infrastructure.
| Month | Risk | Impact | Probability | Mitigation |
|---|---|---|---|---|
| M05 | Model performance below baseline expectations | High | Medium | Hyperparameter tuning, architecture adjustments, ensemble methods |
| M10 | CRM integration complexity higher than expected | High | Medium | Start with simplest CRM, add complexity incrementally |
| M11 | Integration testing reveals blockers | High | Medium | Early integration tests, fallback paths |
| M07 | Integration testing reveals incompatibilities | Medium | Medium | Early integration testing, modular architecture |
| M08 | Hybrid deployment complexity | Medium | Medium | Incremental integration, sandbox testing |
| M13 | API integration complexity higher than expected | Medium | Medium | Incremental integration, sandbox testing |
| M18 | SAP integration authentication issues | Medium | Medium | Early auth testing, OAuth fallback |
Contingency: If model performance targets are not met by MS2 (M6), fall back to ensemble approach combining multiple smaller models. If CRM integration exceeds planned complexity, implement abstraction layer with simplified mappings first.
Risks related to team availability, capacity, and knowledge transfer.
| Month | Risk | Impact | Probability | Mitigation |
|---|---|---|---|---|
| M10 | Resource contention with 4 active WPs | Medium | High | Clear priority matrix, weekly capacity reviews |
| M14 | 4 active WPs strain team capacity | Medium | High | Weekly capacity reviews, task prioritization |
| M01 | WP1/WP2 kickoff delays due to resource availability | Medium | Medium | Early resource planning, backup personnel identified |
| M04 | WP3 kickoff delays due to resource availability | Medium | Medium | Early resource planning, backup personnel identified |
| M13 | Year 2 resource availability | Medium | Medium | Early confirmation, backup personnel identified |
| M14 | WP5 kickoff delays from resource contention | Medium | Medium | Priority matrix, capacity planning |
| M15 | WP3 knowledge transfer incomplete | Medium | Medium | Documentation checklist, handoff verification |
| M21 | WP4 knowledge transfer incomplete | Medium | Medium | Documentation checklist, handoff verification |
| M02 | Team availability constraints | Medium | Low | Cross-training, flexible task assignment |
Contingency: Peak resource contention occurs during M10-M15 when 4 WPs are active simultaneously. Mitigation includes a weekly capacity dashboard and pre-assigned priority rankings per WP. Knowledge transfer between ending and starting WPs must begin 1 month before handoff.
Risks related to deliverable quality, GO/NO-GO criteria, and validation processes.
| Month | Risk | Impact | Probability | Mitigation |
|---|---|---|---|---|
| M04 | GO/NO-GO criteria not fully met | Critical | Low | Weekly progress tracking, early escalation |
| M12 | GO/NO-GO criteria not met | Critical | Low | Contingency time, accelerated completion |
| M11 | WP2 deliverables not ready for MS3 | Critical | Low | Weekly progress tracking, early escalation |
| M24 | 6 deliverables in 152h creates bottleneck | High | High | M22-M23 preparation work, parallel processing |
| M09 | Q3 deliverables quality concerns | High | Low | Early draft reviews, peer review process |
| M12 | Deliverable quality below standard | High | Low | Early quality reviews, peer validation |
| M16 | 300 forms validation incomplete | High | Medium | Prioritize high-impact forms, parallel processing |
| M19 | 300-case validation incomplete | High | Medium | Parallel processing, prioritize critical cases |
| M21 | D4.1-D4.3 quality not meeting acceptance | High | Low | Early quality reviews, peer review process |
Contingency: All GO/NO-GO milestones have contingency time built into the preceding month. If MS1 criteria are at risk, the steering committee can grant conditional approval with remediation deadlines. Final month (M24) bottleneck is mitigated by starting deliverable drafts in M22-M23.
Risks related to partners, market conditions, and external dependencies.
| Month | Risk | Impact | Probability | Mitigation |
|---|---|---|---|---|
| M06 | Pilot partner onboarding delays | High | Medium | Streamlined onboarding process, dedicated partner liaison |
| M07 | Training approach scaling issues | High | Medium | Early performance testing, fallback architectures |
| M09 | CRM access delays block WP4 kickoff | High | Medium | Early credential requests, sandbox alternatives |
| M11 | 500 document milestone at risk | High | Medium | Buffer documents, prioritize processing |
| M14 | Form format variability higher than expected | High | Medium | Flexible schema design, format abstraction layer |
| M20 | Security audit reveals critical findings | High | Low | Pre-audit internal review, remediation buffer |
| M20 | TRL 5-6 evidence insufficient | High | Medium | Document all demonstrations, pilot feedback collection |
| M03 | Document volume insufficient for statistical significance | High | Low | Prioritize high-value documents, request additional samples |
| M02 | Pilot partner recruitment slower than expected | Medium | Medium | Expand outreach network, leverage Wecan industry contacts |
| M22 | Open benchmark dataset licensing issues | Medium | Medium | Early legal review, Creative Commons default |
Contingency: Pilot partner risks are the highest external concern. The project maintains a pipeline of 8-10 potential partners (via Wecan network) to ensure 3-5 can be confirmed by MS4. Document volume targets include 20% buffer. Security audit is scheduled with 4-week remediation window.
| Trigger | Response | Owner |
|---|---|---|
| Task completion < 80% in any month | Escalation to steering committee | WP Lead |
| GO/NO-GO criterion at risk (4 weeks before) | Remediation plan required | Project Manager |
| Resource contention > 110% capacity | Task reprioritization, overtime approval | Project Manager |
| Pilot partner drops out | Activate replacement from pipeline | Wecan |
| Security vulnerability (critical) | Immediate patch, re-audit scheduled | FHGR + Wecan |
Source: Risk registers from 24 monthly execution JSON files in the execution/ folder