Vendor Risk Assessment
Evaluate any AI/SaaS vendor across 6 risk dimensions. Outputs a scored report with go/no-go recommendation.
When to Use
- Onboarding a new SaaS or AI vendor
- Annual vendor review cycle
- Evaluating build-vs-buy decisions
- Due diligence for partnerships or acquisitions
- Compliance requirements (SOC2, ISO 27001, GDPR)
How to Use
The user provides vendor details (name, product, website, any available documentation). The agent researches and scores the vendor across 6 dimensions.
Input Format
Vendor: [Company Name]
Product: [Product/Service Name]
Website: [URL]
Use Case: [What you'd use it for]
Data Sensitivity: [low/medium/high/critical]
Additional Context: [Any docs, certifications, or concerns]
Assessment Framework
6 Risk Dimensions (each scored 1-10)
1. Security Posture
- SOC2 Type II certification?
- Penetration testing cadence
- Encryption (at rest + in transit)
- Access controls and authentication
- Incident response plan
- Bug bounty program
2. Data Handling & Privacy
- Data residency and sovereignty
- Data retention and deletion policies
- Sub-processor transparency
- GDPR/CCPA compliance
- Data portability (can you get your data out?)
- AI training opt-out policies
3. Compliance & Certifications
- SOC2, ISO 27001, HIPAA, FedRAMP
- Industry-specific (PCI-DSS, HITRUST, etc.)
- AI-specific (EU AI Act readiness, NIST AI RMF)
- Audit frequency and transparency
- Regulatory track record
4. Financial Stability
- Funding stage and runway
- Revenue indicators (public or estimated)
- Customer concentration risk
- Acquisition risk
- Pricing stability history
5. Operational Resilience
- Uptime SLA and historical performance
- Disaster recovery plan
- Multi-region availability
- Dependency on single cloud provider
- Support responsiveness and escalation paths
- Change management process
6. Contractual Terms
- Termination and exit clauses
- Liability caps and indemnification
- IP ownership clarity
- Auto-renewal traps
- Price increase limitations
- SLA breach remedies
Output Format
# Vendor Risk Assessment: [Vendor Name]
**Date:** YYYY-MM-DD
**Assessor:** AI Agent (AfrexAI)
**Data Sensitivity Level:** [low/medium/high/critical]
## Overall Risk Score: [X/10] — [LOW/MEDIUM/HIGH/CRITICAL]
## Dimension Scores
| Dimension | Score | Risk Level | Key Finding |
|-----------|-------|------------|-------------|
| Security Posture | X/10 | LOW/MED/HIGH | ... |
| Data Handling | X/10 | LOW/MED/HIGH | ... |
| Compliance | X/10 | LOW/MED/HIGH | ... |
| Financial Stability | X/10 | LOW/MED/HIGH | ... |
| Operational Resilience | X/10 | LOW/MED/HIGH | ... |
| Contractual Terms | X/10 | LOW/MED/HIGH | ... |
## Recommendation: [APPROVE / APPROVE WITH CONDITIONS / REJECT]
## Critical Findings
- [Finding 1]
- [Finding 2]
## Mitigation Requirements (if Approve with Conditions)
1. [Requirement 1 — deadline]
2. [Requirement 2 — deadline]
## Research Sources
- [Source 1]
- [Source 2]
Scoring Guide
- 9-10: Excellent — minimal risk, enterprise-grade
- 7-8: Good — acceptable for most use cases
- 5-6: Moderate — proceed with caution, mitigations needed
- 3-4: Poor — significant concerns, conditional approval only
- 1-2: Critical — recommend rejection or major remediation
Overall Risk Calculation
- Average of 6 dimensions, weighted by data sensitivity:
- Low sensitivity: equal weights
- Medium: Security 2x, Data 2x
- High: Security 3x, Data 3x, Compliance 2x
- Critical: Security 4x, Data 4x, Compliance 3x, Financial 2x
Research Process
- Check vendor website for security/compliance pages
- Search for SOC2/ISO certifications and trust pages
- Check status pages for uptime history
- Search for breach history or security incidents
- Review pricing page for contract terms indicators
- Check Crunchbase/LinkedIn for financial stability signals
- Search for customer reviews mentioning reliability/support
Pro Tips
- Request the vendor's SOC2 Type II report directly — if they hesitate, that's a signal
- Check their status page history (statuspage.io, etc.) for real uptime data
- For AI vendors specifically: ask about model training on your data, output ownership, and hallucination liability
- Compare their security page to competitors — vague = red flag
Need help managing vendor risk across your entire stack? AfrexAI builds autonomous AI agents that monitor vendors continuously — not just at onboarding. Visit afrexai.com or book a call: calendly.com/cbeckford-afrexai/30min