Change Risk Prediction (CRP)
CRP is a stand-alone, human-in-the-loop AI platform that predicts change request risk in real time using historical ITSM data. It gives confidence scores, factor explanations, and prescriptive mitigation steps to help teams make faster, safer decisions.
Discovery
We conducted interviews and a survey to uncover current pain points and opportunities for AI-driven risk prediction.
Research Activities
- 8 semi-structured interviews (Change Managers, Owners, ITSM Analysts)
- Survey of 38 ITSM practitioners
- Competitive review: ServiceNow & BMC
- Affinity mapping & synthesis
Research & Insights
User Personas
Primary personas: Change Manager (prioritizes approvals & confidence) and Change Owner (needs clear remediation guidance).
Pain Points
- Information is fragmented across CMDB, ticket history, test logs
- Risk scores are often black-box and not trusted
- Manual checks create high cognitive load and slow CAB decisions
- Users need prescriptive steps — not just a score
Opportunities
- Unified risk detail panel with top contributing factors
- Explainability: show confidence & factor breakdown
- Prescriptive mitigation steps and similar-change comparisons
- Real-time scanning & smooth ServiceNow/BMC integration
Ideation
We defined user flows and prioritized screens to reduce time-to-decision and increase trust in AI predictions.
Primary User Flow
- Login (SSO)
- Connect Data Sources (ServiceNow / BMC)
- Initialization & historical data training
- Landing: Risk overview & filters
- Detail panel: score, confidence, factors, similar changes, recommended steps
Key Screens
_show.png)
Note: these images are placeholders — replace with your exported mockups or Figma assets. They are responsive and will scale to fit.
Wireframes & Mockups
From lo-fi wireframes to hi-fi UI, we iterated on the explanation UI, confidence indicators, and prescriptive guidance components.
Risk Overview
Landing page with filters, upcoming timeline, and quick risk triage list.
Detail Panel
Selected change shows score, confidence, top factors, similar past changes and step-by-step mitigation.
Outcome
The CRP design focuses on trust, clarity, and actionability — it reduces evaluation time and provides human-in-the-loop controls for critical decisions.
Impact (projected)
- 40–60% faster risk evaluation
- Higher confidence in approvals due to explainability
- Fewer failed changes with prescriptive mitigation
Next Steps
- Export hi-fi screens to Figma and create component library
- Build prototype for pilot integration with ServiceNow
- Run usability tests with CAB participants
E-Commerce Platform
Reimagining online shopping with intuitive navigation and seamless checkout
