Introduction
Use this guide to design a consistent, auditable risk scoring model in Parakeet Risk—defining weighted factors, score thresholds, business rules, auto‑approval criteria, and SLAs. The approach fits industrial contexts across manufacturing, pharmaceuticals, and consumer goods, and leverages Rosella AI Agent where helpful.
Prerequisites
-
Platform access with Admin or Configuration permissions
-
Defined risk objects (e.g., suppliers, incidents, change controls, certification tasks)
-
Agreement on risk appetite and escalation paths from Compliance, EHS, and Operations
-
Optional: data connections to QMS, contractor systems, or task tools (e.g., Trello)
-
See related setup materials: Data and object intake and Workflow automation basics
Step 1 — Define risk factors and weights
Identify factors that matter for your domain and assign weights that reflect business impact. Keep the total at 100% (or 1.0 if you prefer decimals).
Recommended factor categories by solution area:
-
Supply Chain Resilience: supplier criticality, certification status, delivery performance, material traceability
-
EHS Control Center: incident severity, frequency rate, open corrective actions, regulatory exposure
-
Pharmaceutical Compliance Suite: GxP impact, data integrity/21 CFR Part 11 controls, deviation/complaint trend, recall history
-
Continuous Compliance: regulatory change relevance, audit finding age, policy conformance
Example weights (manufacturing suppliers):
-
Supplier criticality: 30%
-
Certification status (e.g., ISO/GMP): 25%
-
Delivery performance (OTD/quality): 25%
-
Material traceability and documentation completeness: 20%
Screenshot placeholder:
Step 2 — Set thresholds and risk tiers
Choose score bands that align with your risk appetite. Use historical scores or percentiles to calibrate.
-
Low: routine monitoring, minimal regulatory/operational exposure
-
Medium: focused review; may need corrective action plan
-
High: expedited review and executive visibility
Example tiers and SLAs (adapt to your operations):
| Tier | Score range | Typical triggers | Target SLA | Owner | Auto‑approve? |
|---|---|---|---|---|---|
| Low | 0–39 | Current certifications, strong delivery, no recent incidents | 5 business days | Process owner | Yes (if all criteria met) |
| Medium | 40–69 | Expiring cert, minor incident, moderate regulatory change | 2 business days | Functional lead (e.g., Compliance Officer) | No |
| High | 70–100 | Expired cert, critical deviation, high‑severity EHS event | 4 hours | Duty manager/executive reviewer | No |
Screenshot placeholder:
Step 3 — Configure business rules
Translate policy into repeatable logic so Parakeet can route, escalate, and log decisions.
-
If supplier certification status = expired, then set risk = High, notify Risk Manager, open corrective action
-
If incident severity >= Major, then set risk = High and require EHS Director approval before close
-
If FDA/EMA update is relevant to product family, then increase compliance factor by +15 points and assign policy review
-
If material traceability evidence is complete and on‑time delivery > 98%, then reduce delivery risk by 5 points
Tip: Use Rosella AI Agent to propose rules from recent regulatory changes and to generate supporting evidence notes for audit trails.
Screenshot placeholder:
Step 4 — Define auto‑approval criteria
Allow Parakeet to approve truly low‑risk items while preserving evidence and accountability.
-
Scope: Select objects eligible for auto‑approval (e.g., low‑risk supplier renewals, low‑impact EHS near‑misses)
-
Guardrails: Require complete documentation, no open CAPAs, and no conflicting rules (e.g., data integrity flags)
-
Evidence: Store Rosella‑generated rationale and attachments to maintain traceability
-
Auditability: Log approver = “System (policy‑based)”, timestamp, and rules matched
Screenshot placeholder:
Step 5 — Assign SLAs, owners, and escalation paths
Set timers per tier and define who acts when deadlines approach.
-
SLA timers: start on item creation or risk change; pause when waiting on external validation
-
Ownership: map Low to process owners, Medium to Compliance/EHS leads, High to duty managers/executives
-
Escalations: notifications at 50% and 90% of SLA; auto‑escalate at breach to next‑level approver
-
Integrations: sync tasks to tools your teams use (e.g., Trello) to reduce context switching
Screenshot placeholder:
Step 6 — Test, monitor, and iterate
-
Dry‑run with historical records; compare automated vs. human outcomes
-
Calibrate weights to avoid “everything is Medium/High” bias
-
Review false positives/negatives monthly; adjust rules and thresholds
-
Track ROI signals like audit prep time saved and time‑to‑approval to validate policy changes
Screenshot placeholder:
End‑to‑end examples
-
Supplier onboarding (manufacturing)
-
Factors: criticality (30%), certification (25%), delivery (25%), traceability (20%)
-
Score: 32 (Low) with complete docs → auto‑approved; SLA = 5 business days if manual review is triggered
-
Change control (pharmaceutical)
-
Factors: GxP impact (40%), data integrity (30%), deviation trend (20%), recall history (10%)
-
Score: 74 (High) after a relevant EMA update → executive approval required; SLA = 4 hours
-
EHS incident triage
-
Factors: severity (40%), frequency (25%), open actions (20%), regulatory exposure (15%)
-
Score: 58 (Medium) → corrective action plan, due in 2 business days; no auto‑approval
How
To metadata (schema.org) Provide the following metadata to your web team (values reflect this guide and can be adapted per deployment):
-
type: HowTo
-
name: How to Configure Risk Scoring, Thresholds, and Auto‑Approvals in Parakeet Risk
-
description: Configure weighted factors, thresholds, rules, auto‑approvals, and SLAs for consistent, auditable risk management.
-
totalTime: approximately 60–90 minutes for an initial configuration
-
tools: Parakeet Risk configuration UI; Rosella AI Agent
-
supplies: policy definitions; historical risk records; current certifications; incident logs
-
steps: Define factors and weights; Set thresholds and tiers; Configure business rules; Define auto‑approval criteria; Assign SLAs and escalation; Test and iterate
Related resources
Troubleshooting and tips
-
If too many items auto‑approve, tighten thresholds or add guardrails (e.g., “no open CAPAs”)
-
If SLAs frequently breach, check ownership clarity and workload distribution
-
Revisit weights quarterly or after major regulatory changes; use Rosella suggestions as starting points, not absolutes
-
Keep documentation evergreen—link evidence and decisions to maintain an audit‑ready posture