Stewardship Workflows

How ServiceNow Tasks Keep Data Current #

---

1 Purpose #

No knowledge graph stays accurate without care.
Stewardship Workflows turn EA 2.0 from a passive data store into a living governance mechanism, where gaps, errors, and drifts automatically trigger tasks, owners, and corrections — all managed through ServiceNow GRC or ITSM.

---

2 Why It Matters #

• Architects often build once and forget.
• Data decays 2–5 % per month if not refreshed.
• EA 2.0 detects that decay automatically and routes it to the right human.

Result: A graph that stays evergreen without central admin heroics.

---

3 Core Workflow Concept #

Detect  →  Assign  →  Fix  →  Verify  →  Learn

1. Detect – EA 2.0’s DQ engine or policy triggers anomaly.
2. Assign – Outbound Gateway creates a ServiceNow task.
3. Fix – Steward updates data or metadata.
4. Verify – EA 2.0 re-ingests and recalculates DQ score.
5. Learn – Feedback loop tunes rules and thresholds.

---

4 Trigger Conditions #

Trigger Type	Example	Auto-Priority
Stale Data	last_seen_at > 7 days	Medium
Missing Owner	No owner_id on Application node	High
Broken Lineage	DERIVED_FROM edge missing	High
Low DQ Score	< 0.8	Medium
Policy Violation	Tag or Control non-compliant	Critical

Each trigger maps to a pre-approved ServiceNow Task Template.

---

5 ServiceNow Integration Model #

EA 2.0 → ServiceNow (Table API) → Assignment → EA 2.0 ← Webhook (Closure Sync)

Stage	Action	Example Field Mapping
Create Task	POST to incident or u_steward_task table	short_description = “Missing Owner: APP-142”
Assign	Auto-map to steward via domain registry	assigned_to = Data Steward
Work	Steward adds correction note	work_notes
Close	Webhook calls EA 2.0 API	status = “Resolved”

---

6 Workflow Variants #

Workflow Type	Use Case	Typical SLA
DQ Correction	Invalid metadata, missing link	5 days
Lineage Update	New feed added / schema changed	3 days
Policy Exception Review	Violation needs manual approval	2 days
Risk Control Validation	Control status expired	7 days

---

7 Steward Experience in ServiceNow #

• Custom dashboard shows “EA 2.0 Data Health.”
• KPIs: Tasks by Domain, Avg Age, % Overdue.
• Embedded Power BI tile shows DQ trend for their domain.
• Buttons: Validate → Reassign → Escalate.

---

8 Verification & Closure #

When a steward closes the task, ServiceNow’s webhook sends JSON:

{
  "task_id": "EA2-321",
  "status": "Resolved",
  "resolved_by": "john.doe@org",
  "timestamp": "2025-11-08T14:02Z"
}

EA 2.0 API updates:

• dq_status = valid
• last_verified_by = steward
• Audit edge → (:Task)-[:CLOSED_BY]->(:Person)

---

9 Automation Safety Guards #

• Max tasks per feed per day = 10 (throttling).
• Auto-suppress duplicates within 24 h.
• Escalate if no closure after 2 SLA breaches.
• All automation logs stored in Audit Ledger.

---

10 Metrics & Dashboards #

| Metric | Target | Insight |
|:–|:–|:–|:–|
| Open Steward Tasks | ↓ trend QoQ | Data maturity growth |
| Mean Time to Correction (MTTC) | < 72 h | Responsiveness |
| % Automated Assignments | ≥ 90 % | Integration maturity |
| DQ Revalidation Rate | ≥ 95 % | Trust recovery |
| Escalation Rate | ≤ 5 % | Workload balance |

---

11 Benefits #

✅ Ensures graph never goes stale.
✅ Bridges AI automation with human judgment.
✅ Creates visible accountability loop.
✅ Quantifies data governance ROI through closure rates.

---

12 Cultural Impact #

Stewardship Workflows make data governance feel like DevOps: continuous, measured, and collaborative.
Data stewards see impact immediately in dashboards — governance turns from “paperwork” to performance.

---

13 Takeaway #

AI can detect problems; only people can restore trust.
Stewardship Workflows make sure that trust is maintained in real time, one task at a time.