Operations Case Triage Assistant

What it does

The Operations Case Triage Assistant runs an incoming technical-operations queue (NOC / support / incidents) through a sequential combination of four established frameworks — judgment is anchored to named canon, not gut feel:

1. ITIL 4 Incident Priority Matrix (Impact × Urgency). Each case is crossed on a 3×3 matrix by how many users/systems are affected (Impact) × how fast it is growing / whether work has stopped (Urgency) → a P1-P4 priority and target SLA window.
2. FMEA — RPN (Severity × Occurrence × Detection, 1-10 scale → 1-1000). Captures SLA breach risk: severity of the breach × likelihood/time-in-queue × difficulty of early detection. RPN ≥ 200 bumps the priority a tier, RPN ≥ 336 raises an urgent-escalation flag. Severity 9-10 cases are surfaced regardless of RPN.
3. Eisenhower Matrix (Important × Urgent). Maps the P-band into four routing buckets: Act now · Schedule · Delegate/route · Noise (close) — correct team assignment derives from this anchor.
4. Pareto / 80-20 root-bottleneck clustering. In multi-case mode, finds which 20% of causes (service/customer/error-code) drive 80% of cases; flags the root bottleneck instead of triaging each ticket in isolation.

It then quantifies the SLA window with SRE MTTA/MTTR + Error-Budget (remaining = target − elapsed), optionally gives an Erlang-C "this staffing can't hold this SLA" capacity signal, and produces a 0-100 Operations Health Score (SLA-window health 35% + priority distribution 25% + routing accuracy 20% + bottleneck concentration 20%).

When to use it

• At shift handover, queue pile-up or an incident storm, when you need to decide "what do I look at first, who do I assign what".
• To surface SLA-breach-risk cases first — which fire before you fight the fire.
• To flag the root bottleneck (Pareto) beneath recurring case clusters.
• To quantify a queue's instantaneous triage health and separate out what needs escalation.

Method / frameworks

ITIL 4 Incident Management (AXELOS/PeopleCert) · ISO/IEC 20000-1 incident prioritization · FMEA/RPN (IEC 60812, AIAG-VDA S×O×D scale) · Pareto Principle (Juran) · Erlang-C queueing theory · Google SRE SLI/SLO/Error-Budget & MTTA/MTTR · Eisenhower Important×Urgent matrix. No threshold is invented: if a number is unknown a typical range is given or requested from the user, and gaps are written explicitly into the gaps field.