ReviewDesk
← Back

ReviewDesk guide · By the Ingenious AI team · Published 14 November 2026 · Reviewed 14 November 2026

How should human-review cases be assigned across a team?

Generate a free adverse-outcome letter — no sign-up Build a free pre-use notice See pricing

The short answer. Assign every case to a named person whose competence matches the decision type, balance queues by age and weight rather than by count, and make reassignment a clean, recorded, stigma-free move. Assignment is where review quality is decided before any reviewing happens: the right file in the right hands with time to judge it. ReviewDesk's SLA workflow carries assignment on the case itself, alongside its deadline and history.

What is review case assignment?

Case assignment is the allocation of human-review cases to individual reviewers: who owns this case, on the basis of what competence and capacity, with what visibility to the team. In ReviewDesk, assignment is a case property in the SLA workflow — the assignee is named on the case, deadlines travel with reassignment, and the decision history records who decided.

Why named owners, never team queues?

Because deadlines are felt by people, not by groups. A case owned by 'the review team' is owned at the exact moments nobody is looking at it, which is when cases age. A case owned by a named reviewer appears in one person's morning queue with its clock showing; its ageing is somebody's visible fact. Named ownership also makes the decision history meaningful — 'decided by' is a person whose judgement the record can stand behind — and makes quality sampling possible, because consistency can only be measured across identifiable reviewers.

Teams still matter: the inbox owner's board shows all assignments, and cover is arranged by moving cases between named people, each move recorded. The unit of accountability is the person; the unit of resilience is the team. Systems that confuse the two — group ownership for comfort, individual blame at audit — get the worst of both. ReviewDesk's model is explicit: the assignee is on the case, the case is on the board.

How should cases be matched to reviewers?

By decision competence first: hiring reviews to people who can judge the role's real requirements, underwriting reviews to underwriters, tenancy reviews to those with letting authority. By independence second: not the person who made or configured the original decision, and where the original decision-maker is the only competent reviewer — small teams, specialist niches — the case records that fact and the review proceeds as a distinct, documented act, per the review standard. By load third: competence without capacity produces rushed review, which is the failure mode assignment exists to prevent.

Develop range deliberately. A bench where only one person can review a decision type is a bench one resignation from crisis. Pair junior reviewers with sampled oversight, rotate decision types quarterly among the competent, and let the decision history be the training corpus — new reviewers who read fifty closed cases learn the house standard faster than any induction deck can teach it.

How should load be balanced — count, weight or age?

Weight and age, with count as a sanity check. Cases differ enormously: a namesake correction may take fifteen minutes; a contested underwriting review may take half a day across sittings. Balancing by raw count quietly overloads whoever draws the heavy cases and teaches reviewers to prefer light ones. Weight by decision type using your own completion data — ReviewDesk's real-case metrics give you the observed durations to weight with — and prioritise within every queue by age and deadline proximity, so no case becomes the one everyone steps around.

Publish the balancing rule. Reviewers accept hard assignments from a system they can see is fair; they resist, quietly and effectively, assignments that feel like favouritism or punishment. The board's transparency is the fairness mechanism — everyone's queue is visible, weights are known, and the owner's reallocations are events on the record rather than whispers.

How does reassignment stay clean?

The case moves whole: file, the person's request, the clock with its provenance, the history so far, and a one-line handover note from the original assignee — what they had established, what remains. The applicant's status view shows 'in review' throughout; internal choreography is not the applicant's concern. Reassignment reasons are operational vocabulary — capacity, competence match, absence, conflict — recorded without commentary on the previous owner's performance.

Watch reassignment frequency per case. A case reassigned three times is not being managed; it is being avoided, usually because it is hard, sensitive or blocked externally. The owner's board should surface multiply-moved cases for a decision about resourcing, not a fourth move. The escalation path in the escalation guide is where such cases belong, with the blocker named rather than the case circulated.

How does assignment serve consistency?

Consistency is an assignment product before it is a training product. If similar cases reach reviewers with similar standards, outcomes cohere; if routing is random, outcome variance measures your routing, not your reviewers. Sample across reviewers monthly — same decision type, similar facts, compare outcomes and note quality — and feed the findings into assignment: pair divergent reviewers on calibration cases, clarify criteria where honest judges diverge, and record the clarification where the next reviewer will find it.

The decision history is the instrument panel for all of this, which is why the decision-history guide treats it as a first-class record. Assignment data plus outcome data plus note sampling is the complete quality system for a review bench; most organisations run the first, glance at the second and skip the third, then wonder why their review quality is a rumour rather than a measurement.

What changes as the bench grows?

Specialisation, supervision and a real owner role. Past a handful of reviewers, route by decision-type pods, give each pod a lead who samples quality and holds the escalation queue, and keep the inbox owner as the cross-pod balancer watching age and clocks. ReviewDesk's workspace holds the structure without new tooling: assignments, deadlines and histories live on the cases, and the boards aggregate at whatever level leadership needs to see.

Resist the enterprise temptation to add approval layers. A review that needs a second signature for ordinary cases doubles latency and halves ownership; reserve second-pair review for defined high-consequence types and for sampling, not for everything. The point of the bench is distributed judgement with recorded accountability — the architecture this whole programme keeps choosing, because it is the one that survives contact with volume.

Your five actions for review case assignment

  1. Assign every case to a named reviewer — team queues are where deadlines go to be unfelt.
  2. Match by competence, then independence from the original decision, then real capacity.
  3. Balance by case weight and age using your own completion data, and publish the balancing rule.
  4. Move cases whole on reassignment — file, clock, history, one-line handover — and surface cases moved three times for a resourcing decision.
  5. Sample across reviewers monthly for consistency, and let the findings tune assignment and criteria.

Worked example: a two-person bench that stopped dropping cases

A fictional insurer's review load lands on two handlers — Amara and Dev — allocated by whoever opens the inbox first. The result is predictable: Dev, faster to the queue, accumulates the heavy underwriting reviews; Amara clears corrections; two cases age unseen during Dev's leave because they lived in his personal list. The fix is structural, not personal: cases are assigned on arrival by type and weight, both queues sit on one board, and leave cover is a recorded reassignment of live cases with handover notes.

A quarter later the same two people carry twenty per cent more volume with no overdue cases, and the monthly sampling shows their outcome notes converging in quality — because each now reviews a deliberate mix, and calibration cases are part of the assignment rhythm. The lesson generalises: most 'reviewer performance' problems are assignment design problems wearing a name badge. Fix the routing, and the people you already have become the bench you needed.

Frequently asked questions

Why must cases have a named assignee?

Deadlines and quality attach to people, not groups. A named owner makes ageing visible, the decision history meaningful and quality sampling possible. Teams provide cover by recorded reassignment between named people.

How should cases be matched to reviewers?

Competence for the decision type first, independence from the original decision second, real capacity third. In small teams where the original decision-maker must review, record that fact and conduct the review as a distinct documented act.

How is load balanced fairly?

By weight and age, using observed completion times per decision type — not by raw count, which overloads whoever draws heavy cases. Publishing the rule and the queues makes the fairness visible.

What makes reassignment clean?

The case moves whole — file, request, clock with provenance, history — with a one-line handover note, recorded as an event. Cases reassigned repeatedly should escalate for a resourcing decision instead of moving again.

Is this guide legal advice?

No. ReviewDesk is templates and workflow tooling, not legal advice. Assignment is described as ReviewDesk's SLA workflow implements it; consult counsel for your obligations.

Keep reading: What should a review case's decision history record? · How does the 3-year evidence vault export a complete decision file? · What should you ask a software vendor before you rely on their tool for consequential decisions? · ADMT duties explained for software vendors and the businesses that deploy their tools · How do you set up your first ADMT workflow, step by step? · All ReviewDesk guides · ReviewDesk home · Try the free desk · Free letter generator · Free notice builder · Pricing

ReviewDesk is built by Ingenious AI. Templates and workflow tools — not legal advice.