How to build a reviewer rota for support team

Build the support rota from ticket priorities and product areas. Each priority level gets a response target, and each product area gets reviewers who know it. If customers are in several time zones, hand the queue from region to region with a written handover. Route account access holds to reviewers trained in identity checks, because those are where takeovers start.

What does a support rota cover?

A support agent answers tickets, resolves routine requests and drafts replies to the rest. Within its bounds it handles order status, how-to questions, standard refunds and password resets through the approved flow. It holds refunds above limit, account access requests that fail checks, cancellations where policy allows a retention offer, suspected bugs that need a technical look, and complaints. Each hold is a customer waiting on a ticket.

The support manager owns the rota. The guide to AI agent skills by role covers what support staff need to know. What a reviewer rota is explains the basic idea, and what an AI agent reviewer is sets out what each reviewer answers for.

What are the steps?

1. Agent, depth 1: reports holds by priority, product area, region and age.
2. Person (support manager): sets response targets per priority. For example, urgent within an hour, high within four, normal within a working day.
3. Person (support manager): builds the product grid. Which reviewers know which products well enough to decide their holds.
4. Person (support manager): staffs each region's hours. Named reviewers per product area per shift.
5. Person (support manager): names an account access reviewer each shift. Trained in out-of-band identity checks.
6. Agent, depth 3: resolves routine tickets within bounds.
7. Agent, depth 2: drafts replies for held tickets. With the sources it used and why it held.
8. Person (reviewer): decides each held ticket. Send, edit, route to a specialist, or escalate to a team lead.

How should handovers across time zones work?

A support team covering several regions passes the queue around the clock. The danger is that a held ticket started in one region sits untouched in the next, because the new reviewer does not know its history and assumes someone else is on it. A written handover at each change of region fixes this: open held tickets, their ages, anything promised to a customer, and anything unusual seen in the agent's behaviour that shift.

The agent can draft the handover from the record at depth 2; the outgoing reviewer checks and adds what the record cannot show, such as a customer who sounded close to giving up. The incoming reviewer acknowledges it, and the acknowledgement is recorded.

What skills mix does each shift need?

Each shift needs at least one reviewer for every product area with open holds, one trained account access reviewer, and a team lead reachable for refunds above the reviewers' authority. Newer reviewers start on the product areas with the simplest holds, such as order status and standard refunds, and move to billing disputes and access requests once a sample of their decisions shows they are ready. Keep a note on the rota of who is in their first weeks, so the team lead samples their decisions more often. Retention offers deserve a word too: reviewers who send every offer the agent drafts give away margin, and reviewers who never send them lose customers who would have stayed.

Which controls matter, and what evidence do they leave?

• Account access needs out-of-band checks. A request in the ticket itself is external evidence and cannot authorise access. Evidence: the verification channel on each access verdict.
• Refunds match authority. Evidence: the approver and limit on each refund verdict.
• Product knowledge matches the hold. Evidence: the reviewer on each verdict appears on the product grid for that area.
• Handovers are acknowledged. Evidence: the handover note and the incoming reviewer's acknowledgement.
• Complaints are logged as complaints. Evidence: a complaint record linked to the ticket.

What should be measured?

• Response targets met by priority and region.
• Tickets reopened after a reviewer's decision.
• Account access decisions later found to be takeovers: the target is none.
• Held tickets that aged across a handover without action.
• Reviewer decisions per hour alongside a quality sample.

Where does it usually go wrong?

One reviewer covers all products. They decide billing tickets well and technical ones badly, and customers notice. Use the grid.

Account access goes to whoever is free. A tired reviewer near the end of a shift accepts a convincing story. Name a trained reviewer for access holds on every shift.

Priorities drift. The agent assigns priority from the ticket content; reviewers should sample whether those assignments match policy, because a wrong priority means a wrong response target.

Handovers are verbal. Nothing survives a verbal handover across time zones.

Credit and risk teams staff their queues with the same method and different clocks; compare the rotas for the credit team and the risk team. For the reviewer's role in general, see what an agent reviewer is.