Count the messages that need a decision
Before adding a second assistant, separate the inbox into messages that need a decision and messages that need only a response. Pull two weeks of conversations and tag each one by the action it requires. A shared inbox often looks overloaded because a small number of decision-heavy messages block the queue. Adding a person does not fix a decision bottleneck.
Record who currently makes each decision and how long the decision takes. If the same owner reviews every refund or exception, the queue will wait for that owner no matter how many assistants answer email. The audit should show whether the constraint is message volume, decision authority, or missing information. A second assistant helps only with the first and third.
Find the unowned conversations
List the conversations with no assigned owner or no next action, and note how long they have been in that state. An unowned message is not the same as an unread message, because a message can be read and still have no one responsible for the next step. Count the unowned conversations separately from the total backlog. That number is the clearest sign that ownership, not staffing, is the problem.
Check whether the unowned messages cluster around a particular label, customer, or time of day. A cluster may point to a missing rule, a handoff gap, or a queue that nobody checks after a certain hour. Record the cluster and the person who should own it. The audit should end with a proposed owner for each cluster, not a general request for more help.
Measure the time to first reply
Measure the time from arrival to first reply, and separate automated acknowledgements from a real human response. A fast auto-reply can hide a slow answer, so record both. Look at the median and the oldest conversations rather than the average alone. The oldest messages show where the process has actually failed.
Compare the reply time with the customer promise and with the business's own service expectation. If the promise is shorter than the observed time, the gap is the real problem to solve. Record whether the delay happened before or after the message was assigned. That split tells you whether the queue or the decision is slow.
Map who owns each label
Write down every label or folder in the shared inbox and name the person who owns the next action for that label. Include the backup owner and the escalation path. A label without an owner is a place where messages go to wait. The map should be short enough to review in one meeting and specific enough to assign work.
Check whether the labels match the work the team actually does. An old label may collect messages that belong in a different queue, while a new service may have no label at all. Remove or rename labels that no longer describe the work, and update the routing rules at the same time. The inbox structure should follow the work, not the history of the company.
Test a two-person handoff
Before hiring, simulate a two-person inbox by asking a colleague to take one label for a day. Watch how the handoff works, where the second person needs context, and whether the customer receives duplicate replies. Record the questions the second person asked. Those questions are the training material for the new assistant.
Use a shared assignment rule rather than a private understanding of who takes what. Require the owner to acknowledge a reassigned conversation before the first person stops watching it. Check that both people can see the same customer history. A handoff that works in a test is more reliable than a promise that the new person will figure it out.
Set a rule for duplicate replies
Two assistants can reply to the same message when the assignment is unclear, and duplicate replies confuse the customer. Write a rule that only the assigned owner replies, while the other person adds an internal note if they have context. The rule should be visible in the inbox and repeated during onboarding. Duplicate replies are a symptom of unclear ownership, not just a training issue.
Record every duplicate reply and review the cause. If the cause is a slow assignment, fix the routing. If the cause is a missing label, add the label. If the cause is two people watching the same queue, change the schedule. The review should focus on the process that allowed the duplicate.
Decide with the audit evidence
Use the audit to choose one of three actions: add an assistant, change the ownership model, or fix the decision path. If most messages wait for one approver, a second assistant will not help. If the queue is large but decisions are fast, a second assistant may be justified. Write the evidence and the chosen action in one page.
Set a review date and a measure for the chosen action. If the business hires, check whether the unowned count falls and the first-reply time improves. If it changes ownership instead, check the same measures. The audit is useful only if it changes a decision and the result is checked later.
If the queue and decision rights are already defined but need a steady operator, outsourced operations support can turn approved examples, decision lines, and review rules into a draft your manager can check. The manager still owns exceptions and final approval.
For the next step, compare this workflow with Design a weekend coverage rota for a Philippines-based support team and Write a handover note that survives a shift change. For access and oversight, use the authoritative planning reference alongside your own policy.
