Operational checklist to audit manual conversation assignment in WhatsApp: steps, criteria and warning signs

What this checklist is for

If you handle WhatsApp with a team and assign conversations by hand, sooner or later one slips through: a message left unanswered, two agents answering the same thing, a customer who waited forty minutes because no one took it. This checklist is for auditing how you are assigning today, finding the weak points and leaving behind a procedure that anyone on the team can follow and verify.

It doesn't matter whether you use the WhatsApp Business app, WhatsApp Web or a tool with a shared inbox like Wando: manual assignment has the same risks and is audited with the same logic. The difference is that in a shared inbox you can see the status of each conversation and who has it assigned; in the regular app, not even that.

What manual assignment is and when it makes sense

Manual assignment is when a person decides which agent answers each conversation. It can be the business owner, a coordinator or the first person who sees the message. It makes sense when the volume is low, when conversations require judgment (for example, routing to a specific salesperson) or when the team is small and everyone knows each other.

But it has a cost: it depends on someone watching the inbox, knowing who to assign each thing to and doing it in a timely manner. If there is no written criteria, each agent assigns however they see fit. And when the volume grows, the first thing that breaks is precisely the assignment.

Before auditing: what you need to have on hand

• Access to the inbox or app where messages arrive, with permission to see conversations assigned to other agents.
• The team list: who handles conversations, at what hours and with what responsibilities.
• The last 48 hours of conversations, at minimum, to review real cases.
• A document or spreadsheet to note the findings (the checklist itself works).
• The assignment criteria you use today, if it exists. If it doesn't exist, that is the first finding.

Checklist: what to review at each stage

This checklist is meant to be applied in a 30 to 60 minute session, with the inbox open and a real case for each point. It's not for reading from memory: it's for checking off.

Stage 1: Base setup

• Do all the agents who respond have their own account or user in the tool? If two people share the same WhatsApp session, you can't audit anything: you don't know who answered what.
• Are the business hours defined and known by the whole team? If there are no hours, there's no way to know if a response took too long.
• Is there a written assignment criterion? It doesn't have to be a hundred-page manual: it can be a three-line list. If it doesn't exist, write it during this same audit.
• Does the team know what to do with a message that isn't theirs? If there's no procedure for reassigning, everyone will handle it as best they can.

Stage 2: Review of recent conversations

• Take the last 20 incoming conversations and check: do all of them have a response? If any was left unanswered, that's a sign of an error.
• Of those that have a response, who responded? Was it the right person according to the criterion you defined?
• Are there conversations that two different agents responded to? That indicates there's no clear view of who has each conversation.
• How long did the first response take in each case? Note the times: later you'll define an acceptable maximum.
• Are there conversations that were assigned to someone who isn't on shift? That's a classic: the message arrives, the first person who sees it takes it, and if that agent left without reassigning, the customer waits until the next day.

Stage 3: Edge cases

• A customer writes twice: does the same agent who responded the first time see the second message? If there's no way to know, it's a context problem.
• A customer writes during a time when no one is assigned: is there an automatic message that notifies them? If there isn't, the customer doesn't know if they'll be attended to.
• An agent gets sick or leaves halfway through the day: who takes their conversations? If there's no handover procedure, the conversations are left orphaned.
• A message is for sales but reaches customer service: is there a way to forward it without the customer having to repeat everything?
• A customer asks to speak with a specific person: does the current agent know who to forward them to?

Stage 4: Warning signs

These are the signs that manual assignment is failing. If you find any of them, it is not an isolated case: it is a symptom that the process needs a change.

What to do when something fails

When you detect a failure, the first step is not to shut everything down: it is to contain the specific case and then correct the process. The sequence is this:

1. 1Respond to the message that was left unattended: the coordinator or the available agent does it, even if they are not the originally assigned one.
2. 2Notify the team through the internal channel (WhatsApp group, Slack) that there is an unattended conversation and who took it.
3. 3Log the failure in an incident spreadsheet: what happened, when, who should have taken it and why it did not happen.
4. 4Check whether it is an isolated case or a pattern: if it happened more than once in the week, the problem is with the process, not with a person.
5. 5Adjust the process: if the person who left did not reassign, the end-of-shift procedure must include mandatory reassignment.
6. 6Verify the next day that the adjustment worked: review the conversations from the last 24 hours again.

When it is worth switching to automatic assignment

Manual assignment has a ceiling. When volume grows, the cost of coordinating by hand outweighs the benefit. Some signs that you have reached that point:

• You spend more than 30 minutes a day deciding who answers each message.
• Messages slip through even when the whole team is working.
• Customers complain about delays or about being answered by two different people.
• The team doesn't know who to turn to when a conversation gets complicated.
• Conversations get lost when someone is away.

At that point, a tool with automatic assignment based on criteria (like Wando, which also suggests replies with AI that the team approves with one click) can solve the underlying problem. But the audit is still useful: the assignment criteria you defined by hand are the same ones you'll configure in the tool.

How to measure whether assignment improved

After applying the changes, define three simple metrics and review them once a week:

Frequently asked questions

This section answers the most common questions when auditing manual assignment.

How often should I audit?

If the team is small and the volume is low, a monthly audit is enough. If you notice signs of error (unanswered messages, double replies), audit as soon as they appear. After any process change, audit again a week later to verify that it worked.

What do I do if I don't have a tool with a shared inbox?

You can still audit, but you will have limitations: you won't be able to see other agents' conversations or the full history. In that case, the audit is limited to reviewing the messages that come to you and asking the team how they are handling the cases. If the volume justifies the audit, it probably also justifies a tool.

Is manual assignment always bad?

No. For a business that receives ten messages a day and is handled by a single person, manual assignment is all it needs. The problem appears when there is more than one agent, rotating shifts, or variable volume. That's where manual assignment without a written process becomes fragile.

What happens if an agent made the mistake?

First, contain the case: reply to the message and notify the team. Then talk to the agent to understand what happened. Most of the time it isn't negligence: it's that there was no clear procedure or that the tool didn't show the conversation status. The goal of the audit is to improve the process, not to punish people.