How to distribute customer conversations across a team

Every request should have a state, a person responsible once assigned and a next action. A shared inbox without these rules may merely make disorganisation visible to more people. Start with simple queues, assignment criteria and a destination for exceptions; then choose a tool that demonstrates that journey.

By Lyntev. Updated on .
About 5 minutes to read.

A request that stays connectedThe same reference follows the request and prevents duplicates.

Translation published on . Original Portuguese guide; source revision .

By the end, you can:

  • Distinguish a queue from a person responsible
  • Prepare a shift handover with context
  • Test duplicates, absences and requests without a destination

Define what you are distributing

One conversation may contain more than one request, and a request may continue over several contacts. Decide how the team distinguishes these objects. If a customer asks about a repair and then requests a new quote, two tasks with different states may need to be retained. Assigning only the contact to a person can hide work that remains pending.

Observe real examples privately and turn them into fictitious scenarios for designing rules. Identify the input, expected result and who can decide. Do not collect additional customer data merely because the tool offers fields. The design should make work understandable to the team and reduce requests to repeat information already supplied.

Separate queue, person responsible and state

A queue gathers work with a common characteristic, such as new enquiries or post-delivery support. The person responsible takes on a specific task. The state indicates the request’s situation: unassigned, in progress, awaiting information or completed. These fields answer different questions; a topic label does not demonstrate that someone has taken responsibility for replying.

Use a few queues and states with observable definitions. “Important” or “being handled” can mean different things to each person. Write what allows entry to and departure from each state. Microsoft documentation treats classification and assignment as separate distribution stages; the principle helps organise the process without implying your tool has the same features.

Separate queue, person responsible and state
ElementFictitious exampleVerifiable rule
QueueNew quote requestsRequests not yet commercially assessed
Person responsibleDesignated reception staff memberCan view and take on the next action
StateAwaiting photographThe missing information is identified
Next actionReview received materialHas a person, reason and review time

Choose criteria the team can fulfil

Assign by availability, competence, continuity or another explicit need. Rotating requests equally may seem fair yet be unsuitable if only one person can answer a particular situation. Giving everything to the specialist can also create an invisible queue. Record priority, capacity and the alternative destination when nobody is available.

Distinguish genuine urgency from a word written by the customer. A short commercial deadline needs assessment; a safety risk may require a specific procedure outside the normal flow. Avoid rules promising immediate service without coverage. Distribution should help find a suitable person, not merely change a name on screen to make a request look resolved.

Prepare a handover that avoids starting again

When handing over a conversation, include what the customer wants, what has been checked, the commitment made and the next action. There is no need to copy the entire history into a long note if the person can consult it with appropriate access. A short summary should distinguish customer statements from facts confirmed by the team.

Define when responsibility ends for the previous person and starts for the next. If the destination is only a queue, the request still awaits acceptance. Confirm that someone reviews the queue at the start of the shift. An automatic transfer message does not prove another professional received the context or took responsibility for continuing.

Try changing the person responsible

Receção Azul, a fictitious assistance business, receives a quote request in the morning. The first person identifies the equipment and asks for a photograph. Before leaving, they pass the request to the afternoon team with the state “awaiting material”, an indication of what is missing and a commitment to review after receipt. They do not mark it completed simply because they replied.

When the customer sends the photograph, the team checks whether the task reappears in the right place and someone takes it on. The test uses fictitious data and does not require real customer messages. The expected result is visible continuity, including what happens if the submission arrives during handover. No percentage benefit is presumed from this example.

Test concurrent actions, absence and missing destinations

Ask two people to try to take the same request during the test. Check whether the tool prevents conflict or the manual process detects it before replying. Then test the absence of the person responsible and the arrival of a topic belonging to no queue. There should be a triage destination rather than a limbo with nobody responsible.

Review notifications too. Alerting the whole team about everything can bury important signals; alerting only an absent person can stop operations. Confirm who sees the pending request and how a delay will be detected. Document tool limitations and the necessary procedure without confusing a planned integration with a capability already available.

Measure unassigned work and continuity

Monitor unassigned requests, time until acceptance, repeated handovers and tasks past their review time. These indicators show distribution obstacles. Count the first reply and resolution separately; a quick greeting can coexist with a forgotten quote. Use proportionate data and avoid publishing individual rankings without context about the type of work.

Analyse samples with the team to understand causes: an ambiguous rule, excess volume, a skills gap or incomplete information. Changing tools can help, but does not always resolve responsibility design. Make one change at a time and confirm the request remains identifiable when it changes person, channel or shift.

Document the rule and how to test it

The worksheet should let a new person know where to start and how to ask for help. Record who may change priority and who reviews the unassigned queue. You do not need a complex organisation for a few requests; you need a process that matches the team and does not rely on memory alone.

Review the design when shifts, people or services change. Reuse the conflict, absence and missing-material scenarios before extending automation. The acceptance criterion is the observed journey, not a statement that the system is intelligent. Good assignment ends in a next action someone can carry out.

Worksheet for your project

Object distributed: conversation or request:
Queues and entry criteria:
States and exit conditions:
Assignment criteria and capacity:
Who reviews unassigned work:
Acceptance and shift-handover rule:
Minimum summary and commitment already made:
Destination for absences and unknown topics:
Tests with two people, delay and a new message:
Indicators and person responsible for review:

Sources and criteria

Content prepared with AI assistance and research using technical sources, with checks of the relevant examples and calculations. The references support the guide’s decisions; examples identified as fictitious let you try out the method.

Wellinton Oliveira

Written by

CEO of Lyntev

Technology with context, original analysis and sources you can consult.