Customer service with contextالعربية

Every customer conversation has an owner, the right context and a reviewed answer

Give the agent conversation ownership, the customer's record and a draft from approved knowledge to review, so routine answers take less searching and nothing is sent without a person.

Built on

Where included

Does this sound familiar?

Customer messages arrive on more than one channel and land with whoever is free. Agents search old chats and documents for the same answers, handovers lose the thread, and nobody can say who owns an open conversation.

Who is involved

  • Service agent
  • Service team lead
  • Knowledge owner who approves answer content
  • Quality reviewer, where included

From trigger to outcome

Follow the work from the first trigger. Each arrow names what passes along it and carries its evidence status in words, so a planned step never looks like a working one.

Products in this map

Trigger-to-outcome map: Every customer conversation has an owner, the right context and a reviewed answerEach arrow is labelled with what passes and its evidence status. The table after the map lists every connection in words.Trigger: A customer starts aconversation on an activated messagingchannel, or in website chat once it isreleased (website chat is demonstratedin testing).Website chatCustomer writes iniConnect InboxAgent owns the conversationConversation + ownerDemonstrated in testingWhatsAppCustomer writes in, once activatediConnect InboxAgent owns the conversationConversation + ownerConfigured per deploymentiConnect InboxAgent owns the conversationCRMCustomer record beside theconversationConversation linked to the contactConfigured per deploymentApproved knowledgeContent your company approvediSupport AISuggests a draft to reviewApproved answer sourcesConfigured per deploymentiSupport AISuggests a draft to reviewiConnect InboxAgent owns the conversationDraft for a person to reviewConfigured per deploymentCRMCustomer record beside theconversationiSupport AISuggests a draft to reviewCustomer history informs the draftPlannediConnect InboxAgent owns the conversationiQualityWhere includedReviewer samples conversationsSampled conversation for reviewDemonstrated in testingiSupport AISuggests a draft to reviewiQualityWhere includedReviewer samples conversationsAI drafts included in qualityreviewPlannediQualityWhere includedReviewer samples conversationsiSightWhere includedProtected team summaryProtected team summaryDemonstrated in testingWhat your business gainsLess searching for routineanswers, with clearresponsibility and continuity athandover.

How to read the lines

  • Solid line: Demonstrated in testing
  • Dotted line: Configured per deployment
  • Dashed line: Planned
Each connection in words
From:Website chatTo:iConnect InboxWhat passes:Conversation + ownerDemonstrated in testing
From:WhatsAppTo:iConnect InboxWhat passes:Conversation + ownerConfigured per deployment
From:iConnect InboxTo:CRMWhat passes:Conversation linked to the contactConfigured per deployment
From:Approved knowledgeTo:iSupport AIWhat passes:Approved answer sourcesConfigured per deployment
From:iSupport AITo:iConnect InboxWhat passes:Draft for a person to reviewConfigured per deployment
From:CRMTo:iSupport AIWhat passes:Customer history informs the draftPlanned
From:iConnect InboxTo:iQuality (Where included)What passes:Sampled conversation for reviewDemonstrated in testing
From:iSupport AITo:iQuality (Where included)What passes:AI drafts included in quality reviewPlanned
From:iQuality (Where included)To:iSight (Where included)What passes:Protected team summaryDemonstrated in testing

Who acts at each hand-off

The people involved, in the order the work passes between them.

  1. Customer

    Writes in through an activated messaging channel, or through website chat once it is released

    Result: A conversation in the shared inbox with an owner and a status

    Next owner: Service agent

  2. Service agent

    Opens the conversation beside the customer's contact and replies: on WhatsApp they can review a suggested draft or write their own reply; website chat, demonstrated in testing and not yet released, is designed to be answered by the agent directly

    Result: A reply a person approved before it was sent

    Next owner: Customer

  3. Knowledge owner

    Adds and approves the answer content that drafts may draw on

    Result: Approved knowledge the team can rely on

    Next owner: Service agent

  4. Team lead or quality reviewer

    Samples conversations for review and agrees a coaching action, where quality review is included

    Result: A coaching action based on real conversations

    Next owner: Service agent

What each product contributes

Each product's part in this workflow, with its current status.

  • iConnect InboxControlled rollout

    Brings conversations into one shared place, each with an owner and a status, and links each one to the customer's contact.

  • iSupport AIControlled rollout

    Suggests a reply draft from your approved knowledge for the agent to review, edit or discard. A person approves every reply.

  • CRMControlled rollout

    Holds the customer's record so the agent can see who they are talking to. Drafts do not use this history yet: that connection is planned.

  • iQualityRoadmapWhere included

    Where included, is designed to let a reviewer sample conversations, record findings and agree coaching actions.

  • iSightRoadmapWhere included

    Where quality review is included, is designed to show managers a protected team summary rather than individual detail.

What your business gains

  • Less searching for routine answers: the draft starts from approved content
  • Clear responsibility: every open conversation has an owner
  • Continuity at handover: the next agent sees the conversation and the contact
  • Coaching based on real conversations, where quality review is included

Less searching for routine answers, with clear responsibility and continuity at handover.

A useful measure for a pilot

Share of conversations with a named owner, and the time from the first message to the first reply a person approved.

What this requires

This workflow needs the shared inbox, the customer record and these items set up for your company:

  • An activated channel: WhatsApp activated with your own WhatsApp Business account, or website chat once it is released. The WhatsApp status is Integration built; activation and verification pending. Website chat is demonstrated in testing and not yet released. Drafts are prepared for WhatsApp conversations only; website chat is designed to be answered by a person.
  • Approved knowledge: answer content your company adds and marks as approved. Only approved knowledge is used for drafts.
  • Your own AI provider key for drafts. Usage on that key is on your account with the provider.
  • Agents and team leads set up with roles, and a rule for who owns new conversations.

Products this workflow runs on

Can be added

Evidence and scope

Website chat reaching the shared inbox with a human reply, sampling conversations for quality review and the protected team summary are demonstrated in testing on code not yet released to customers. WhatsApp, approved knowledge and drafts depend on setup for your deployment. Drafts informed by customer history, and AI drafts in quality review, are planned.

  • Demonstrated in testingthe connection passes automated tests in test environments, on code not yet released to customers.
  • Configured per deploymentit depends on setup for your company, such as an activated channel or approved knowledge.
  • Plannedit is not built yet. It is shown so you can see the intended workflow.

Each product shows its own status, and none is described as released for general use. Your demo confirms what applies to your deployment.

A suitable starting option

Service team

Compare what this starting option includes, what it requires and what can be added. Packaging is agreed per deployment.

See this starting option: Service team

Questions about this workflow

Does the draft use the customer's history from the customer record?

Not yet. Drafts draw only on your approved knowledge. Using customer history to inform a draft is a planned connection, shown as Planned on the map.

Are drafts prepared for website chat?

No. Drafts are prepared for WhatsApp conversations only. Website chat is demonstrated in testing and not yet released; it is designed to be answered by a person.

Can an agent reply without the draft?

Yes. The agent can ignore the draft and write their own reply. Nothing is sent until a person approves it.

Are AI drafts part of quality review?

Not yet. Quality review samples conversations handled by people. Including AI drafts in quality review is planned.

See a conversation handled end to end

Bring a few typical customer questions. We show how they would arrive, who would own them and how a reviewed answer is prepared in your setup.

Request a tailored demo of this workflow