Call Deflector
Deflect phone or web chat conversations to Whatsapp for easy re-engagement
Introduction
The Call Deflector app enables your bot to answer incoming chat messages on your social channels. It can be used to configure any type of social messaging assistant. By default it is configured to forward messages to an employee (escalate) while sending an automatic reply, but it can also be configured for automatic answering on recognised questions.
On other channels than Whatsapp it will deflect to Whatsapp when it takes too long for an employee to join and respond. This way your audience is forced to continue in Whatsapp which makes them re-engageable, even when they are offline.
Getting started
Channel requirements
The app requires a chat channel to handle incoming messages. It can also handle the phone channel to redirect callers to a chat based channel.
Installation process
Define opening hours
The app uses your opening hours to determine whether or not you are online. Via 'Settings - Opening hours' you can modify the availability of your bot.
Enable a chat channel
The Call Deflector requires at least one social chat channel, such as Whatsapp.
Using the Call Deflector
App settings
The only settings required for the app are the opening hours. Based on the opening hours the Smart: Escalate node can redirect users to Whatsapp for easier re-engagement.
Roles
Agent
Operator
Supervisor
Producer
App workflow
The app is configured up with 4 flow files
Playbook
Inbox
Phone2wa
Web2wa
Playbook
The playbook allows starting the conversation on 3 separate channels in order to elicit the user intention. You can see this by inspecting the little badges at the main entry point, this will be telephony, chat and whatsapp channel related situations. The user, depending on calling or chatting will then try to answer the questions created (in this case no questions are trained) therefore the unknown flow will trigger the auto reply flow.
Inbox
The Inbox will open the conversation on Web using a say node and a prompt in order to elicit the user intention. You can see this by inspecting the little badges at the main entry point. The prompt will then try to answer the question (in this case no questions are trained) therefore the unknown flow will trigger the Escalate: smart flow.
However, on Whatsapp the conversation is never started by the bot (unless a template message is sent) but only by a Contact. Therefore, the Escalate: smart kick in directly when a message is sent that isn't automatically replied by a dialog trigger.
Phone2wa
The Phone to Whatsapp (Phone2wa) flow deflects incoming phone calls to Whatsapp. The app asks the user to which phone number it can send a whatsapp message:
Once the connected has been closed, the caller will receive a Whatsapp message based on the number the call has given up to be contacted on. See the Ask: phone for reference. After the caller reacts to the incoming Whatsapp message, the conversation moves from speech to chat.
Web2wa
The Web to Whatsapp (Web2wa) flow, pushes the chat conversation to Whatsapp. When it takes too long for an operator to join the chat, the user will be shown a QR code. If the user scans the QR Code, their phone will open Whatsapp, allowing them to continue the conversation there.
The deflection to Whatsapp is useful for two reasons:
It allows you to re-engage more easily. People don't tend to keep their browser tab with the chat widget open all the time. Whatsapp offers a familiar experience, and users know they can close a conversation.
On web, users expect to be answered directly, whereas on Whatsapp they're used to a more slow chat.
FAQ
Languages
The app will determine the language of the user based on their phone number. The app falls back on the default language of the app in case the language cannot be determined (e.g. French or Dutch in Belgium).
It's possible to add an "Ask: Language" node to allow users to switch to their preferred language.
Opening hours
The app will let the user know if they're reaching out within or outside office hours. This helps set the expectation on how long to wait for a reply.
Resources
Last updated
Was this helpful?