Catégorie
|
#donnéespersonnalisée
|
Objective
Develop a connector between your CRM tool and iAdvize to enable the escalation of leads in your CRM.
What would I need?
Required:
- The iAdvize GraphQL GraphQL API
- Creating an iAdvize application
- 1 iAdvize Webhook :
Possibly needed:
Use Case
Here is an example of an ideal app to integrate iAdvize into a CRM.
I would like to be able to integrate iAdvize into my CRM and thus
- escalate leads/contacts in my CRM
- gather the context of visitors and their conversations in my CRM
Process
1 - Prerequisites
- Prerequisite 1 : Recovery of your API key
All the resources accessible from the GraphQL API are visible in our documentation.
→ GraphQL:
The authentication uses temporary and revocable access keys.
Please note that the lifetime of the key is 24 hours
You can generate an access key by calling the url mentioned in this link with a user email and a password.
A link must be made between the conversation and the information available in your CRM. This can be done via the email. The visitor's email should then be automatically requested at the beginning of the conversation by a bot.
- Prerequisite 3 : App creation
To create this interconnection between iAdvize and your CRM, you will first need to create an iAdvize App.

Go to the next step to understand how to use them!
2 - I want to trace/create leads and gather context on my visitors and their conversations in my CRM
The webhook system allows external applications to subscribe to events (via callback URLs) to receive real-time updates. When you create your application, you can subscribe to a list of events.
In this example, I use the webhook to:
- create leads/contacts in CRM
- gather context concerning visitors and their conversations in my CRM
For more information on our other available webhooks:
3 - I want to extract key information from my conversation and save it in my CRM
Preliminary step
To facilitate the analysis of the content of the conversation and the search for key information to be brought up in your CRM, you must identify precisely the questions that will be asked to the visitor. This way, you will know that the visitor's messages following these questions correspond to the elements to be brought up in your CRM.
iAdvize allows you to use Markdown syntax in messages sent to visitors. You can therefore use the following tip to facilitate the identification of your questions in the content of the conversation:
For each of your questions (which is waiting to be answered in your CRM), you can prefix it with the following markdown code:: [](Qn) (where n is a unique question number in the conversation).
ℹ️ This markdown normally corresponds to a link where the text is supposed to be indicated between the square brackets. As there is no text in the square brackets, the link will be invisible and will not appear on the visitor's side.
Here is an example of the conversation content you could retrieve:
Agent side view:

By typing Q1, the agent sees his pre-recorded question appear - he can then select it

The agent can also use the library of pre-recorded questions/answers by clicking on the "list" icon
Visitor side view:

The visitor does not see the markdown - he only sees the agent's pre-recorded questions
After analyzing the conversation, you can easily locate the Q1 and Q2 questions and you know that the answers are immediately after each of these questions. The answers thus located can then be saved in your CRM.
⚠️ The format of the responses is a "conversation" format, not a "form" format. This may result in long visitor responses and/or multiple pieces of information in a single response.
Note that you can use this trick either :
- by a bot (for example, pre-qualification)
- by determining them beforehand and creating them in iAdvize's pre-recorded answers so that the agents can use them during the conversation.
-
- Example: For the question [](Q1) What is your interest model? - The operator can start typing in the Q1 dialing field and iAdvize will automatically propose the corresponding pre-recorded answer.
It is then possible to simplify the life of advisors by adding the sequence of questions in the pre-recorded answers in iAdvize by integrating the markdown to list the questions "Q1" "Q2" "Q3" ...
⚠️ Note that the email retrieval must be automatically done by a pre-qualification bot - it cannot be done by an agent, otherwise the email will not be included in the visitor resource fields. However, it is also possible for the agent to update the visitor's email during the conversation by entering it in the "Profile" part of the conversation. The email will then go back into the email field of the visitor resource.
Extract visitor responses from the conversation
1) Recover the conversation id
Recommended solution
- Once the conversation is over, you can retrieve the id of the conversation via the end of conversation webhook v2.conversation.closed
Alternative solution
- If you want to retrieve the content of the conversation in real time (i.e. while the conversation is taking place), you can subscribe to the conversation start webhook v2.conversation.pushed to retrieve the conversation ID. You'll have to make regular requests because we don't have an event indicating the reception of a new message. (polling)
2) Retrieve conversation content from this conversation ID
Thanks to the conversation id, you can retrieve the content of conversations via the conversation* query
*the content of the conversation is in the messages field

If you want to retrieve conversations from a particular agent (or bot), you can subscribe to the conversation.closed webhook v2.conversation.closed and filter on the ID of the operators involved.
4 - Installing your App within the iAdvize Admin
Remember: this App can be used in one or more iAdvize projects (SID).
- Go to the iAdvize admin and search for your application in the Marketplace and install it. Follow the displayed steps and complete the installation.
- Once your App is installed, you should have a screen similar to this one telling you that your App is installed:
