Our system allows for the automation of managing messages from multiple sources, such as WhatsApp, email inboxes, API integrations, and other channels.
Each received message represents an inquiry or problem from a customer. The system automatically assigns the message to a specialized AI agent, which generates a response and executes a defined action.
1
You will need to have access to the WhatsApp Business API:
Verification of the Facebook Business Manager Account: Make sure that your Facebook Business Manager account is verified. This involves providing information and documentation about your business for validation.
Verify a dedicated phone number for the API and set up your business profile on WhatsApp. (This phone can only connect to GuruSup as an official Meta Business Solutions Provider (BSP))
Go to gurusup.com
2
3
You need to be registered. If you are already registered, log in.
If you are not registered, sign up with your email and a password or with Google.
4
Click on the settings section.
Click on "Integrations"
Click on "Login with Facebook" to grant permissions. To integrate WhatsApp, it is necessary to grant permissions in Meta Business.
In this process, follow the guided steps of "Login with Facebook"
This process will be subject to the country from which you operate and to the previous configuration you already have in Meta.
Remember to perform this validation with the company's account (Meta Business Panel) that you want to associate with WhatsApp.
5
To start creating the agents, we will send a test message to the inbox.
If you have already completed the integration with WhatsApp, you will be able to send a message.
If you have not completed it yet or prefer to integrate the flow with an email, send an email to the GuruSup mailbox associated with your account.
(This is a unique mailbox to which you can set up a forwarding from your Gmail account, Outlook, or another email provider to automate email messages.
6
In the test message, we will simulate a real case, a typical problem of our clients.
7
When the message enters we can see relevant information such as the subject, the first lines of the body, the source of the message (in this case, redirecting to the GuruSup inbox), the language in which it has entered "Spanish", date and time, and user data.
As you can see, my configuration is in English, so the message directly comes in with the translation. You can manage your translation language in "Settings".
8
In this panel we can see two spaces: on the left the conversation thread and on the right the panel from where we will manage the responses, fine-tune the agents, and be able to leave the automated process in "Auto-answer" mode.
At this point, we do not have any agents created, so "Triage" is assigned, the agent responsible for detecting which agent will resolve the client's issue.
9
When we write the message that we would manually provide to this message, we are giving a set of instructions to start creating the agent.
We enter the response in the text box
Click on "Create"
10
11
We reviewed the agent's generation and modified it to perfect its behavior.
12
The agent can generate a draft for validation and submission. In this process, when a draft is generated, the user will be able to modify the text, and it will make the agent "learn" to respond.
When we consider that it responds the way we would, we can change the behavior so that it answers without any type of review.
Upon finishing the editing, click "Save" to keep the changes.
And you can repeat this process as many times as you want to create agents or create and edit them directly from the "Agents" tab.
14
With this foundation, you can start automating responses.
To go further with GuruSup, you can also connect databases with APIs, endpoints, or no-code platforms like Make or Zapier that will allow you to integrate with multiple tools.
We left a video here with a more detailed step-by-step guide and some automations using Make.