Email Bots: How to Automate Your Mailbox
Email bots are Visor.ai‘s new product solution. It’s another option of automation interactions between companies and customers. However, it’s present in the mailbox of your Customer Service.
It came to the public at the beginning of 2020 and, like chatbots, uses Artificial Intelligence.
Tranquilidade is one of Visor.ai’s largest customers. When they heard that we were expanding their solutions, they immediately wanted to be part of this new stage.
The insurer knew that the solution was at an embryonic stage and that it would need improvements.
Since the chatbot worked so well, on both the website and WhatsApp, they weren’t afraid and were excited to invest in Visor.ai solution once more.
That’s how the pilot project came about in partnership with the Portuguese insurance company.
Chat and Email bots: Differences between Channels
The two great forms of written contact with Tranquilidade are through chat and emails.
The insurance company receives about 6000 contacts a month, through chatbot, and has an efficiency of automatic responses that is around 70%.
In addition to these requests, it receives thousands of emails every month in its mailbox from the most varied subjects.
However, each channel has its specific type of information.
In chat, the questions are more direct. They are more general information requests. In other words, customers or potential customers ask questions about insurance, such as how to join, the coverage of each plan, etc.
In the email, the questions involve sending more specific documents and queries. Usually, contacts are made by people or companies already insured; clients who wish to change personal data or the insurance policy (which requires proof), or report occurrences (car accidents, for example).
The Technologies behind Email Bots
Both chatbots and email bots are directly related to the written language.
For the bots to respond, it’s mandatory for technologies to process this kind of language.
It’s necessary to employ areas of Artificial Intelligence, such as Natural Language Processing (NLP) and Machine Learning.
The NLP is concerned with the interactions between humans and machines, namely, problems related to the understanding of Human Language. In this case, the written inputs.
On the other hand, Machine Learning is in charge of “teaching” the computer how to make accurate predictions from previously acquired data.
The key to Machine Learning is the size of the database. The more examples there are, the easier it is for the bot to recognize certain types of elements and then respond.
These two subjects are essential for discrimination in writing. However, in emails, document attachments are often received.
This technique allows the system to recognize images and videos and turn them into code that the computer understands.
With the recognition of characters, it’s possible to train the bot to perceive a pattern and, from that, recognize a type of documents. This training guarantees a high level of image identification.
Imagine that you teach your bot what an Identity Card (ID) is. You show it what the characteristic elements are and fill the database with numerous examples.
When your system receives an ID, the computer crosses it with the information you gave it and then recognizes that specific type of document. It can even extract identifying elements, such as the ID number or the person’s name.
It’s with these three technologies that your email bot identifies the requests from your clients and the attachments that they sent.
Email bots and Other Systems
Salesforce is a CRM (Customer Relationship Management) solution that allows all departments of a company to manage their customer relations.
In this case, Tranquilidade uses Salesforce to handle incoming emails, so it was essential to have access to the whole process.
Visor.ai, by offering a very adaptable and scalable solution, has easily integrated its bot into Salesforce.
The integration of the email bots was carried out directly in the insurance company’s mailbox and Salesforce. Thus, it’s possible to feed the bot with real and up-to-date information, always taking into account all the security procedures and data confidentiality of the insurer and its clients.
In this way, it was possible to reach the focus of this project: to automate the processes of email sorting between the company and its customers. Namely, to which category the email belongs, if the attached documents are those necessary for the execution of the request, among other actions.
Email bot Expected Results
Tranquilidade has now part of the email handling automated.
In this first stage of the launch, it starts with the processes of opening car accident claims, change of IBAN, and request for the issuance of letters to the Statutory Auditors.
However, there are already prospects of extension to other processes and email channels.
This new solution is also an excellent help for customer service assistants in sorting requests.
Instead of receiving the emails in a random and uncategorized way, they have immediate access to filtered cases by category. This categorization allows not only automatic processing but also internal routing of claims to the departments responsible for each type of request.
The automation of interactions brings impressive results to companies that have large volumes of standard emails. That is repetitive emails where the response or resolution process is always the same.
This solution doesn’t intend to reduce the number of contacts to the insurance company. But rather to reduce the process until the queries reach the indicated department.
Tranquilidade still has thousands of contacts in its electronic address. But now has a more automated and less time-consuming treatment for its employees.
This way, it increases the productivity of your team and the resolution time for each case. It also provides a better service and increases the satisfaction of your customers.
At this moment, the email bot of Visor.ai performs the following procedures:
- Reading the subject and text in the body of the email
- Reading of attached documents through the Optical Character Recognition System – OCR
- Automatic responses with requests for missing documents needed for a given action
- Categorization of emails according to subject and forwarding to the correct department.
The automation of emails meets Visor.ai’s goal: to offer an increasingly robust, multi-channel solution.
Our mission is to put an end to repetitive interactions and tasks that can be programmed. Thus, your customer service team has more time for more complex problems.
If you are interested in this new automation option, don’t hesitate and contact us!
Design: Ana Rolo Text: Filipa Perdigão