Get to know the Ellevo case

Ellena, the new born bot that became global

Founded in 2006, Ellevo invests in innovation and in the last years, has dedicated resources to process automation and artificial intelligence. Ellevo is an Independent Software Vendor (ISV) that develops and provides a service management solution for medium and large companies through its Service Desk platform, where more than 15 million emails and 3 million calls travel monthly. It is used to manage requests for Customer Service Hotline (CSH), Shared Services Centers (SSC), Industry, Supply Chain and Facilities, Human Resources, Finance and Ombudsman. Since 2017, it saw in artificial intelligence, an opportunity of transformation for its customers and for itself. 

The company's headquarter is in the city of Blumenau (SC) and there are business units in São Paulo (SP) and Florida (USA). Your customers are PwC, Grupo Positivo, Estácio de Sá, Renault, BRF, Natura, Senior, Bunge, Compass/GRSA, Chilli Beans, Energisa, Sotreq and Banco Alfa, among many others. 

The common demand among customers

They had identified a common demand between all these customers: the difficulty of meeting the growing demand of the consumers, and, also, managing, training staff and giving feedback to thousands of services requests every month. A great part of this demand is repetitive and recurrent, such as renewals of contracts, resolution of doubts about product, deadlines and amounts of payments, instructions for access to services, etc.

This demand occurs in all companies and requires each one to have a large group of attendants operating the Service Desk platform and interacting with consumers. However, a great part of this demand is repetitive and recurrent and, although it has a substantially impact on the routine, it does not create revenues and its costs increase with time.

This proved to be a great opportunity for Ellevo, that created a solution based on chatbot with machine learning and cognitive search capabilities, called Ellena. Able to interact with consumers through voice or text commands, both through the company's web portal and through the apps Telegram, Facebook Messenger and mobile app.

Integrations and applications

The chatbot Ellena also integrates with Ellevo's leading business applications, carrying out real-time ERP and CRM data research as well as transaction documents such as contracts, manuals and invoices.

Ellena was designed and built using the Microsoft Bot Framework, so that consumers could seamlessly communicate in their own native language, AND Azure App Services, to easily connect multiple applications without infrastructure interferences. In order to guarantee this experience to the users it was necessary to integrate other services to the framework, such as LUIS Language Understanding, Bing SpellCheck, Azure QnA, Cognitive Search and Speech to Text.


Another key point was the integration with the Facebook and Telegram messaging apps with multilanguage support, which enabled it to serve platform consumers in other languages ​​and on different channels and media, without the need to download specific applications.

For a better user experience, the machine learning-based service LUIS was adopted, which identifies the general meaning of a sentence in natural language and corrects the sentences in which the bot could not find a correspondence or an explicit sense, indexing a feeling for that specific phrase. This has amplified the ability to communicate with consumers without having to worry about specific technical terms or vocabulary nor to repeat different sentences to be attended, since, regardless of the form of the request, there will always be a relationship established in LUIS between the phrase and its respective sense.

Still, all the textual interaction of the chatbot goes through Bing Spellcheck for validation of the user's writing in order to find synonyms and correct the input of information that is made by the user. The input of voice requests, on the other hand, is sent to the Speech to Text cognitive service for real-time transcription and analysis and ensures that the flow is fully met regardless of the channel adopted by the user. Speech to Text is integrated into the chatbot via API and transcribes the text present in the audio messages.

In order to relate frequently asked questions and answers and to keep the knowledge base constantly updated during customer service, learning new answers as the calls are made on a daily basis, Azure QnA Bot Service was used.

Finally, a small detail of great value, the icing on the cake, which made the user experience exciting and assertive was the use of Azure Search's Cognitive Search functionality to enrich with additional information such as images, videos and texts from multiple sources as the Service Desk’s platform, Company’s ERP and even the internet. All the available content is analyzed and classified, extracting key words from it and feelings, among other features of natural language and integrating this content into the conversation in natural language assisted to create the best consumer experience.


In just one case of use, of a call center operation with 400 attendants, chatbot Ellena was included in the service process as the first contact point to receive and handle the demands of consumers before directing them to an attendant, which led to an estimated reduction of 40% of repetitive (by scripts) calls being handled by humans. Furthermore, the fact that consumers meet their own demand led to a nearly 30% reduction in the opening of calls from the Shared Services Center (SSC). The effect was so noticeable that the chatbot became the first contact point for service in all channels available to consumers, from the company's web portal to social networks.

In another case, the company’s Brazilian unit, which is based in the USA, started to test the tool in other Latin American units, being necessary to adjust the service to another language (Portuguese to Spanish). The preparation time for this change would have been quite long due to idiomatic issues, were it not for the simplicity of multilanguage support, the creation of a contextual dictionary of synonyms, and the grouping of scripts by affinity, which allowed the launch to occur within a few days.

The result was so consistent that a new pilot is being conducted at a unit in India, where Ellena is being tested in English. The success of the project in India has a strategic aspect: It will be the Proof of Concept for use in the American market in the second half of 2019. 

Beyond conquering customers in other countries, Ellena was able to reach other continents in less than a year.

Ellena, as well as the entire Service Desk platform, is multitenant, multilanguage and repeatable and can engage to consumers in any and all industries, from small to large, whether B2B or B2C, connecting via APIs to any company's systems. Users can access through voice or text commands, the company's web portal and Telegram, Facebook Messenger. A customizable mobile application option is also available.

Because the chatbot receives end customer’s demands and detect requests, list and present alternatives of solutions, it is ready to be deployed in any segment. It is currently used to meet the internal demands of large corporations, such as human resources, Information Technologies, Telecommunications, financial, commercial and shared service centers, but it can also be applied to supplies and logistics, industrial production and service areas such as Customer Service Hotline (CSH), marketing and education.

The movement

The #DesligueiMeuDataCenter started as a customer migration project from Teevo’s datacenter to Azure, but we understand that this is only the beginning of a great market movement and clearly signals to customers that it does not make sense anymore to invest on local infrastructure and datacenter.

The focus of IT managers should be on looking at business, how to add value to organizations and use the technologies available to ensure the differential in a market that is increasingly competitive. As long as IT areas are paying attention to infrastructure, there is no room to think about how to adopt platforms to effectively perform digital transformation.

We have no doubt that in a few years it will be very rare to find customers who still make efforts to manage on-premises environments. To ensure the survival of these organizations, the focus should be on the understanding of how technology can be used to transform business.

We are sharing a message to the market that the journey to the cloud is inevitable, even for companies offering datacenter services through local infrastructures, and Teevo was Brazil's first datacenter to carry out this movement.

Count on TEEVO to shut down your Data Center.

Get in touch and get information, ask questions and find the solutions you need.