Weni comes up with Push - a multilingual interactive chatbot
Jul 12 , 2017
On this page
Weni has developed Push – a web platform that facilitates smart campaigns by designing communication flows and automated responses easily and intuitively. Building on their experience in developing mobile apps, Weni's platform will use natural language processing and support multiple languages with an open database and collaborative for organizations and companies to create chatbots.
We hope to encourage other developers and companies from around the world to collaborate, contribute and create solutions based on our platform.
How would you describe your solution?
Weni is developing Push – an intelligent, interactive robot that can provide people with relevant, tailored information based on questions presented by people around the world. Push’s web platform will enable people to talk to the robot, ask questions and seek information, interactively.
By developing Push, we aim to address two main challenges:
The market for chatbot applications is growing rapidly, however, it is difficult to create contextual models of artificial intelligence from scratch without advanced technical knowledge.
Existing systems are not able to work well across local dialects and less commonly spoken languages.
Geared to address these challenges, our solution harnesses the power of open-source and crowdsourcing technology, to create a rich and multilingual repository of pre-trained knowledge bases. Our solution is also designed with a ready model for others to use with ease – providing the ability to translate to other languages and exporting to other tools like wit.ai and api.ai.
What is unique about your solution and what is the competitive landscape like?
There are several tools for creating chatbots and some of them are open-source. However, these tools were mostly created for developers – making them less intuitive for non-developers to use. Push will allow anyone to create their very own smart chatbot and build communication flows and automated responses – easily. Our solution will also develop a diverse, rich set of pre-trained knowledge bases in various language and themes, including a functionality to add and update languages and knowledge bases.
How did you come up with your solution and what inspired you to form your company?
Weni specializes in application development for Android and iOS. In 2016, we started developing chatbots and applications using natural language. While testing numerous tools in the market, we realized that great effort was required to create a chatbot – having to train models on every new project. Moreover, we found limited space for us to share our learnings with the community and to learn from others. Throughout the year, more platforms that supported Portuguese (our native language) were developed – however, none of them could be set up with our own server.
During that time, we also had an opportunity to work with the U-Report team in various countries. This helped us understand the importance of building a dynamic and global product that speaks many languages and serves a wide range of stakeholders.
This experience gave us the inspiration to start developing, testing and validating a solution that serves this need.
Why open-source your solution?
The main artificial intelligence tools available are already open source. We aim to be a player in sharing information on how to develop solutions based on natural language (versus artificial language or computer code). Through making our code open source and available, we hope to encourage other developers and companies from around the world to collaborate, contribute and create solutions based on our platform.
In addition, utilizing crowdsourcing systems allow us to develop an active community of individuals working on improving the creation and translation of pre-trained models.
How did your team come together? What is your team’s MO and drive towards the problem you’re trying to solve?
We are from Maceió – Brazil, one of the most violent cities in the world and with great social inequality.
We always had a vision to work on products with great social impact, keeping the world more connected. Working on a project that changes the lives of people around the world is what brought us together: from the very first application that we created Protect Brazil, a free tool that allows everyone to engage in the protection of children and adolescents to four years later, building Push, a platform will use natural language processing and support multiple languages for organisations and companies to create chatbots.
Use of the Investment
What do you plan on doing with UNICEF’s Venture Fund investment and how will you use that to leverage raising follow-on investment?
UNICEF projects around the world have the opportunity to transform the lives of millions of children. Being part of the UNICEF Venture Fund portfolio provides us with the support to solve real world problems.
We will use the seed funding received from UNICEF to accelerate the development of our solution, tap into the Fund’s networks and generate concrete social impact use cases.