Rasa just turned two and we’re excited that our awesome community has hit a few important milestones:
- Every minute, somebody downloads our open source tools to build contextual AI assistants.
- Downloads have tripled since our last announcement in April.
- We’ve seen great contextual AI assistants across industries and our community, including a DIY version of Google Duplex.
- Our community members doubled and we see local user groups popping up in many different countries.
- More than 200 contributors have made our tools better and more accessible by writing code and content.
Weren’t chatbots meant to be dead?
The wave of chatbots that was predicted two years ago didn’t happen and the reaction has been headlines that read: “chatbots are dead”
At Rasa, we see the opposite. Our open source community is growing faster than ever. How is that possible? We see two underlying trends that continue to drive the growth of Rasa, especially in the enterprise:
- Enterprises (and their customers) want AI assistants, not FAQ chatbots, but they’re difficult to build: Since Google demoed Duplex in May, every developer, product manager, and executive wants their own. Duplex showed what an AI assistant can do if it can handle contextual conversations and go beyond a “level 2” FAQ chatbot. These capabilities are extremely hard to build if you’re not Google, and that is where Rasa comes in.
- Fully controlling and owning your IP and data is crucial, and can’t be done using cloud APIs: The early days of chatbots were filled with prototypes, proofs of concept, and innovation projects. Intellectual property didn’t matter. But when Fortune 500 companies go into production, IP and data privacy (e.g. GDPR, HIPAA) become hard requirements, and those are difficult to meet using the cloud tools provided by the big tech companies. Instead, companies run Rasa on-premise or in a private cloud.
At Rasa we believe that to get the most from conversational AI, you need a product team that combines the efforts of developers, conversational designers, product managers, copywriters, and data scientists.
Our mission is to provide product teams the best open source tools to build and run great “level 3” contextual AI assistants in production, without handing over all their IP and data to the big tech companies.
Every minute, somebody downloads our open source tools to build contextual AI assistants
Our downloads have tripled to over 300k since our last community update eight months ago. This growth shows developers’ and companies’ immense appetite for building contextual AI assistants. We see exciting use cases across all industries from healthcare to banking, from startups to the Fortune 500. Over the coming months, we’ll publish more of these customer stories.
Today, we wanted to highlight a few great examples from the community:
- A DIY Google Duplex: Our hero contributor Josh built a “Google Duplex”-style assistant using Rasa and Twilio. It helps salons to take appointment calls automatically and shows how easy it is to connect Rasa with other developer tools like Twilio.
- AI Assistant for Primary Care: Circle Medical is building a new kind of primary care practice that uses AI to improve the patient experience and deliver care more efficiently. They build an AI assistant that facilitates communication between patients and physicians. It helps patients refill their prescriptions and get information about their medical results automatically.
- Onboarding Bot for Docs: A small product team here at Rasa started working on a Sara - an AI assistant that helps new Rasa developers get started through a conversation - the code is open source, so it also provides a more advanced example bot for Rasa users to learn from.
你好, Olá, Hallo, नमस्ते, Hello - Our growing global community is becoming local
Rasa users are not just consumers of an API but actively take part in our community. We’re excited to have a new home - our Community Forum. Since its launch four months ago, almost 2000 developers signed up and are supporting each other in building AI assistants.
In October, our community met in Berlin for the first Rasa Summit - we had over 300 attendees joining in person or via livestream - big thanks to community member Edouard from neobank N26 for presenting their work.
Our community has been global from day 1 and we love to see Rasa being used in the majority of countries worldwide. Since our last update, we’ve seen this global community forming local user groups in many different areas, for example:
- India 🇮🇳: We have local meetup groups in Mumbai and Bangalore that are entirely community-run and organized by our hero contributor Soumya. There is even a hackathon dedicated to Rasa happening and organized by our hero contributor Bhavani.
- China 🇨🇳: Our Chinese community is organized on QQ, a messaging app, by our hero contributor Xiaoquan, and is working on local meetups. There have been amazing efforts to translate our docs into Mandarin, too.
- Brazil 🇧🇷: We also have a booming local community in Brazil which has its own Facebook group run by our hero contributor Matheus and is doing a great job in translating Rasa resources in Portuguese.
- US 🇺🇸: There are Rasa user groups in New York City and San Francisco.
- Germany 🇩🇪: We had our first Rasa User Group Berlin in December in hometown and will continue to run them on a regular basis.
We’re really excited to see those local meetups forming - if you’re interested to start one in your city, let us know by posting on our forum.
Our contributors continue to push the boundaries of AI assistants every day
Contributors are the heart of our community and help to make our tools better every day. We’re stoked to have seen so many high-quality code and content contributions - we doubled the number of contributors to over 200 since our last update in April. We cannot mention every one of course, here but wanted to highlight a few to give you some idea of the amazing people involved:
- Souvik Ghosh - our hero contributor who wrote some really great tutorials on developing with Rasa, made multiple code contributions and is doing an amazing job in helping others on the forum.
- Great team effort by Arthur, Izabela, Matheus and another Matheus in creating and contributing a RocketChat connector for Rasa.
- Our contributor Tom Wollnik who made one of the most recent contributions to Rasa for making it easier to configure training policies using the yml files.
We’re very excited to have such a strong and active community of makers all over the world. In our next year, we’ll be working to double-down on our community, open source tools, and machine learning research. AI assistants are in their earliest days so let’s get back to work - there is still a lot of stuff that needs to be developed to make AI assistants that work for everyone. Be part of it and join our community of makers! Want to be even more involved? We're hiring!