Our Mission: Make great assistants easy to build
At Rasa, we envision a world where software is accessible to every end user, and every developer can build great AI assistants.
Conversational interfaces have the potential to allow everyone to use any piece of software by expressing what they want in their own words. This is the opposite of how most applications work today, where users have to translate what they want into a valid sequence of button clicks or commands.
A level 5 assistant could in theory allow you to express in your own words that you want to change your cell phone plan without understanding how to use a clunky web portal. You could book a doctor’s appointment without knowing which specialist is the best fit, or edit an image in Photoshop without hours of prior training. This would radically reduce the friction of interacting with the world around us and make it more accessible.
Bringing the potential of conversational interfaces to everyone
Conversational interfaces aren’t a new idea. So why aren’t conversational interfaces ubiquitous, highly developed, and part of our daily lives? The technology doesn’t deliver on that promise yet. It’s still very hard to build good assistants, and nearly impossible to build great ones.
Whoever wins the race to build this technology can set the standards and defaults for years to come. So, what kind of future will we have? Will the core technology be hidden behind an API, or will it be open to modify and scrutinize? Will the best assistants only run on one company’s platform, available only in the most commercially interesting languages? Will assistants be built with ethical considerations in mind?
Depending on how this plays out, many end users and developers* could get left behind. We envision a world with AI assistants that work for everyone. Concretely, this means that (1) every end user can use any piece of software, and (2) every developer can build great AI assistants independently.
For that vision to become reality, more folks have to be able to build better AI assistants. It’s simply not possible for one company to build all use cases, especially use cases that don't attract as much commercial interest or attention. That’s why our mission is to empower developers to build the use cases that matter to them and make great assistants easy to build.
*developer = broader definition of people who make software
Making great assistants easy to build
To get there, we have to balance our efforts across two themes: pushing what the most advanced developers can build, and reducing the effort required to build great AI assistants. Focusing exclusively on one of these is a dead end. If we only focus on pushing what is theoretically possible, we don't build for all developers; if only on reducing effort, we impede those who can help us push the boundaries.
Finding the right balance will require us to stay patient and say “no” to a lot of exciting projects, features, potential customers, and opportunities. Committing to these five key principles will help make these hard decisions:
Provide open source code & free education
We want to help everyone working with conversational AI improve their understanding and skills with well-maintained, documented open source software and free education. Controlling their own data, acquiring knowledge, and creating intellectual property gives developers independence.
Nurture an active & diverse community
We don’t have all the answers and it takes a combined effort to push the boundaries of conversational AI. Our inclusive, respectful, and open community allows every developer and partner to learn, collaborate, and contribute. Observing the most advanced teams in the community is a way for us to see the future earlier.
Ship applied research to solve real problems
AI as a field is moving fast, with impressive results coming from universities and big tech companies. Our research team is driven by real challenges and datasets from our developer community and enterprise customers, allowing us to bridge the gap between paper and production.
Go high-touch with enterprise customers
Most enterprises are new to conversational AI, which is our opportunity to educate them, and vice versa. Working closely together allows us to deeply understand challenges and use cases in the real world, find solutions, and scale those back to the wider community.
Empower diverse, inclusive, multidisciplinary teams
Language data isn't just a sequence of bits. It directly reflects the society and culture of the people using it. We believe that diverse, inclusive, multidisciplinary teams are needed to build the best assistants. To serve those teams, Rasa has to be diverse, inclusive, and multidisciplinary as well.
It is our responsibility to ensure that everyone who wants to build assistants can do so independently, and to prevent others with closed ecosystems, walled gardens, and restrictive APIs from winning the technology race. Why do we think we can pull this off? Because we’re able to dedicate all our resources and energy to these five principles like no one else can, creating a virtuous cycle between them.
Focusing on our mission will create enormous value for our community, customers, and the world. Capturing just a fraction of the value we create will make us a very valuable company, too. We’re just getting started. Let’s do this! 🚀