Ava — Making conversations accessible for the deaf

Thanks @adrienm for hunting us - this is one of these special moments we've been really looking forward to! Hey guys! Ava is an unusual product, and we're humbled to be here. We just released the 1.0 on the Appstore and Playstore, and we'd love to get feedback & thoughts from all the community - it's our first time here (hi 👋!). I grew up the only hearing person in a Deaf family. When you get to be in the middle of these 2 worlds, you see a ton of issues. We started Ava to solve what we believe to be the most essential problem today for 400M deaf & hard-of-hearing people: understanding and participating to group conversations. Lunch, business meetings, hanging out with friends can become a nightmare to follow when you only rely on lip-reading. Only other solutions involve a professional interpreter/captioner and cost $100/hour. Talk about an expensive coffee meeting 👀. Ava aims to make the crazy idea of a 24/7 accessibility to conversations, possible. Ava connects the smartphones of a group of people who downloaded the app, and turns them into a smart microphone system. With some speech recognition & speaker identification magic✨, Ava can then show who says what in less than a second to the deaf person. We use it everyday for meetings, standups, lunches. We have to: our cofounder & CTO is himself deaf. But because it just needs smartphones, Ava is flexible, simple, and works in many other situations. The most interesting thing was to realize the range of situations our early users decided to use Ava for and that we hadn't predicted. Stargazing with friends because reading lips in the dark is impossible, communicating from the other ends of a house, understanding their doctor, captioning a parents/students meeting... And sometimes they even stretched to its extreme 😁 :) While Ava isn't always getting what people say right, it's a definite leap from no accessibility at all. And it's just the beginning: Ava gets smarter as time goes, learning to distinguish your voice better everyday 🤓. We're excited to share the 1.0 to the world today, and to announce a $1.8M seed round, led by visionary investors who support us in this vision of a truly 24/7 accessible world. Our wish? To make Thanksgiving dinners actually great again for 15M deaf & hard-of-hearing Americans this year. Learn more about it here: https://medium.com/@AvaScribe/in... If you know a deaf/hard-of-hearing person, pass it around please: this little app might actually change the life of someone! We'd love to hear what you think of Ava - we'll be here all day! Cheers, Thibault & the Ava Team

@t_duchemin @adrienm @avascribe It's really refreshing to see someone in the tech world make a product to solve such a huge social problem. Great work!

@t_duchemin This is so very awesome! I had a dream of building something like this 15 years ago when the tech didn't exist to make it happen! So great to see you make this a reality! Congrats!

@mattbrian Thanks for your questions. The app requires a certain version of the Android operating system on your phone. It seems that your Moto G4 Plus is not updated enough. You can try doing a software update, otherwise I'm afraid it won't work for that device. Right now it's not available in Brazil yet. Feel free to leave your contact information behind on our website (www.ava.me), so that we can keep you posted on any updates. Thanks again!

@pieter_doevendans @mattbrian please send us an email at team@ava.me, we'll ping you when we have dates for Brazil. Obrigado!

thank you @cyrilpaglino for the #CODA ❤️!

@ravsydney It's been an amazing journey for us. We're learning a lot every day! While building Ava, we've gotten to understand the challenges, but also the beauties of deaf and hard of hearing cultures. In the end of the day, we're all the same human beings, and our 'different abilities' should not impact our social and professional opportunities. It is easy to take our every day for granted, and forget that we all have to make an effort together for equal access. Thanks a lot for recognizing this, and for your kind words!

@hameto_ we love BeMyEyes too! that's exactly it for us as well: how using social dynamics in place to solve a problem together (our logo "&" reflects this thoughts of shared action). Great question re: this important use case of a "networking" moment. One of the answers is that you'd diminish the accuracy by behaviorally putting the microphones too far from the speakers. So it's not that much the identification who is really the issue here. The trade-off is: keeping accuracy at a high level or asking to the rest of the group to download an app. For now, Ava isn't optimal in these situations, but we have a few ideas we'll pursue on (& maybe come back here for!)

@lonefur yes, we're starting for familiar conversations, with relatives/close people. - The use case you describe (to random strangers) is a big one. Of course, we're not into asking the other to install the app then (unless it's an important/1 hour convo), but the way in the video Mando uses it for the doctor should work OK in quiet settings. - iPhone 5S isn't great, also processing is very slow - Crashes: on us. Sorry about this. 1.1 will be fixes :p

@wilhempujar Can't help but love your bluntness! We're trying our best to not get lost in the maze.

@adamscochran very much aware of it! let's say the mobile version is the first brick of the system . would love to discuss more about it - we have a few ideas around this.

@clemenslm Thanks for the question! We're already available in a few English-speaking countries (AUS, UK, CA, South Africa). The availability is limited by the different languages the speech recognition system supports. The more users in a country, the better. Our wish? To have something ready for Christmas time in some countries where we're working with communities already, depending on bandwidth & how today goes ✨

@cecimetropolis great question! first we had to learn how to communicate super effectively with about 40% of our early users, who were Deaf: we learnt ASL - it was easier for me than for Pieter, who self-taught in 3 months from scratch :) we tend to have a few key principles we believe in, pick a few very important features we want to build, and try to really focus on nailing the experience. key principles: 1. We go for distributed: because our system is made of everyone's devices, it forces us to want to design the most seamless way of connecting everyone with this constraint. For example, we have no account creation if you're hearing and join the first conversation, or that it takes a tap to connect 2 people on Ava already 2. Designing for very non-tech savvy people: there's all ages, deafness being sort of randomly distributed. We always have them in mind 3. Focus on the core functionality Regarding the testing, we made a Slack community with deaf/hard-of-hearing users (so tech-savvy feedback) and kept testing also face to face with normal users, look at their behavior, etc. We don't have a lot of features right now, but they should work well. For example, saving transcripts isn't something we do. We could, but what matters right now is to nail the actual following of the conversation :) Hope this helps!

@jon_lay it's been a true honor to build Ava and get your feedback all the way. You & @wearehanno are also super inspiring us to build a great culture every day 😉

@gknook Hey Guido, indeed, we're not available in Germany, but only English-speaking countries for now. Happy about it because our servers are being hit like crazy 😅 We'll give a timeline a bit later to open for the other countries.

@jdbt let us know how Thanksgiving goes this year for him! 👍

@streatstyle for sure, we use the principles of machine learning. The key is to make the way to improve our technology scalable & multi-linguistic. We'll unveil couple things about it in the future!