ATU798 – Fluent with Jason Suhari

Hi, this is Jason Suhari and I’m the founder of Fluent, and this is your Assistive Technology Update.

Hello, and welcome to your Assistive Technology Update, a weekly dose of information that keeps you up to date on the latest developments in the field of technology designed to assist individuals with disabilities and special needs. I’m your host, Josh Anderson, with the INDATA Project at Easterseals Crossroads in beautiful Indianapolis, Indiana. Welcome to episode 798 of Assistive Technology Update. It is scheduled to be released on September 11th, 2026.

On today’s show, we’re super excited to welcome Jason Suhari from Fluent on to tell us about this great new tool. We’ve also got a short story about how smart home technology helped individuals in a marriage. And a new way to use ASL to access your Google Pixel phone. As always, listeners, we thank you for listening. Let’s go ahead and get on with the show.

So our first story is a kind of a feel good story from over at the BBC, so from our friends across the pond. It’s titled Technology Has Given Me My Husband Back. It’s written by Vikki Irwin from the Suffolk Political Reporter, and it talks about the monks, not like monks, but Heather and John Monk. It says they’ve been married for 42 years, but they just got some technology which has made Heather’s life a lot easier by helping out John.

Says that John’s 71, has Parkinson’s and a medical condition that affects his memory. So he has a hard time using a regular phone, along with a lot of other things around the house. Said his wife Heather, who’s 63, kind of felt unable to leave the house, even for short little trips. But it says that they got some assistive technology from the Suffolk County Council and its partners.

So the system they implemented is called Cassius. It uses sensors, electronic tablets, pressure mats, and other technology to help individuals live independently. So they’ve helped a lot of folks kind of around the area with these tools. It says now Heather is easily able to go out and run errands or just pop out for a little while. Says thanks to the video care phone and smart speaker, John’s able to easily contact folks. He just taps a picture on the screen and it automatically gets a hold of them. And he also wears a fall pendant that activates if he falls over or presses it if he needs help. So it really gives John and Heather both a whole lot more independence.

So we’ve talked a lot about remote supports on here. We’ve had a lot of different companies and folks come on and talk about it. But really I think this just kind of stresses the importance of assistive technology and how it can really help with independence, not just of the individual, but of their caregivers, their friends, their family, their whole circle of influence can find more independence.

And it actually says that since implementing this, John is actually doing a whole lot more. Kind of playing music, singing along, and just able to feel more independent himself. Not just Heather feeling more independent that she can actually be able to go out and do things. So I’ll put a link to this over in the show notes. And luckily I do get the pleasure of hearing a lot of amazing stories on here about how assistive technology can really help folks out. As soon as I found this one, I thought I would share it with all of you.

We got a story today that comes from us over from Google DeepMind’s blog. It’s titled Putting Sign Language AI into User’s Hands. So I’m not going to lie, I do my best to keep up on assistive technology and accommodations and those things that come out all the time, but I must admit it is an impossible task with just everything coming out all the time, all the great stuff that’s out there. So this was actually brought to my attention by our head of deaf services here at Easterseals Crossroads, Katy. And Katy, just let me know, have you heard about the new Google Pixel 11 and that it offers sign language translation? I had not. So I went and found this story so that I could share it with all of our listeners, but the new Google Pixel 11 phone is actually able to interpret sign language.

So this is something I know at least I’ve been really looking forward to for quite a while. I was wondering when this kind of step in artificial intelligence would be there. So the blog post starts talking about artificial intelligence ability to process spoken language, and just how much better that has gotten, especially over, well, it says the recent decades, but I would say within the last five years it has gotten so much better, not just understanding the human voice and being able to change that, but also being able to understand folks with maybe non-standard speech. So they’ve really come a long way.

But Google DeepMind is introducing a massive multilingual sign language to text or SL2T translation model. Now this SL2T will allow sign to text dictation in Gboard and Live Transcribe on the Pixel 11. It’s going to start with just ASL to English, but it does say that more devices and additional languages will follow. According to the blog post here, it says it’s kind of the same way that a hearing user would use dictation to speak instead of typing, but it allows deaf users to sign to their phone anywhere that they would normally type.

Some examples it gives, says you can sign to search the web, draft messages or documents, ask Gemini to solve queries or execute tasks. And says that in live transcribe, you can sign responses in conversation instead of having to type back and forth. So if you’re someone that uses ASL or listen to some of our shows here with folks talking about ASL and translation, ASL and English are not the same language. So if you have someone who is deaf or hard of hearing that you’re trying to communicate with, ASL is going to be their preferred language. So this gives you such a great tool just because the individual is able to sign, the phone changes that into text, you could easily communicate, or like it said, if you want to search or anything like that, instead of having to sit there and think and type everything out, just sign straight into the camera on the phone and all that information is going to show up right there as text.

What I really love is I’ve talked to folks, and they’ve been like, “Why don’t we have this kind of thing yet?” So the story gets a little bit into some of the challenges with really doing sign language translation. And I guess now is probably the time I should throw something in here that this definitely does not replace interpreters in any way for anything. If you are using Google Translate or Live Transcribe or something like that, it’s still going to be one way kind of communication. It’s still not going to be the full kind of communication, but still it’s a great step in the right direction.

So like I said, it talks about just how this has taken a very long time, and it kind of gets into it a little bit, and it just says the progress in sign language AI has been slow. Partially because building AI for sign language presents complex challenges and because there’s widespread misconceptions about how the languages themselves work.

Getting a little deeper, it says that compared to spoken language transcription, sign language translation presents two core challenges. First, transcribing speech is a matter of performing a sequential mapping from sound to text in the same language, whereas sign languages are independent, natural language with their own grammar and lexicons. So in other words, you’re not doing a one for one. I’m texting my wife, and I say, “I love you.” It may not be the exact same word for word kind of translation. So there’s a lot more that goes on behind the scenes.

Second, it says, “The model must learn to see and understand physical movement.” And for any folks who kind of know sign language, I mean, we always think it’s moving the hands, it’s things like that. But sign language actually conveys meaning through the hands, the arms, the torso, the head, the face, so many different, different things. So you’re not just tracking the movement of fingers or hands, you’re tracking all of these different, different things to really get context and figure out what somebody is trying to convey.

It gets very deep into how the SL2T actually works. And I won’t get into this completely just because, I don’t know, it gives you a lot of different things. But really if you think it is using that camera to interpret markers on the face, the torso, the arms, the hands, the fingers, all these different kind of data points, and then put those all together for the translation. So if you think, that’s a lot, that’s a lot of data and information really coming in order to get everything. It says the model is trained on over 100,000 hours of data across more than 50 sign languages, with roughly a quarter of the data in American Sign Language.

So that is a lot of data to really take in and be able to use. Also, if you think, that has to be challenging because people’s faces look different, people’s hands, arms are different sizes. I know in talking to folks, there is the chance that there’s kind of a regional dialect, just as there would be with spoken words. So a lot of things to interpret.

But as we’ve seen with artificial intelligence, I mean, this might be the first kind of iteration of this, but it sounds like Google really took their time putting it together. I’m sure that even if it’s not great right out of the box, it will get better, especially as more folks start using it, as it learns more and as it just gets a whole lot better. It says they did build a lot of things in here to kind of optimize how it works, to be able to prevent hallucinations, and also to ensure fairness for the 10% of signers who are left-handed because that’s going to change some things up too.

So it sounds like they really took a whole lot of data in there. And as you get down here, and something we’ve talked about on this show a lot, especially with the founders, the creators, and the folks that make things, this was built 100% with the deaf community from the very start of the process all the way through real world letting it out into society. They established an AI Sign Language Advisory Committee with different organizations and really just making sure, again, that they’re doing a collaborative approach to make sure that sign language is accessible and that these tools are accessible to the individuals who actually benefit and will use them. It gets down into acknowledgements and where it’s going.

But anyway, I’m very, very excited to get to try this out. I know some of our folks in our deaf services department are excited to try it out too. It’s just another really great form of accessibility when we kind of look at accessing our phones and our devices and everything else. Like I said, this is only available on the Google Pixel 11. I will say with Android, they always seem to give the best tools to the Pixel, which I mean, I can’t blame them. I’m an Apple user, and you only get iOS on an Apple phone anyway. So anyway, very, very excited again to get to try this out.

And for our listeners out there, if you have used this tool on a Pixel 11, what were your thoughts? Did you like it? Did you find it useful? Do you find it helpful? Do you think it still needs work? And are you hoping that it gets better as time goes on? But I think this is a huge step forward. I can’t wait to see where it goes with just making things a little bit more accessible for folks who use ASL to communicate, and then on to folks who use other forms of sign language all around the world. So I’ll put a link to this down in the show notes so that you can go check it out. Hopefully it’ll be available on more phones here very soon to be able to help out even more folks.

Listeners, next up on the show, we’re very excited to welcome back Ale Gonzalez from BridgingApps with an app worth mentioning.

This is Ale Gonzalez with BridgingApps, and this is an app worth mentioning. This week’s featured app is called YNAB, which stands for You Need a Budget. YNAB is a personal budgeting app that helps users take control of their finances by giving every dollar a purpose. Users can create spending categories, set savings goals, track expenses, and monitor their progress all in one place. The app also syncs across devices, making it easy to stay on top of your budget wherever you are.

Our favorite features were the goal setting tools and simple budgeting approach. YNAB encourages users to plan ahead, build healthy financial habits, and make informed spending decisions. It can be especially helpful for individuals looking to improve money management, save for future expenses, or create a more organized financial routine. YNAB is available for both Android and iOS devices, and offers a free trial with optional paid subscription plans. For more information on this app and others like it, visit www.bridgingapps.org.

Listeners, if you’ve been listening to our show for a while, you know that I am a big fan of alternative access methods. I love when there’s options for folks to access their technology in the way that they prefer. With choices comes freedom and independence, and I’m a big proponent of both. Today, I’m excited to welcome Jason from Fluent to the show. And he’s here to talk about how Fluent can open up a world of control using AI, and I cannot wait to learn more about it. Jason, welcome to the show.

Yeah, thanks so much for having me on.

Yeah, I am excited to get into everything, but before we start talking about Fluent, could you tell our listeners a little bit about yourself?

Yeah, definitely. I’m Jason. I’m a fourth year data and computer science student at the National University of Singapore. I’m originally Indonesian, and I moved to Singapore to study, but now I’m here in San Francisco to go pursue this startup dream of mine.

Awesome, awesome. Well, onto that startup dream of yours, I guess, start us off with telling where did the idea for Fluent come from?

Well, Fluent was actually more of a pet project rather than a startup at first. One of my family members had Parkinson’s and he lost control of his hands. And so I kind of wanted to help him get back to his usual working flow. And I built a prototype for Fluent actually just for my first customer, which was my uncle.

Oh, wow. Very cool. Very, very cool. I always love finding out where the inspiration comes for things. Jason, what is Fluent?

Well, Fluent is a way for you to control your computer with just your voice. And the way we do this compared to other products, like for example, the Dragon or Tobii Dynavox, or anything in that kind of segment is, instead of controlling the individual actions on your computer, you can simply tell your computer what exactly it is that you want to do. And instead of having to go through all the different clicks, the drags, the movements of your computer, you can just tell it exactly what you want it to do, and it’ll figure it out.

Oh, very cool. Can you give us kind of an example?

Yeah. So an easy example of this would be, instead of trying to go to my email by, for example, clicking the start button, searching Outlook and searching for a particular email, which takes a lot of clicks, I can just say, “Hey, Fluent, check my email for any missing assignments,” and it’ll do exactly that.

Oh, wow. Very cool. Very cool. So I can really… I’m sure the learning curve is much less. Some of the other tools that you mentioned, there’s a little bit of a learning curve to get everything down, get the correct voice commands, the other kinds of things. Sounds like Fluent’s just kind of able to work with the user.

Yeah, totally. I mean, I’ve talked to a lot of people in the space, and they love the existing products and I appreciate those products as well, but as someone who’s tried those products alongside my uncle and some of my other patients, well, I realized that simply the learning curve is too steep.

Yeah, yeah. No, it really is, and it can be a huge barrier for individuals, for sure. Jason, I know you’ve been working on Fluent for a while. How has it changed since you first kind of came up with the idea?

Well, to be frank, I was a bit more ambitious in kind of the bad way in the beginning. I had a lot more inputs other than voice, like for example, gaze or switches or haptics or gestures. And that really sounded cool on paper, but what I realized after talking to so many patients, so many doctors, so many people actually using these products is simplicity really is the best. And what we’re going towards right now, not just in the disability or motor impairment space, is using voice as an alternative to the keyboard and mouse.

Nice. Very nice. Yeah, and I love that you were trying all those different input methods and just came back to the one that a lot of folks can use, and it can definitely fit and be able to assist.

Yeah.

Jason, I guess what’s on the horizon for Fluent? Where do you see it going from where we are now?

Well, it’s currently just a tool in my laptop, but the obvious next step here is to get people to actually use it. And on that front, I have around 1,000 pilot users right now all across the US and Southeast Asia, because that’s where I live. And we have a bunch of different partner organizations, around 70 in fact. And that’s accessibility centers, rehab centers, different doctors, hospitals, and accessibility networks.

Nice. And are they just providing you feedback, test cases, and everything else?

Right now, it’s kind of a mix of them giving me feedback and them being direct users. I like to think of it as the middle line here because you can’t really get good feedback unless you’re getting it from someone who’s using your product.

Most certainly, especially folks that need the product as well. I mean, as somebody who could probably benefit from it, just talking about myself, I could always give you feedback, but if it’s something I rely on on a daily basis, it makes a big difference. And I can probably give you much better feedback if you’re using it with folks like that. So Jason, I guess, how is Fluent able to adjust to the user? You kind of said it could do a lot of different things. I can just tell it what I need to do. It can go through the different steps. I guess, how is it able to adjust to different needs and different commands?

Well, it’s funny you say that because we actually have a pretty new feature, which is personal context. So instead of just assuming that everybody’s the exact same, our AI actually stores on completely local storage. So that means none of the data goes to me because I don’t really need that data anyways. Well, or at least the AI knows things about you. So for example, what conditions you might be dealing with right now or maybe any potential tasks that you might be handling. So for example, if I have some important emails that I’m really waiting out on, instead of scouring through every single email when I tell it to look for important emails, it’s going to remember those emails and specifically look for those emails.

Nice. Very cool. And I guess just because it’s kind of a big topic and just to stress this, so the AI, all that data stays local on my device and isn’t going out there into the world, is that correct?

Absolutely. And because this is an AI that does things, another additional layer of security that we have is just the guardrails. So instead of letting your AI roam free, potentially delete all your files, send a breakup text to your girlfriend, I mean, it prevents all those things from happening before they actually happen.

Nice, nice. Well, good, good. Yeah, you said it gets to know you and everything. I guess you don’t want it making decisions for you on the backend or anything like that. Jason, as you’ve kind of been building Fluent and you’ve had folks using it and everything, taking feedback, can you tell us a story of someone’s experience using Fluent that maybe, I don’t know, surprised you, sticks with you, or something like that?

I’ll tell you a moment where I felt like it really did click outside of the kind of accessibility niche. And I know that’s not really what we’re here for, but something really interesting was when I actually gave this product to one of my friends who isn’t really struggling with any motor impairments or any similar conditions, just a regular student trying to get his work done. And instead of feeling like this was a kind of bother, or maybe a less efficient alternative to the keyboard and mouse, he actually found that it was a quicker, faster, more natural way of doing his work because it was kind of just like telling another person, “Hey, can you do this for me?” And relying on that person to finish that task.

Very cool. Very cool. So Jason, as far as, I know this can kind of work on your computer, can it work on any computer? Is it Windows, Mac? How does it work on that end?

Yeah, I have versions for all of Windows, Mac, and Linux. I will say that Windows is the most stable just because I’ve done the most testing on Windows. But if anyone wants to try it out on Mac or Linux, feel free to let me know and I’ll send you a copy.

Awesome. If folks do want to find out more about Fluent, or even kind of request a copy, what’s a good way for them to do that?

Yeah, well, we’re actually shipping a landing page pretty soon. It’s fluentforall.com. It should be live right now, but I’m trying to make it look prettier.

Perfect. And is that Fluent, F-O-R-A-L-L, .com?

Correct.

Okay, perfect. Perfect. We will make sure to put a link down in the show notes so that folks can easily find that, find out more, as well as even be able to request a copy so they can try it out on their own device.

Yeah, that sounds great.

Jason, what did we miss? What did we forget to talk about with Fluent, with the journey of making it, with who it could assist and everything else? What did we miss?

I guess one thing that I would love to talk about is my time here in San Francisco. Because I’m here as an Indonesian studying in Singapore going to San Francisco to build this product. Well, I actually got chosen for an accelerator here called Founders Inc., which is a pretty good accelerator with essentially 99 other people in the same office also building their own startups. Some of them actually also building in the accessibility space. There’s someone doing EMGs for controlling your computer, which is pretty similar to what I’m doing.

And I think just this general experience of being around other young entrepreneurs who are really interested in changing the world and making it a better place, and I mean some of them making a lot of money as well, it’s just been a great motivator and a reminder that if you really do want to do something and build something and make something that is important to someone else, you can just do it.

Nice. And I guess that’ll lead me into kind of a question. Jason, let’s say that we got listeners out there that see a need, be it a family member, a personal experience kind of like you had, or maybe something they just kind of see out there. What kind of advice would you give them as they maybe look to make these new accommodations and new things to assist people?

Yeah. I think biggest advice here is that you can always just get into something even if you’re not from that kind of background. So specifically if you want to make a program specifically for your needs or if you want to make say just generally anything that would help you out, you can always search up guides on how to make a program or you can use AI even.

Nice. Awesome. Awesome. Yeah, definitely. If you see needs out there, for sure folks, most of the folks we have on here maybe didn’t start in that field, or even did and then just found a mission or found something where they were able to go out and make something to be able to assist folks. Well, Jason, thank you so much for coming on the show. Thank you for building Fluent, and we look forward to kind of watching your journey and hopefully some of our listeners can go out there and get it and start using it and have much better access to their technology.

Yeah, that would be the dream. Thanks so much for having me on.

Do you have a question about assistive technology? Do you have a suggestion for someone we should interview on Assistive Technology Update? If so, call our listener line at 317-721-7124. Send us an email at tech@eastersealscrossroads.org, or shoot us a note on Twitter @INDATAProject. Our captions and transcripts for the show are sponsored by the Indiana Telephone Relay Access Corporation, or InTRAC. You can find out more about InTRAC at relayindiana.com.

A special thanks to Nikol Prieto for scheduling our amazing guests and making a mess of my schedule. Today’s show was produced, edited, hosted, and fraught over by yours truly. The opinions expressed by our guests are their own and may or may not reflect those of the INDATA Project, Easterseals Crossroads, our supporting partners, or this host. This was your Assistive Technology Update, and I’m Josh Anderson with the INDATA Project at Easterseals Crossroads in beautiful Indianapolis, Indiana. We look forward to seeing you next time. Bye-bye.