#112: Q Branch: Transforming Courts with AI & Automation in Maricopa County

Today we have Aaron Judy, chief of Innovation, AI at Maricopa County, clerk of Superior Court. Aaron, welcome to Tech Tables. Appreciate it. Sorry for the delay. We missed our last opportunity. And why and why didn't you just tell the audience why you were delay? Why didn't you come on tech tables? What, you know, I was hours away and then, and my wife and I had a baby, so, I literally texted you from the hospital.

It's not gonna happen. I remember saying that [00:01:00] morning, unless this baby shows up, I'll be there and son of a gun. It showed up. Yeah, it was when we were very happy that you had your priorities in order and that you were with the child and not at a live podcast event. The nice part about shooting today's episode recording it is that everything's already done.

I already love it. We already got the podcast title. I don't normally write the podcast title until after the event. Or until after the podcast. But for the live events, I typically will write them before. So today's podcast title is Q, branch Building. What's possible? Obviously, Q Branch refers to James Bond oh oh seven.

Uh, I'm a big fan, the r and d division within m i six. Your division within Maricopa County Superior Court is structured on the innovation side. That allows for POCs and exploration of what's possible. I absolutely love that now, but before we jump in today, Aaron, why don't you walk us through Q Branch, which I love being a huge James Bond fan and maybe just how you're thinking about Maricopa County Superior [00:02:00] innovation side.

Q branch. We go by a few things. Coincidentally, Q Branch is the plaque. That's actually the name played on my office door downtown. It says Q branch on it. Q branch, skunk works. The house of ideas. These are all, uh, AKAs, aliases and monikers. W our role is the, we saw this technology, we follow technology, emerging technology very closely.

We may come up with an idea on our own, or somebody may come to us and say, Hey, we have a need. We need to fill this gap. We need a solution to do X. And we may jump in there. Sometimes we sit in other people's meetings and just absorb and listen in and say, Hey, we could do that differently. We pair our technologists with subject matter experts, our technologists, we try to keep pure in that.

We don't saddle the innovation group with operations. I'll elaborate on the operations there just a little bit cuz there's also a caveat to that, but we focus on the technology. I am not a court expert. Don't tell anybody I said that. I am not a court expert. I'm a technologist, but I'm paired with an innovation [00:03:00] delivery manager.

He is my sounding board. He is a subject matter expert on everything that has to do with the court. He will tell me if I'm crazy in advance. He's the canary, right? He's also the individual that start contacting teams for change management, right? Hey, these things are coming. We're thinking of these technologies.

He's, he will organize focus groups. He will build recipient teams, the catching team that's going to start leveraging a technology. He's built our contact center, for example, who curates and is the human handoff for our conversational AI as one example. I love that. Today's podcast episode is sponsored by Sentinel One.

Sentinel One redefined cybersecurity by pushing the boundaries of autonomous technology with its singularity, XDR platform. Sentinel One is the leader in endpoint protection and beyond. And thank you for Sentinel for showing up for the live podcast event in Phoenix. Love to have you there. Love Steve Bell who was there and you can learn more about [00:04:00] sentinel1@sentinelone.com.

All right. Now we're gonna hit off all of our podcast questions I'm very excited about. Let's start with RPA Robotic Process Automation. Recorded great podcast episode with Krishna Ethel, who's the head of AI and Cloud at the state of Texas. We dived into what the state of Texas is working on with this cloud and AI COEs.

That's, I think, believe episode 40. And before we jump into the, into the three big AI initiatives that you're working on, could you maybe just give us a baseline difference between rpa. Into ai, how are they different? And then how does AI supercharge? Sure. Rpa, um, years ago there was no such thing as rpa.

There was just automation. We were writing scripts to do things, and that was our first automation. When we started making chat bots, chat ops, those types of things, we just plugged our automations into that and that became what we thought was Automation Today, rpa. Is a standalone system. By definition, I think Gartner calls it that a standalone system dedicated to either event driven or AI [00:05:00] driven automation, pre predefined processes, or they have a great definition of it.

Really, we think about it, is the last mile. And I can, I, when we get into some of the projects, I, I can talk about that. But for us rpa, it, it has to be. It has to be headless, right? If I asked you to press an extra button today and all your work will be done, you'll be like, oh man, I gotta press this extra button.

It sounds like more work. You don't wanna do it. And ownership, who's gonna do it? Do you own your R P A? Those are the two biggest hurdles is its intrusiveness and who owns it. Yeah, no, that's really great. I am a huge fan of automation, end of scripts. Anything I can automate and do in my own business I am doing, and trust me, there's a lot of automation in the background and no one sees.

I love that. So when you know, why do you think there's a lot of like fear regarding technologies? Like RPA or AI or ml, you name it. Why do humans [00:06:00] themselves close themselves off to the art of the possible? This might be more related to the end user than the technologists per se, but the end users are still there.

Um, I think you and I have talked before, fears and institution of culture, how you celebrate and recognize failure. It will structure fear and your tolerance for risk, right? Ultimately, obstacles, expectations, obligations, explainability, ownership. All these little bits are what stop people obstacles. If you don't have the skilled.

Talent in your shop to do some of these things. It becomes an obstacle. Expectations and obligations. You want to do cool stuff, but this judge is expecting you to deliver this modification to this system, or this customer wants you to do something. You can't go do the cool thing and deliver it first, or else somebody's gonna be mad that you didn't deliver their thing.

Explainability if it is a black box. Now mind you, it has to have some [00:07:00] kind of productization, right? I have to be able to touch it and look and show everybody, even a layman, to be able to access it. But if I can't explain how it works. I'm not gonna put the risk of my public career. I don't want to be in the headlines for this thing spouting hate at somebody and ownership, right?

Who is ultimately going to be in charge. If something is said by a conversational ai, who owns that? I'll leave it to. Those are some really great questions. Now let's dive into the three initiatives. So conversational ai. AI system reliability and computer vision and AI are what you are working on. You maybe just sure touch upon those three.

Right now actually, our cloud contact center that we've got, we've got an omnichannel, a conversational ai, Cleo, who's plugged into the web, into Amazon's voice. First devices a Google's or, uh, she's got text, email, you name it. We try to plug her into it. She's in Facebook. We have human handoff for, uh, many of those channels.

So if she can't answer your questions, she will get somebody they can. But our biggest venture into that [00:08:00] right now is certifying our voice channel. We want to replace our traditional press one, press six IVR system with just a natural human experience. What can I help you with? What can I answer for you?

That means we have to enable certain services to be, to be voice first. I'd like to look up a payment. Those types of things. You need to lead individuals through those experiences. Our sre, Alfred SRE is a system reliability engineer. Traditionally, this is a human being that would receive alerts, maybe, excuse me, triage incidents with your public facing services or website or something to that effect.

In this case, we have an AI that receives our alert emails, tries to act upon those emails and will send its results or necessity for remediation onto the on-Call Human for review if necessary. And right now we're looking at expanding its capabilities into other systems, letting it baby. Our mission critical systems.

So if something happens with them, it can at least [00:09:00] try to triage that first line of support. And then our last one, which is probably the biggest one, it's gonna be huge savings for us, is our intelligent capture project. It's a two phase project. The first phase is supposed to realize something like 33% in labor reduction bar coding and human sorting.

And phase two, something like an additional 40%, which is profound when you think about it. That's one of our core competencies is document management and fiduciary for the courts themselves. So that's our main line of what we do. We are looking at using NEX to drive rpa. That last mile, what we were talking about a minute ago, I want nex.

To drive a website that there is no API for that a human traditionally has to drive and make decisions based on these inputs we're going to, we're gonna use what we have right now is a quorum of AI that agree on the document. Several pieces of AI look at the same document and say, I see this is a case number.

We both agree that's the case number and that helps with [00:10:00] the routing. Once we've extracted all those features and have all the key components, we're gonna input those into another system and relieve a great portion of, of human effort. Yeah, there are a lot of really fantastic use cases from you're on the ground with the court system.

I'm also thinking like in the medical field stuff where there's like crazy. An unbelievable amount of documents, specifically like excessive and duplicative like documents everywhere. Every time I go to the doctor's office and show my health insurance card, which by the way, in California, you can have on your iPhone, but no doctor ever takes it and they have, you still have to hand over your plastic piece and then they scan it and make a photocopy of it and then file it.

And it didn't change from the weird two weeks ago where I showed up to the same place. That's right. It's really weird. But creating all of [00:11:00] that, and especially like in the court system, I can only imagine is probably creates some serious headaches for humans, does that. Labor reduction and document management is huge and so if you can get that off the plate where then the human can actually focus on.

The more important tasks that'll be huge for Maricopa County Superior Court. Curious, like what innovations have the most promise from what you're seeing in the gov tech world right now? Simply what excites you the most? Um, I would've to say composable architectures Legos. That's what I, it's not the Ironman, it's not, that's not Iron Man actually kinda is in a way.

I use the software from it to build r r two or R four. Mine's R four anyway. Yeah, composable architecture. That means I can take the best of breed of technologies and stack them like Legos. If I decide tomorrow that there is a better product out there, I'll just unplug it and plug in a new one. That's something, when you're looking for technologies, now look for those [00:12:00] with.

Great interoperability, well documented APIs a lot to offer a good community behind it. Uh, computer vision, that's what our document management thing, that's really what we're looking at, is having a computer look at and understand what a document is. That's not just the words, that's the shape of it.

Where the header is, that's fingerprinting the document, right? Where are these layouts? If there's a picture on there, yes, is this a horse or is it a cat? That kind of thing. Deeper into, it's a hard problem to solve. Is this document, does this document contain human input? It's easy to pass a document to a service and get the characters off of it, but to say, Bly, yes or no, does this document have human input on it is actually harder than you think?

It's, and that's one of the things we're trying to solve now, because if I can solve that, I don't have to hand off to more expensive cloud services if I can solve that locally. It doesn't have any human input. Don't give it to the handwriting service. Yep. Yep. No, definitely. And what I was gonna go, what ideas do [00:13:00] you have that aren't prototypes, but could have promise in?

Uh, you know, I've got this lofty idea for, uh, an augmented reality campus experience with indoor navigation. Bruce's got some great products for indoor navigation, and I would like to see a, imagine, imagine you're late for court, right? This could be your freedom, your livelihood, your children. Right. You need to be there.

Some other circumstances made you late. Imagine having your phone, and this all stems from experience I had with the customers. The gentleman came through security only to get his effects back and say, is there a courtroom be in here and no, there's no courtroom. Be in this building. Imagine if he was able to hold his phone up and have a way point down, pointing down the street, saying three blocks up on the left.

That's the building you want. And then when he gets inside, he or she. They get inside and are able to navigate the building down the hall, make a right. I'm at my courtroom. That's significant. That's changing the world right there. Yeah. I read like this, I've not had to [00:14:00] walk through a courtroom, but I can picture, uh, similar analogy would be like with business services, where in, for example, in Santa Barbara, I have to go file a document with the city or I've gotta pay taxes.

And this is a very thing where I am trying to. Make a payment because the city of Santa Barbara doesn't take online payments and I Tell me about it. I know I am walking through a dungeon to find this room. This example would be fantastic if I could actually take my phone out. And navigate my way through the number of government offices that are in downtown Santa Barbara.

This would be fantastic. So I love the augmented reality use case. I think it's pretty popular too, just with you take on the consumer side, whether that's putting IKEA furniture in your house, whatever the metaverse is, right? Different shoes. Yeah. Yeah, exactly. Probably is gonna take some time, but [00:15:00] I think there's definitely some promise there.

Okay. My next question is from Shauna Rogers, chief Digital and date Officer at the State of Texas Attorney General's office. So I had her in Austin a few weeks back now when she was actually at the live podcast event. So Shauna asks, How do you measure adoption and relevance in technology to your constituency?

How do you measure what's working and what's not measured use? And that is through observability from the beginning. And my secret sauce to, that's New Relic. If you ever get a chance to take a look at New Relic that is observing the application, the services behind it, the servers that they're on, and any other important kinda infrastructure that's underneath of it.

That's my easiest way. If I can say, look, it's getting used a lot, it's getting a lot of traffic. I can correlate that to value in some way. We have user personas where we try to map out what this user is as an internal external, what their value prop might be, and I can tie those two together. So [00:16:00] really trust me, this is what I struggle with.

This is what I have to do. So yeah, I measured use and observability. Those are my secrets laws. Measured use and observability. I wanna add to that. Love it. Sometimes you have to adopt a technology just for social credibility, right? You remember, let's see, back in the the late nineties when two people in the office had email and you were like, when is the customer ever going to email me?

That's ridiculous. And now people think of our email as archaic. So there might be some of these technologies that you're like, when am I ever gonna do that with the customer? It might become mainstream. That's it. Yeah, I wasn't in the office, Aaron, because I was in school. I'm a little bit young enough with email.

I was not at the time, by mainstream, not my hit email was, I think, pretty regular, pretty, pretty mainstream. But when I was in school it, it was definitely not mainstream yet. And it was, I had the old, like originally it was like the old AOL dial up, right? Yeah, yeah, yeah. So this is [00:17:00] before Google and Google Maps or anything.

It was a fascinating time to have a computer, especially as a young kid. I had a tough bun tinkering around and messing around and no Tandy. Tandy computers. That was my first one. Um, oh, Tandy. Oh, that was your first one. Nice. I'll have to go look that one up. I'm a little rusty on all the old computers now, but I only know it's okay.

Apple. That's all right. I played Oregon Trail. Yeah. Yeah. Oh, I love it. Sure. All right, so let's wrap up our podcast with the future of. Digital humans. First, what is a digital human? And second, what is the business application of a digital human? Doesn't have to be necessarily with your job in the Superior Court.

Uh, and then Wolf, if you could talk about it, expand upon a little bit with Cleo. Digital human is a virtual avatar that represents a human as closely as possible. It may reach into what's called the uncanny [00:18:00] valley, where it can look a little robotic or soulless doll's eyes look, but the goal is to get a very humanistic interaction with somebody else.

The more we push customers away, don't come down to the office. Just use my chat service. Don't. Don't use your traditional methods. We get away from the humanization of things. A lot of folks just like seeing a smile when they're on the worst day of their life. Sometimes they want to hand a paper across the counter and just have you go, oh, okay.

What you're asking me isn't crazy. Get feedback at times from humans even when you're not getting feedback. Conversely, not all generations like text, you and I were just talking about it. I was Gen X, I was one of the first texters. I was one of the first Doss users for the generations that came before me.

Value human face-to-face interactions. They may not like a chat bot, they may not like typing or texting things. So how do I reach those customers? Conversely, gen Z, gen Y and Gen Z, they want that same human video interaction. They want more immersive video [00:19:00] interactions, and that's where digital humans come in.

I am trying to figure out what is the value of a smile. Honestly, if I can put a smile on my service. I can talk to you and understand where you are empathically, how you're standing, how your body posture is. Are you sad if I ask you for a death certificate? Don't keep smiling at me. Don't keep with the same boring text.

Have a solemn look. Feel sorry for me. We're humans. We want that. Most of what we exchange in communication isn't coming out of our mouths, and ultimately that's where I want to go with Cleo. My problem is the value prop there. I created a couple different prototypes with a few of the very mainstream top of the line technologies.

Both prototypes have a digital human interface with optional chat interface and the ability to understand, because human sway, we move in space so it can follow you. It understands where you are. Imagine coming into a lobby. This is just my pie in [00:20:00] the sky. Imagine coming into a lobby and being greeted instead of the person at the desk by a full body avatar.

Hi, what can I help you with? I have this form. Oh, let me see that form. You need this desk over here, and then maybe it's not taken number, maybe it is. Bam. Hold your phone up and take this QR code. And I. I've got a lot of lofty ideas. Your phone has a scanner in it that rivals a lot of nineties technology.

You can or automatically rectify, rectify documents and scan them in. Imagine if you had our lobby in your phone with a digital person there that's helping lead you through that process. That's amazing. Yeah, there's a lot of really great ideas. I love what you said about the risk to humanization and what is the value of a smile.

This might be the most underrated. Question of the last, I don't know, decade of. Of the world, and this is a fantastic question. I heard a speaker who's got a book coming out and the book is coming out later this year in [00:21:00] October. It's called Unreasonable Hospitality. I like that by this guy was like a five star Michelin star restaurant owner, and he gave this talk an unreasonable hospitality.

I think I might have some notes on it from the conference and it was just so good. And this is what I think about what is the value of a smile? Because when you have an experience, if you're a citizen, And you interact with the government and you have an experience. Typically it's, to be honest, it's not really a good experience.

Most folks that I interact and consume services with in Santa Barbara are tired, overworked, and I have empathy and perspective, so I understand that and I try and give them a smile actually, because. What do you want? I'm like, I just want to pay you. No. And so I've gotta come with the exact paperwork, like I'm filing this and that.

And I think what is the value of a smile is a fantastic question. And if you could up the level where the service of hospitality is incredible and off [00:22:00] the charts, if it's unreasonably amazing, I think that just makes for a better world and a better envi, better environment in general. We measure customer, we measure customer satisfaction, even though.

Most of our customers are probably court ordered. They're not happy to begin with, but we want their experience. Yeah. Yeah, that's a great example. Another example I've thought about having before Covid, I traveled to a number of countries before, and when you land in the airport, it's actually a very handy thing is if there was something like this, that you could just press English.

Yeah, and it, it would just be like whatever language it is, you're your liaison. It was there waiting on you to be like, okay, let's get you, cars are over here, food this way. So they have the airport, they have the basic signs, but it's like little more detailed stuff. I need an at t m because the country I'm in doesn't operate on plastic or it's not like a really.

Use high usage plastic country. Yeah, I need cash. Where's the nearest atm? [00:23:00] Or I need to make it to wherever I'm staying. And so I think there's, that's one use case. I think there's always a, did you see Nestles use cases? Uh, Nestle be a digital human to help people bake cookies. Believe it or not, most of their calls are about the instructions on the back.

They also had to make it a female because most. The customers prefer, uh, apparently to have a female baker. It's an interesting That is. Yeah. Awesome. Nestle Digital. I'm actually, I'm, yeah. If the audience hears me typing, I'm. Throwing that in to look at this right now. Yeah. Her name is Ro. It's her name after the lady that made the recipe.

That's hilarious. I'm not mistaken. Ruth is the original lady that made the recipe. I love all the titles to the articles. Nestle's cookie AI looks creepy as hell, but could improve your baking. That's hilarious. That's really funny. Yeah. Do you know I'm thinking of cooking now that you brought this in. If anyone in the [00:24:00] audience or yourself has ever ordered HelloFresh or No.

Or Blue Apron or some service like that. So it's like a food delivery service where all the ingredients are out and they give you a card to cook. If you're not that smart like me and you struggle with cooking, you tick out the card and everything doesn't work cuz you're like, oh, I'm throwing all this together.

But if I had something on my phone that could display and walk me through, YouTube's like the closest bridge to that. But even then, I'm really liking this, this AI opportunity. But Ruth is pretty funny. I'm gonna have to go. Uh, let's see. Autodesk has one. There's many companies that have them now. Even Disney's experimenting with one.

I got to see a preview of Ray. So it's actually Ray, it looks like her. Okay. Even her hair's waving in the wind and you can talk and interact with her. Ask her Star Wars questions. It's it's amazing. Clearly, I'm a Star Wars fan. Oh wow. Yep. Yeah, we're a, me and Jack are huge Star [00:25:00] Wars fans. We're dialed in.

We are, we are waiting. Our daughter went to camp, but we are waiting to watch the last O one. I'm gonna make a low when she comes back. Yeah. Yeah. So I'm gonna have to check out that. We saw, I just recently went to Disneyland with Jackman and we saw, we saw Ray, uh, yeah. In life, the human version being chased by the storm troopers, but digital humans, I might have gotten a preview to it.

Okay. So awesome. This is a ton of fun. At some point, I'm gonna have to stop by your offices. I want to take a photo with the Cube Branch Slack. That looks pretty awesome. So where and where can people find you? Aaron, what do you like to hang out? Uh, Dan, for business stuff, if you're into nerdy things, I probably Instagram augmented artist.

YouTube never works. 79, that's probably the best places. The Never Works com has links to it all. The never. Works the.com. Awesome. Well we link to that. We will link to that in the show notes [00:26:00] and please reach out to Aaron on LinkedIn. I'm sure he's stoked. I know I can almost guess my guy, Krishna from Texas is gonna wanna connect with you as he is also working on a lot of innovation RPA in cloud tech there.

Okay, cool. I'll definitely make sure to connect you guys. Awesome. Well thanks for coming on Tech tables there. I really appreciate it. Appreciate it. B

Speaker 2: Hey, what's up everybody. This is Joe Tossett from techtables. com and you're listening to the public sector show by tech tables. This podcast features human centric stories from public sector CIOs, CISOs, and technology leaders across federal, state, city, county, and higher education. You'll gain valuable insights in the current issues and challenges faced by top leaders.

Speaker 2: Through interviews, speaking engagements, live podcast tour events, we offer you a behind the mic look at the opportunities top leaders are seeing today. And to make sure you never miss an episode, head over to Spotify, Apple Podcasts, hit that follow button and leave a quick rating. Just tap the number of stars that you think this show deserves.