Extra: Customer Service with AI

So who are you?

Yes, thank you for having me here, Robert. My name is Akrit Prasad. I'm one of the co founders and CEO of a company called Aptig. And Aptage was born out of my pain points in my past life working as an operator for about 20 years. And I spent several years working at companies like Salesforce, Appdynamics, companies that build technologies that most people use today. And for us, one of the biggest learnings that I had in my operating experience in these companies was how do we empower our customers to get the best experience possible? And I had an experience once where we had updates that happened in our product that led to lots of customer escalations and problems. And it's challenges like that that inspired me to build a company like Aptig for empowering customer service teams.

What kind of businesses would use Apt Edge and what kind of people in those businesses? Because I've been to places that make baseball bats to planes. They're very different.

Yeah. So we initially built Aptage for the type of business and industry that myself, my co founder, we came from, which is more the technical support audience, more B2B SaaS companies. But what we've seen is that as we started with some of those companies initially, the value of what we're solving started to resonate with many different kinds of companies as well. And so we ended up actually signing contracts and bringing on customers that are in the manufacturing industrial space, such as Hexagon AB as well. We also started to see interest from education tech companies, from companies that are in more of the retail industry, where they have a lot of internal knowledge or social or media industries. So we've actually started to see interest from a lot of different kinds of companies. But our core customer base originally was more on the B2B SaaS, technical support audience. And the problem that we're solving, we've started to see is a much wider problem that really every frontline support team member faces.

So I've worked a few jobs. What is AI doing for these people now? You know, because I, I know if I need to clean install my TV, I can go to GPT4 and ask it, how do you clean install? And it, it lays out the instructions for how to do tech support at, at that level. What is it doing inside companies for people?

Yeah. So when you think about generative AI technologies, a lot of what we see with ChatGPT and these products, right, is information that's, that's learned from the history of the Internet. So these products like ChatGPT that we can use as a consumer is trained on the history of the Internet to find answers to any question we have. And so it's kind of that step function forward of an experience where instead of searching in a web browser and then finding 10 articles, then reading through those articles and queuing, getting to an answer, a system like ChatGPT just gives you that answer right away. So that's a huge area of value. Now what we're focused on with APT Edge is how do we provide that same type of experience in the world of a customer service rep that are dealing with frustrated customers calling in, trying to get their issue resolved as quickly as possible. So the way that we look at this is imagine if you could take that same experience and the ability to surf through all the knowledge and create answers and empower it on your enterprise domain knowledge, the knowledge that your frontline customer service reps have to access day in and day out. Yeah. And today what we find is the average rep is looking through 5 to 15 different systems for any customer facing issue to get to the right knowledge and the right answer. And so for us, it's using AI as an enabler, as almost it's a copilot for these frontline support teams to find information faster, to get to answers faster and reduce the time they spend today searching in different systems to be able to get to responses.

So if I'm a CTO at a company of some scale, you know, has 500 employees or more probably, and I have a few of these systems where there's different databases, you know, maybe we bought a company and they had a different system and yep, we didn't completely change over. How do we hook those systems up to your system and how, how do you, how do you ingest? And then we have a lot of problems, right, like oh, what happens to all that data? You know, how do we trust that the data's staying in the firewall and stuff like that we'll talk about.

Absolutely. Yeah, absolutely. And so those are all extremely valid areas that every company has to think about when they look at technologies and solutions, especially in the world of AI products. But the way that we thought about this problem was that the status quo in the current state of these frontline customer service teams is they've got knowledge in many different places and they have to go and look through those areas of knowledge one at a time when they're trying to get to information and then they have to parse that information using their cognitive energy and then use their creative energy to create a customer facing response. They can send out. So our approach was always how can we assist in that cognitive load and that creative load? By being able to look through systems of knowledge in one experience and create responses dynamically and instantly. And that was really the inspiration, the way that we plug into these systems of knowledge. We have pre built connectors so we can deploy and be up and running in less than 48 hours. So being an AI native product that was built around this problem from day one, we can actually connect to your systems like Salesforce, Zendask, Confluence, SharePoint, even upload raw PDF data that's got context and information, or Slack or teams conversations where it's more unstructured data. We built a product where we can connect to these structured, unstructured systems of information immediately, be able to scan, crawl through that information in real time to find where an answer is or where an answer may be in sections of knowledge across these systems and then create a customer facing response that's personalized and presented to the user right where they're working, whether they're working inside a system like Salesforce or Zendesk or any of the embedded experiences that we can meet the teams right where they work.

Yeah, that's really interesting. This is a big cultural shift for companies, isn't it?

It is. There's definitely a big cultural shift. But we also see there's been some macro cultural changes. Covid brought in a change where teams had to work in a culture of remote embracement. So they couldn't. If they're running into an issue and they're not able to solve the problem, and maybe there's an article somewhere tucked away, but they can't act. They weren't able to find it because it's 1, 1 of the 5, 10 places they have to look. They might go ask a colleague for help. Now pre Covid, we lived in a world where you could look across the shoulder and ask your colleague, hey, I'm seeing a trend around this problem. Have you seen this too? Once Covid happened, everybody moved into a remote setup. So it actually became even more difficult to get to knowledge and answers because you couldn't just tap into somebody else. And so we thought about that as what if you can arm every person with a collective knowledge of the entire company and make them as smart as everybody else collectively in the company. That's really powerful because now you can make every rep that smart and efficient. That's going to allow them to work better, faster, more productively, but it's also going to allow them to work in an Environment where they can deliver better experiences and be happier. We're actually seeing this interesting pattern where a lot of customer service teams are investing in like a happiness manager. They want to keep their people happy because there's an attrition problem. People reps churn, they go to different companies. You have to onboard new people. At the heart of that is the knowledge problem. If you can't connect information and get teams to answers and knowledge faster, then that whole experience that the reps have with their customers leads to frustrating engagements and leads to low morale. And so at the end of the day, our core focus is the number one responsibility of customer service teams is helping solve an issue, helping solve customers problems. And the root of that is getting to the answer to that problem. And where are the core technology that's empowering that experience, helping them find the answer faster.

This is really important. I visited a new bank down in Brazil and they built their own customer relationship management system. So if a customer calls into the bank, they get routed to the right person. Don't have to wait in a phone tree and stuff like that.

Yeah.

So let's get nerdy. I've talked to a lot of startups and they're telling me about their architectures that have a large language model running everything. Some of them are even starting to use multiple large language models now to argue out things and decide whether answers are correct and stuff like that before they pass them along to do something. Right. Tell me, get nerdy with me. What does that look like?

Yeah, so let me. I'll talk a little bit about the language models themselves because that's a huge area of innovation that's continuously growing and ramping. Language models are. The way I think about them is they're. You have to think about the tasks that they're there for. So some map models are really good at summarizing information. Some models are really good at being able to create responses to contact that you're feeding it. Some models are great at. There's models that do things with images, like the Dall E model, for example, but it's identifying which model is the best for the task that you're looking at and being able to use that in a way that's tied to a practical use case that you're building a product for that uses long edge models. So for us, the language models were always about the delivery layer of the product. And we were familiar with generative technologies even four or five years ago, when it was a bit of an earlier inning of the technology. The way that we thought about this problem was three pieces. There's if I'm a support rep and I need to go look through different systems of knowledge, the first thing I have to do is log into all the systems and search in each system. Then I have to understand the context of what I'm looking for and read through those articles in each of those systems. And then I have to create a response based on that that I can send to my customer. If we break this down, there's three pieces to this. There's a layer of crawling through information and that's an area of automation. We can automate the work of having to look in different places and crawl that. Then there's the cognitive work that a human has to do. The cognitive work of actually trying to find where the answer is and understanding the context, the semantics of it. Then there's the creative piece of it which is now how do I create this answer into a response that I can send to my customer? So it's super personalized. Maybe it's in the language that they want to see, it's a different language like Spanish or they want to see it in a bulleted list, or they want to see it in a really friendly tone or a more formal tone. And so the way that we designed App Edge is solving all three of these in one experience. A system that can connect to the different systems of knowledge through our automation layer and then be able to use our AI to crawl through that and find what the answer is, kind of replacing that human cognitive experience and then use generative technologies to create customer facing responses, solving for that creative labor. That's really the way that the architecture fits together. Now underneath that, there's a lot that's been built around how we integrate these different systems, How a real time, there's a real time connector into Confluence and Jira and Slack and SharePoint. So if there's a new knowledge article that just got added there, our system knows, hey, this new article might be relevant to this new question that just came in from a customer. For a human to have to go through and look at doing that would be very painful. They have to find the system of repository that the article was published and then go find that article, assuming that they know where it is. As humans, we don't know what we don't know. But this is a perfect task for a machine and AI layer that's continuously connected and streaming and listening to all the systems of knowledge to find where an answer is to any given question coming. And that's really how we how we have kind of architected the technology from day one. The generative piece is actually something we have been embedding into our product experience. However, it wasn't until about a year ago that we really kind of saw a breakthrough in how well it could create responses. So even today the apptage AI layers the system that crawls through different repositories of information and articles, tens of hundreds of thousands of knowledge, structured or unstructured, and finds where the answer or the answers are to a question or topic that a customer is requesting an issue on. And based on that, we're able to use generative models to create a customer facing response. So our tech can still find the answer, but our end to end use case is really creating customer responses. And that's where the language models come in. They're great at language manipulation, translation. That whole piece of, of the equation sure is.

It's a new world we're in business people, I mean my wife works for VMware in the conference team and even the tech people who go to the conference, which run, you know, the world's biggest businesses, they're freaked out about AI. They're constantly like, they're worried about their jobs, they're worried about, you know, all the problems that AI can bring, you know, and we can detail them for hours. There's a lot of new problems that this thing can bring potentially. But to companies, they're really scared about their intellectual property. Right. And your customer service people really can give away a lot about your business, your business health.

Right.

So there's a lot of fear there. How do we know that your LLM isn't shoving a whole bunch of stuff over to OpenAI or somewhere else?

Yeah, no, that's a great.

Helping your competitor out. Right, Exactly.

And that's for us, the data safety security trust has been the number one principle from day one. Because when we set out to build this product and this technology, we knew that for us to solve the problem we were trying to solve, we have to connect to the systems of data. We're trying to make it easier for people to find answers. And so we have to connect to the areas where the answers are. And so we had to think about that data experience from day one. And part of that equation is the policy and the compliance side. So we were very early in getting SOC 2 type 2 compliance, HIPAA compliant, running in a very secure cloud environment where everything is encrypted in transit or in rest. So we took a lot of efforts to build that secure environment and on Top of that, we've also invested a lot in the way that we have built our integrations. So our integrations are built in a way where they can be scoped to certain type of information or data in a repository. So for SharePoint, we might only connect to certain sites, not all the sites, or Slack, we might only connect to certain channels that contain information. We don't want to index anything beyond that common knowledge that's needed for customer service teams. So we allow the way that we connect to the information to be very limited and scoped. So it's only tied to the knowledge that's related to what can drive value for the front line teams as they're answering questions and not trying to look into anything else that's not pertinent to the core use case that we're building the product for. And so those are some safeguards that allow us to provide a rich experience, solve a pain point for companies while also being really safe and secure with the data and the access that we have to it.

That's really, really important. Along these lines, the where am I going with, yeah, this is just a new fear, you know, because I don't know if you saw that. I did a psychiatrist therapy session, something like this, having a little conversation for half an hour and ChatGPT was listening.

Right.

She built her, she and another guy built a whole system to listen and write highly detailed notes on the session. And it's pretty crazy. So you start seeing that you're like, oh, this can learn about everything really quick.

One thing that I like to always call out and kind of the more technical domain is there is no like generalized intelligence. Right. There's, there's no like AI that's like, it knows this AI and it's sentient and it's like about to go rogue. That doesn't exist. That technology may exist someday, but it does not exist today. Large language models are really predictive algorithms that predict the next word and the next topic. That based on, you know, a trained model. Yeah. And so they're generating content based on basically a machine learning model. It's generating information to help with that creative labor that we as humans do. Right. I want to be able to summarize this piece of information or I want to be able to create an article from this or plan a trip for me based on everything it knows. So it's really a proxy for that creative human labor to make people more efficient, to make people move to that next level. So people are not doing the rudimentary work anymore. And that's what's exciting to me about this is why do we have to have support reps looking for an answer to a question that's, you know, that's, that's recurring or repetitive? And then why can't we just give them the answer if this has already been solved? Right. And so how do we help them access that knowledge faster? Because their time spent surfing different systems is inefficient time. Yeah. You know, humans should be moved to doing more, more comprehensive, more higher order tasks. And the simple tasks we keep moving to the machines. And this is a very practical use case for a machine AI, right. Can take on some of the simple works of the humans are moving forward and doing more complex things. And that's always been the journey in the past 100 years, I think about the automation era where industrial revolution moving forward into automation. My first job out of college was actually in manufacturing before I went into tech and I saw Frontline, you know, helping create. I was in a steel manufacturing industry and we used to produce, we ended up in. By the time I was there, the company was producing the same amount of steel with 10 times fewer people because those people had moved to higher order functions and they weren't having to do basic work anymore because the machines and automation was doing it. Now this was pre AI, quote unquote, right? There was no AI, it was just automation. But tasks that were manual were being automated so that people can move to doing things that are less redundant and don't require critical thinking. They're moving to tasks that are more critical and essentially we're leveling people up. And that's where I think AI comes in. It's about making people better and moving away from the mundane manual things. And that's how we progress as a society and as using technology.

How long will it be before your system is like her in my ear, If I was working on the support lines at somewhere, I've done that at Microsoft. You sit down at a terminal and start up the system, sign in and then you see a list of calls you can take. You take a call. Now you're talking to somebody, right? Helping them out and you have screens to look at information about what they're trying to talk to you about.

Right.

How soon is it going to be that her is just going to be in my ear going, she needs help with this, Tell her this.

You know, I think it's interesting because we have to think through very carefully how we develop a technology like that. Practically speaking, we're not far from a Technology that can do that because ChatGPT is just the way it's used today as a consumer on using a website logging into ChatGPT, you're using it as an end consumer and it's just you ask a question and outputs an answer. Now translating that into a voice based interaction is just one embedded interface away. So we're not far from that. The ability to create information that is more pertinent to the specific question or problem that a customer has and having context around it that requires a richer interface and a richer intelligence. And so I think we will continue to advance in that direction. Where really these things are like co pilots, assistants, they're here to help us be better. And technology has always done that. Right. Technology from get go we built cars so we don't have to ride horses. We could advance from point A to point B faster. That enabled new things to come up and society advanced as a consequence. And that's the same thing we're seeing here. I will say that interesting automation played a bigger role in automating some of the blue collar work. And now generative AI I think is more, it's more on the white collar work, but it's driving automation efficiency in the white collar space that's allowing that sort of Persona and workforce to advance and move to more complex, higher order challenges and problems that they can work on.

Yeah, I mean AI will eventually help a lot of businesses. My brother in law runs a glass business. Right. And he has a lot of detail he has to keep track of like what, what kind of, you know, glass door is going into the house. Right. And an AI can do help him help with that.

Yeah, yeah. It's any, anything that's also redundant. If there's redundant context or information. AI is a system that can help capture insights and also create responses and context in a more in a way that's creative. Generally creative is generative versus something that's simply kind of a copy paste response. So I think AI presents a way to personalize experiences and solve for the creative labor, at least the generative AI piece. But you have to marry that with practical use cases that are focused on specific tasks. Like for us it's all about how do we help customer service reps get to the answer as quickly as possible, whether they're on a phone, they're on an email, web chat, conversation, whatever the channel may be. Because ultimately that leads to better, happier customers on the other end that are reporting an issue. And if I'm trying to reschedule a flight that got Delayed. The faster I can get that done, the happier I am, better the service airlines company gets, everybody wins. And it's technology in between that's kind of really facilitating that work that humans are manually having to spend cycles doing.

So I turned on GPT a couple minutes ago. Everyone's been listening to us. This is why I start asking these questions. Because ChatGPT can listen and get involved in the conversation. Now it's really crazy. And you can even tell it instructions like, hey chatgpt, can you tell us anything we're missing about customer service in this new AI age where we can use things like large language models to make new systems that can help us work? Thank you. You send it to Whisper API and it converts it to text and then you can send it up and then it reads that real fast and then it gives you, it comes in talks, it says. Your observations about the role of AI in automating both blue collar and white collar work are spot on. As AI systems become increasingly substantial, sophisticated, their applications across various sectors, including customer service, are bound to evolve. In the realm of customer service, AI has the potential to accomplish multiple objectives. Here are some things you may consider. Instant response, multi channel support. And it goes into detail. Personalization, predictive analysis, sentiment analysis, training and assistance, Upselling and cross selling, resource allocation, audit and compliance. They got all that from like two minutes of conversation, right?

And that's the beauty of it. Because it's trained on the history of the Internet, it knows that the general topics of what people. So I think there is 1000% value that it delivers and the way that companies are applying to enterprises or specific domain sequence areas. Because if you use the ChatGPT in a public sense, you have it planned a trip for you or a website, or if you're having a plan like a vacation or a trip, then that might be something where in this case the tolerance for error is much lower. Right. Maybe it planned a trip and it picked a restaurant out of like 10 restaurants. And it's okay, it doesn't need to be the one restaurant because I don't know what I don't know. And so it might be okay. Yeah. But when you're using it for customer service tasks, you can't give the wrong answer. You have to get the right answer every time because that's how you drive that experience. So you have to use it in ways where for us, our system controls, our AI controls the relevance and accuracy and acceptability of answers. We use the generative AI piece to create the personalization layer, the Human like feel of the response that comes from the answer. Got it. But not to create. So you have to use it in certain ways where you deliver that best outcome. And I think that's where if people use it the right way, that's when you get the best value from any technology, including generative AI technologies today.

So you're up and running in days or hours?

In hours. Hours.

Wow.

And then we'll do some training. As with any new product or technology, you have to enable people a little bit, train them. Here's how to use the system in the best way possible. That can take within a week. We'll train folks we usually talk to customer service teams, do trainings and batches once they're trained and ramped up. It's not very different than like going on Google and searching or using ChatGPT. It's a combination of both of those things into your own embedded experience. Right. Where you work and built around your domain knowledge and data and continuously learning from that so that you're able to get the most relevant, accurate experience every time.

What else are you seeing happen in business? Maybe outside of your own business? What's happening? Because I know you go to a lot of conferences and talk to a lot of people. What's happening from your point of view?

But then customer service or do you.

Mean the overall business customer service, something you know, right. That you're seeing something go down because of AI is really changing business or affecting how things are done.

I think. You know one thing that's interesting is when we enter this new macro condition after 2020, 2021, we see the public stocks have compressed a bit based on a revenue to valuation multiple. Every CFO is asking for their company to do more with less or the same staff. You're not getting more budget. Right. But the problem is you still have to keep growing your top line revenue in your business and you still want to maintain good margins. And so you still have all the other pressures, but you don't have the resources. And so you have to invest in technology to empower people to do more with less or with the same. And that's actually been a big accelerant for us because we tell that story all the time. We say, look, we purpose built a technology that makes everybody a superhero. Give them a cape, make them work faster, better. We're all about making people better at the job that they do every day and using technology in a way that's embedded to do that in the best way possible. And so that's something where that's becoming a talk with CFOs and the business executives. I think I would say CFO now are become the most important people in the company because they control the budget. Every company wants to be profitable. That's what the public market wants to see. So that's one, one thematic thing that we've seen. We're definitely seeing a lot of research and go into like, how do we adopt AI technologies in a safe and secure way. So there's a lot of conversations around that that are happening. There's definitely a lot of conversations around, you know, how do we, how do we create experiences that embed AI into different workflows that we have. So every company, every executive is having to think about, okay, if I'm not buying solutions that have some kind of AI capability, then I'm not thinking about the problem the right way. So it's interesting because it's becoming much more, much bigger. Part of the conversation is that AI story in every product. But we are seeing use cases of AI and generative AI being mentioned and being consumed in technologies that are beyond support in sales, marketing, engineering. There's a lot going on that Microsoft is investing in like the copilot experience and GitHub and like, you know, being able to write code. So I think there will be more and more proliferation and infusion of where these technologies go to empower people. And the technology itself is rapidly evolving and getting better based on how people are using it and being able to train it to get better and smarter. So I think we will see that. I think training, empowerment, that'll be a big, big, big motive around this. And the companies that don't embrace it will be the ones that get left behind. It is unfortunately that is kind of the what happened in the world automation world 20 years ago. The companies that didn't embrace it fast enough got displaced by the ones that did because the ones that did were more efficient, more profitable. They could take their business away, they could become, they could be more competitive. And so it is kind of a race in a way to embrace this technology and use it in a way that empowers your business to be more competitive and better.

Perfectly said. I usually end this new series just asking me, what other companies do you find interesting, you know, competitors or otherwise, you know, in this new AI space? And it can be something outside of your business, you know.

Yeah, I didn't thought about that before. I mean there's a lot of exciting, interesting stuff happening.

There's 1100 companies I found just this week on Twitter on X dot com.

Yeah, yeah. I think the way there's definitely going to be a lot of applications of generative AI in gaming and social. So that's going to be interesting to see. There's going to be a lot of applications of it. I see more on the enterprise business to business world but I'm also starting to pay attention to what does it mean for the consumer side. Right. So it's interesting. I think it'll be interesting how it gets incorporated into cars and just hardware kind of technologies. So I'm sure that Tesla will probably be one of the first ones to think about how do we be forward thinking and embracing technology.

They're teaching AIs with video watching now. Watch some videos and learn a new task.

Yeah, they're definitely one of the most progressive forward thinking companies embracing technology to drive value for people. So I think there's that and then there's, you know, there's, there's going to be an infusion of AI in different devices and experiences. We look at it in early, the early onset of it was like the Siri and the Alexa and like these things but you know, they didn't really take off in the right way because I feel like it kind of took off. People still use it but the end use case did not tie to something that was so practical that it tied to like a hard value.

Hey Google, are you there? Of course I am.

It's there.

Yeah, but I need you to stop now. It's starting to have a conversation.

Starting to. It's.

Just starting to get large language.

Model and I think underneath it it might be the Bard model that Google develops.

Well, not quite. Unless it changed this afternoon. Not quite yet. See Chat GPT understands that.

Right, right, right.

And Google doesn't. And Google fix that someday.

Right. And Google's got their own models. They've got the Bard AI model which is all the data. Hey, Google stock please. But yeah, I think that is. I'll be here if you need me. Thank you. There'll be definitely more infusion of those technologies into these home devices. You know I think it's also an interesting privacy topic though at that point because I've been reluctant to use those because I don't know how much of it listens to what I do and say.

Assume it's listening to all of it, then you'll never be shocked. I mean you saw how easy it is for a computer to listen to a conversation and get some value out of it and start doing stuff with that.

Right, exactly. Yeah.

And so my lights have microphones and my Sonos has microphone. My Sony TV has microphone. My camera has five microphones, and we're wearing more microphones.

That's true. Everything is captured.

How many things in the future are going to be listening to us? All of them.

Yeah. And so I think it's. It's, you know, there's something to be said for your. What's the term I use for this?

Privacy.

Privacy. And like your.

It's a new problem coming. Yeah, absolutely true. I mean, I assume we're going to be wearing devices with 14 cameras and nine microphones on them.

Exactly, exactly. So how much of that do we really want to expose? Right. And how much of who we are do we want to give away into, like, the ether, so to speak?

We're going to give it all away. We're going to give it all away. All of it.

It's kind of a scary thought.

It is a scary thought. Luma AI company just came out with a walkthrough app, and you walk through with your iPhone and it ingested my whole house in three minutes.

Wow.

So. And when it ingests the whole house, it ingests every object. Because computer vision can look at that Gaussian splatter nerf and go, oh, there's a Kleenex box right here.

Right.

You know, and know how many Kleenexes I've been eating using since the last time.

Right. Yeah.

You know, and so you start thinking that way.

No, you're right. I mean, it's seeing. It's one thing to listen and then to be able to see. Now you're kind of getting the peripheral and it's kind of. It is invasive. Right. To maybe people. I mean, but I think it's the people's choice. Like, it's going to bring a lot.

Of utility, and that's why you're going to do it.

I think if there's a way to do it in a way that's more limited, like we believe if these systems like Alexa Siri are only listening when you ask it to listen, that's a bit more of a controlled experience. And that's kind of how we have thought about our technology for service teams is connect to the limited systems of information that pertain to the tasks that we built this technology for. And so not being invasive about doing everything right. Because you're right, a lot of things are proprietary to a company, and it's important to keep that data safe and secure and empower that experience for the company, not against the company and this.

AI can see a lot of patterns. We didn't talk about that. What kinds of patterns do you think your systems over time will see as the health of a company? Because if you're watching 50 people taking support lines, you know a lot about that company.

Yeah, the patterns are there. I think the patterns are also specific to the tasks that the product was built to solve for. So the pattern and the intelligence that the system can pick up and learn are specific to helping with that specific task. It's not designed to go into other things where it becomes a bit more invasive and involved and does things incorrectly. It's really designed to get even better at the one thing that was brought in to do, which in our case is helping support teams answer the customer question as quickly and accurately as possible.

Yeah, what a crazy world we're going into. Thank you so much for helping me do this show and get going. Where should we learn more about you?

Yeah, we have a website, Aptage IO and we have a get started for free sort of experience on there. So if there's more questions that anyone has, we'd love to have people come to our website to sign up. We also have an alias or salesptish IO so people can reach out there as well. And thank you so much, Robert, for having me on the show.