conversational AI

DRUID Talks Ep #8 How Can Technology Help the Higher Education Sector Survive and Thrive with Jim Walker

DRUID Talks Ep. 8 explores the 600-year-old higher education sector in the US and how can conversational AI and other techs help it survive and thrive.

Episode #8 of the DRUID Talks webcast features Jim Walker, Advisory Board Member at Roboyo USA, and Subject Matter Expert Kieran Gilmurray. See the full episode and transcript below.

Subscribe for the next episode!

Kieran Gilmurray:  Welcome to DRUID Talks! In this episode, we'll explore the 600-year-old higher education sector in the US. It's reached a tipping point in the era of technology-driven change; it seems designed to resist change. The traditional model has become somewhat unsustainable: there's a rising cost of tuition; there's mounting student debt, and growing concerns around the value of a university degree.

But can technology help the higher education sector in the US survive and thrive? Today, we'll explore these topics with our guest, Jim Walker, a leading 32-year expert, not 32 years old, Jim, it must be said, but 32 years in the industry; an evangelist who has worked at all levels in technology, the public sector, the military and today's topic, the Higher Education sector.

Jim is a robotic process automation expert and pioneered NASA's first in-government deployment of the Washington bot, which is now in our industry, recognized as a seminal moment in RPA folklore. Jim currently sits on the Roboyo USA Advisory Board and is a founding member of the RPA Initiative at George Mason University. Welcome, Jim!

Jim Walker: Oh, thanks for having me. I'm looking forward to today. 

Kieran Gilmurray: Well, Jim, let's jump in because I've got a lot of questions to ask, and I love talking to experts who can reply, just giving me all the stuff they've learned over the last number of decades. Jim, would you mind telling us a little bit about the US Higher Education sector so we better understand its challenges?

Jim Walker: Yeah, so it's amazing when you think about it. If you go back to 1950, about ten years before I was born, there were about 2.2 million participants in higher education. That could be in a two-year degree program or a four-year degree program. That number jumped to about 22 million people around 2020. Right? The number of schools that were helping these students was about 1400 students in 1952 or 4100 schools today.

So, you know, a significant increase in the per-student population at each of these schools. And to your point, we're starting to see this question about: "Should everyone have to go to higher education?"; "Do you really need a degree and something to go do for journalism?"; "Do you need a degree?"; "If you're a good writer, can you just go?".

I would make the argument having a degree myself, that can't hurt to go get some formal training. But thanks to technology, you can get formal training anywhere, anytime these days. And I think the big thing that I kind of am looking at right now is that we learned from COVID that not everybody can learn anywhere, anyplace, that there is now going to be this whole discussion on: "Who is best suited for technology-driven education?" and "Who needs a classroom environment?".

Because I think we've found that not everybody functions well in one or the other. So lots of opportunities for technology right now, I'm looking forward to talking to you about a couple of those ideas.

Kieran Gilmurray: Well, let's dig into a couple of those things here. So, if we look at some of the most popular emerging techs over the last number of years, we're talking about robotic process automation, intelligent automation, and conversational AI. Could education institutes put that technology in, or what are the challenges they might face when actually implementing those technologies? 

Jim Walker: Yes, so they're going to have the typical problem, which is they have this older, in some cases, blended technology platform. Their enterprise is old school stuff, and new students school stuff. And that, by itself, has created a bunch of technical debt for them. You know, their network is built on something, but they're now using WhatsApp, and is that ready for it?

And so you have the technical debt part, the network, the infrastructure. I would say that the schools also have this issue now, where they have two types of students: the digital native and the digital immigrant, right? And so the kid that you see in the back of a car playing with their iPad today, that's a digital native, and they expect everything to be online.

What do you mean I can't change my class at 2:00 in the morning?; I can't access my student information at four in the morning, and I just got off my night shift. And then you have the digital immigrant who is a person my age, the 40 years old that wasn't born with an iPhone in their hand. But they certainly are appreciative of what it brings to them: They do their Ubers with it, They do their movie tickets, whether they get their football seats. And today, you can look exactly where your seats are going to be and which the view you're going to have, so you buy the best ticket you can online. So they're used to that. So you can't just sit back and say: “We're going to stay the old school. We're going to have all these manual processes. We're going to make students come in to pay their fees. We're going to have them come in and request and ask for a copy of their transcript and we're going to get them that transcript in six or eight weeks.”.

So those kinds of issues are forcing higher education to figure out how to be digital. They can push back on it if they want to, but there's an argument that a lot of students are making their decisions are where is the less friction to go to school. I want to take some online classes; I want to go to class; I want to be able to order something on the air for the weekend; I want to be at a register for classes while I'm driving to the beach; I want all of that. And in order to do that, a school has got to become a digital school. 

Kieran Gilmurray: It's interesting that just as you talk about all those things that we can now do digitally. If we were talking about this five years ago, that might have been a "Wow moment", it would have been Amazon to name one company could do it. Now, when we're saying these things, I'm nodding my head, going: "Yes, of course. It's just part of daily life.". And therefore, if there are coming students if they're coming from daily life to suddenly being juxtaposed with an age-old mechanism, then I can understand that that would jar quite a bit and the more technical the institute, the more flexible and available for me, the better.

So, what is the reluctance of higher education or technology or universities in the education sector from actually adopting all of this new technology and coming up to date with all the things you described a moment ago? 

Jim Walker: Yeah, I'm not sure that reluctance is the issue, I think it's rather the resource challenges.

So, there's a whole group of people that would love to see free education. But with free education comes a solid budget. And if you spend more than your budget, you have a problem. There's also the idea we're going to charge tuition, but we've got to keep that tuition down. So the students don't have to take out 100,000$ loans and then complain later that they have to pay for those loans, right?

So there's this cost to saying: “Hey, we went from 1400 schools to 4800 or 4100, and each one of those needs a network and each one of those needs to be digital" and "We're an old school. Let's keep our costs down by not migrating all this new stuff.", right? "Let's not use conversational AI, which makes the students come in because they love us. We're an old school. We have a history, we have heritage, and people will come to us because we're that.".

Well, they might. But, honestly, when you've got inflation going around the world like we have today, you'll be a school that can afford to go to, let me get a degree and get out into the workforce. So I'm not so sure that it's a reluctance. It's hard to think that places that were designed to be innovative and creative and create drama and to create music and to create new business models would say: "Yeah, but we wouldn't want to do that with technology.". I think it's a problem that the CIOs that these organizations have saying: "Everybody wants all this technology. We want my CIO budget to pay for it.", and there's going to have to be a new shared way to pay. But mainly because we're democratizing all the technology also.

Kieran Gilmurray: Yeah, it's interesting that just as you're saying that, Jim, because that's exactly what university was for me, a crucible of sorts, a crucible of intellectual rigour and debate and it was innovative, and you saw professors and students creating the very future in front of you. So that paradox of thought and then potential action behind the CIO budgets, that is extraordinary.

So, are there options, then, maybe affordable options or otherwise? Like, could AI, for example, help the education sector in the US deliver this promise we're talking about of this digital way of working at an affordable rate? 

Jim Walker: Yeah, you'd have to think so. You would... It's interesting to me over conferences these days. We went from, especially during COVID, we had all these conferences that were in-person and then you had big buildings and massive infrastructure requiring people to set it all up, and there was a price for the take; it was call it $1,000. And when we moved to the online format, initially, people wanted to charge that same thousand dollars to attend the virtual event. And I'm like: "Why would I do that?", "$1,000 zoom is meaningful!", you know?, "Give me a ticket for 50 and do it virtual. You don't have all those expenses.".

I think the same thing with education, right? Every class may not need to be in person. It's still going to require an instructor, but an instructor may be able to manage a couple more classes, more than one. Because they are going to be able to use videos, they are going to be able to offer online training that you can go in and get pieces and parts of it.

And, you can make the argument that you could go get 100% of it. You know, if you're a DRUID for conversational AI or UiPath or their academy, there are certain ways, certainly ways to go and get fully certified on a technology. But again, if you're thinking “education”, it's more than learning a skill. It's learning all the skills necessary to be superb at that skill.

And so I would argue that we're going to need to be able to use machine learning to track the students, to find out which students are not going to class and give them a call: "Why are you not in class?"; "Where are you attending?"; "If it's a financial thing that you know, we have this program."; "If it's a life problem, maybe we have a way to help you.".

Maybe it's a young woman who had a baby. Or did you realize that we have this child care program and we can get you on some type of reduced cost for the childcare, but we don't want you? 60% of the people or 62% of the people that attend this college in the US finish it. That means 38% are not finishing. So we can use technology not to track them, to say: "You're not in class!", but to track them, to say: "Since we freed up our back office staff from that repetitive, mundane work, we now have people they can call students and find out how can we help you stay in school. At least, talk to you more about the impact of taking out a student loan. Yes, you need dollars, but you don't need partying dollars. You need your tuition dollars. Let's take those dollars first and get school paid for.".

Kieran Gilmurray: I think if you'd told me that I didn't.. wasn't allowed "partying dollars" when I was a student some years ago, I might have wondered and worried. Someone once said to me, a Professor, Irish guy, Jim Murphy said, still remembering today, he said: "Half of your experience is the learning and half of your experience is the growing.". He did warn me and the rest of the lecture hall, just to be clear, it wasn't just means that if you spend all of your time partying, you'll remember nothing.

You mentioned quite a few use cases there, Jim, and I've seen some others. I've seen some of what I would describe as "proctoring" in AI, validating that whoever's at a distance is the actual person. I've seen a great example from one of the big education companies a couple of years ago that as you went through the curriculum by yourself, the AI was monitoring what you were doing, so it could bespoke tests and it could bespoke the next materials in the next interventions to help you learn, rather than the teacher trying to cope with 160/180/200 students in a lecture hall.

What other use cases exist that technology might be best put in place to try and alleviate some of the other pains in the education system at the moment?

Jim Walker: Well, I would take... So, let's take some of the text that AI. I'd like to take a AI and have it assigned to me to be constantly looking out after my coursework. Because, maybe, I'm going to take a course next fall and I register for it, but I do and then I sign up to take another course, and the AI could say: "Wait a minute, Jim, this course is only offered every three semesters. You've got to get it in now, so that you can graduate on time.". Right? 

Today we have to do all that manually. We have to do that with advisors, faculty members or graduate students who would say they're overworked already and they're going to do that repetitive, mundane stuff is boring. And that's not RPA work, that's artificial intelligence, really thinking through: "Jim's taking these classes"; "He's required to take these classes."; "Here's what the next two years of those courses look like for schedules. We'd better highlight. He's got to take a class now.".

I can use all of the document understanding type of stuff that's out there to scan documents and to formalize what the document is and identify where it goes to get my transcripts. So, I can verify online who I am today using artificial intelligence at RPA. It says: "Here's my driver's license and here's my student ID.". Those two things together say: "That really is Jim Walker. Let's go ahead and mail that transcript electronically to the organization he's applying to work at.". 

And there's another piece we constantly talk about used in the back office for finance or for students that we can change student information online and we can do that kind of stuff. But the other piece of education, of course, is education. It's not just the back office monster. The front of the education are students. And what an amazing opportunity today for two-year colleges, especially junior colleges, sometimes quality in the US, to take a two-year program and say: "First semester is RPA and an introduction to Intelligent Automation. In the second semester, you're going to build your first automation. And, by the way, you're going to interview our HR Department or our Finance Department or our NCAA Athletic Department, and you're going to find the process there that somebody is doing and you're going to build it for them. And when we get through, we're going to evaluate much like the US Air Force does with their robot for every airman, we're going to evaluate whether or not we need to put that in production at this school.".

And so you created a legacy. You're in your second semester, it's a two-year school. You're going to go the next semester, maybe, you do what conversational AI would drew it, and you'd build a bot, a bot that's able to talk to somebody at 2:00 in the morning and help them with their most difficult problem. And if it can't, it says: "Don't worry, we're going to assign you to a ticket. Someone will call you tomorrow morning by 9 a.m.". And sure enough, at age 59, you get a call from a person that says: "Oh, I reviewed your record with a conversational AI last night and I have a solution for you.". Now, you just get this ability and that's just a third semester, and that four semester you're doing much like they're doing with the hardware robots.

My goodness, every high school around here has a robotics course now where they're building the hardware robots to attack each other and all of that. How about we build robots that make this university smarter by having the students develop, now that they know conversational AI, that they know RPA, that they understand all the different kinds of AI that's out there now, that ChatGPT type stuff, and they build solutions for the school? And because the schools didn't pay for them, they can actually put them up on a website, for other schools to have.

And the next thing you know, we jump at a hyper rate, from a manual historic university to a fully automated university that works for the back office, for the professors, for the students, and quite honestly, for the vendors that support that university. All because we put students' minds to work on it, and what did they walk out with besides a certificate, a degree? They built two or three actual automations that are potentially in use at that school. They have a portfolio now to say: "Not only do I have a piece of paper that says I can do this, but I'm also have to prove that I can do this. Here's my project.".

Kieran Gilmurray: Wow, that's kind of like a little... Oh! Bit of a play-back there. That's kind of like a live laboratory, where the university, and I love what you're suggesting, Jim, so they have a production ready automations and if they're shared, all of the universities do, but the students themselves are coming up with, what I would describe as, real world modern in-demand skills, so they're work ready. Because I can imagine if I'm a student, I'm going on, what? 100,000? It's a great institution, but as great as a name I have, I actually want a job at the end of this, and I'm doing that math and I'm looking at the course content. If I come up with real skills that I can talk to about employers, tick for me. And as you say, if there's that production students building automations all at the time across the US higher education sector, then the university is going to benefit, of course, back office and as you mentioned as well, all those other things, the conversational AI, the sports scholarship funding, you name it, it feels like a self-fulfilling prophecy. 

You mentioned conversational AI in the middle of that, and you mentioned one or two examples of students connecting with a conversational AI agent in the middle of the night. Do you think conversational AI is the answer to all of the problems, or is it one of the answers to support higher education institutions today?

Jim Walker: Oh, I definitely think it's one of the solutions, right? If you're an RPA being and you walk in today to a school and say: "I'm here to offer you RPA.", that's only one thing. Now, it comes with things that can be done with: repetitive, mundane work, no question about it. But if you really want it to be smart and intelligent, you're going to be partner with process mining, you're going to process it with document understanding. You could put a voice in front of it or the capability to text in front of it and in the back end, you could hand it to an AI, the day I could do the heavy lifting of the intelligence of it, and then hands it back the output to continue processing.

So I see all of these technologies, I don't see a single one of them today as: "the solution". And I realize everybody seems to be like, ChatGPT to be "the solution". But what I'm not sure we're all ready, right? It just happens to be the that is truly the new silver bullet. But when you take any office in America or anywhere and you say: "I'm going to create a new business.", you bring on somebody for a job and you say: "What skill sets do we need?". And you find the right mix and match of skill sets for the right salary, and you say: "Okay, that's my team.". Well, how about we do the same thing in the virtual world? Let's don't hire all lawyers when we have a call centre. Let's don't hire a bunch of mechanics when we don't have any vehicles. Let's say: "What are we doing?". We're doing a merger with another company. Then we're going to need some process mining and task mining; we're going to need some document understanding; we're definitely going to need some RPA; We're going to use some artificial intelligence, but we're going to need conversational AI, I think, at the front of everything today. I love it when I get a company on a weekend when there's football season and during halftime, I can go in there and chat with somebody. Like: "Oh, it's not that fast because they're servicing more than one person.".

It's halftime on a Saturday! I don't really need speed when I can finish a transaction on a weekend. That's amazing! There's no 8 to 5, but we're not forcing people to work on that. Transactions are probably happening at somebody's house. 

So it's just amazing the opportunities that we have. So, as a student, it's great for me to know that a faculty member at the school can call up the conversational AI and say:"Hey, DRUID bot,I have to cancel class tonight.". And the bot say: "Well, would you like to cancel session one or session two?", "Well, session two.”; “For tonight or for the rest of the semester?"; "Just for tonight."; "Great. Anything else?"; "No.". And when it finishes that, that instructor is able to do that in a second. A note goes to the university saying: "Hey, that instructor canceled another class.". That's good or bad. A note goes to somebody in the building and says: "Could you go put a sign on my door?". Right? Because somebody is going to miss the text message that comes to them. Could also proceed for the text message that you got it and the back is send me back a note later saying every student in that section confirmed that they no class was canceled. No student is going to go out of their way to get the class tonight, be angry that you weren't there. 

You can't do that today. No, the student email is doing it, hoping it'll show up, but when the ones that do show up there were always the ones that had to do the extraordinary effort to get there, so they're the most disgruntled. Let's don't have them disgruntled. Please tell them.

Kieran Gilmurray: I like that. I like what you're describing there as well, because now all of a sudden, one of the things I've seen, Jim, over the last couple of years, is that, if we're being transparent, bots are often underutilized in the back office and front office, never in between. The AI can sometimes be projects or silos within different functions and departments. And, for me, there's nothing more natural than a conversation. And if you could put a conversational layer on top, that surfaced and dragged everything with it, then now we're starting to connect all the islands into one, whatever you want to call it, cognitive business application layer. You're starting to get the robots utilized a lot more, so now you're getting your value for money.

I love those examples. I love to see that happen. What other examples have you seen of conversational AI making a real impact in higher education? Or what examples would you like to put in place to get conversational AI to really work inside higher education universities, colleges or whatever? 

Jim Walker: Yeah, you know, what I'm starting to see is the schools are smart enough to realize that they're able to get off their old chatbot, right? So, let's just agree that a chatbot just for definition was dumb. You ask what time the gymnasium was open, it told you what time the gymnasium was open. That was better than not having it, but that's all I could tell you, right? When you made a phone call to the call system, you had to press "one" to get to this and "two" to get to that, and "three" or even, I don't know, I remember that "three" was what I wanted, so I wait all to go all the way through and go back through it. So, a university being able to say that their students call this number or text this number and ask when the gym is open, and the system will check to see what your geolocation is, and tell you the closest gym to the campus that you happened to be on.

So we have a client down in southern Georgia. There are two campuses in the same town. And so if I need to go to one of the gyms, I don't want to know the time that the gym is open in the other campus. And I could ask the question and it seems like, well, it shouldn't be that inconvenient when it comes back and says: "Well, campus A, the gym closes at seven" and I go: "Oh, I met Campus B" And that would just be smarter because if you called me on the phone, I said: "Jim, what time's the campus open?" I'd say: "Well, what campus you're talking about? B? Oh, is it 8 to 9", right? So this ability for the box to be smarter and this is just amazing that there is dumb today, is there ever going to be. So if you have a chatbot that only speaks 42 languages today,  next month it's going to speak 43. And if you're in the South in the US, today it may not understand what "fixing to go" means, but next year it knows that "fixing to go" is the same thing as "I'm getting ready to do something.". And so it interprets and understands and learns. So these are just going to get smarter. They're going to help me change my class. 

I went to my bank a couple of weeks ago, I was online with a little chatbot, and I was typing in questions and I got so frustrated with it, it would tell me it didn't understand my question. And like an idiot, I just kept retyping the question as if it was going to suddenly realize: "Oh, that's what he's talking about!". With a conversational AI, how about I have a trigger that says if the person actually does what you did and asked the question more than twice, suggest to him that we need to connect with a person. If it's not during the work hours, tell them: "Hey, I can't solve this for you, but I can make it take it for you.". But if it's during hours and there's a call centre: "Hey, I'm going to connect you to a person.". Because remember the old thing about just press the star key and it would connect you with a person, they've got rid of that. We're not letting that happen anymore. And so let's make that bot understand when he says the same thing twice that "I don't have the answer that you're looking for", just connect him with a person. The person should have more time on their hands now, because we have trained them up from answering those tier zero questions. 

So a student that can go in there and change their... pay their tuitions. So let’s say it's the beginning of the semester when you can register for next semester and you want to go register, you just got off of work at midnight, you're tired, but you know, you've got to do it to get your classes. And the first thing the system says is: "Well, well, you have some fines that you have to pay"; "What do you mean I have fines?" Well, if you have a chatbot, it would say: "You have a fine". It would just keep saying that. But if you could have the conversational AI go: "You have this fine and this fine"; "Well, how do I pay those fines?"; Well, if it's a smart system, if you use an RPA with it, go into the finance system, and you put your credit card in, and you can pay those fines. If you have to go down, it could say: "I could schedule an appointment."

But what you don't have is a student who already spent the day in class, cramming for an exam in the evening, went to work at 7 p.m., got off at 1 a.m., and at home, realized that their whole world's crashing on them cuz' they haven't registered and they get stuck because they got a little 400$ fine at the library. And now they're able to using their cognitive conversational AI and using RPA and all these other digital labours, they will solve that problem at 1:15 in the morning and head to bed.

Kieran Gilmurray: I like the sound of that. I sometimes wonder if it's all our fault that we've kind of spoiled conversational AI by writing really rubbish chatbots and by building chatbots that did what we wanted from the inside out, not the outside in. I'm a tremendous fan of a conversational design agent, but someone who can really understand what we are like as people and maybe not repeat what we're like as people. Maybe that's the opportunity to reinvent and redesign how we actually ask or get our customers to engage with us. 

Jim Walker: Well, part of it is we build what we can build when we build it. 

Kieran Gilmurray: That's true. 

Jim Walker: You know, and so it's a challenge these days, I think, to get people to do not call them a chatbot because that's a convenient little way to talk about them, but you do need to appreciate that a whole horse is pulling a wagon, is not a Ford F-150. They both have horsepower, but it's a completely different horsepower, and I would argue that this conversational AI and its ability to do work for me and to shift conversations when I shift conversations or on my phone system if I press “one”, I have to do that one of it. When it gets through, if the conversational AI says: "Jim, is there anything else?"; I'd say: "Oh yes, I'd also like to check on...so-and-so.". It says: "Great, let's go that way.". It's not: Well, let me get you back to zero, and then you press one again. So it is conversational. And I think what we're going to have to do is get really good at watching these conversations and making them as human-like as we can get them, but nobody's going to complain at two in the morning when they just got it we're able to register. No faculty member who's a single parent that has to take his daughter to the hospital real quick is going to be upset that they're calling into it "Bot", and having that bot cancel the class for them because they know it's important to cancel that class, but they also realize the most important thing is that get in the car to go to the doctors. 

Kieran Gilmurray: Really great uses of tech. Jim offered to put a business case together; that higher education institutes are commercial enterprises, but are they looking for a dollar figure to build out here? Because you've talked about students' experience and academics experience and a whole host of other things that don't feel like a hard dollar commercial, you know, business case you build what you're CFO. So, is the higher education business case different, or should it just be put on money, or is there a mix of business cases that you could sell conversational AI to some of the institutions who are listening in going, "Can I afford it?" or "Do I want it?"

Jim Walker: Yeah, so I have talked to several schools now, and early on with RPA, it was the CFO first and foremost. But in conversational AI, the person that's leading the conversations and the ones that I've been in is the VP for Student Affairs, who happens to be bringing in his or her partner, the CIO. And when the two of them are shaking their heads, "Hey, we need this.", they go lobby the CFO. And the reason I think that's important is, in the public sector, whether it's your state, your city, your county, your federal government, in the case of higher education, I don't think that the first thing they think about is ROI, 1$, right?

They have to service everybody. And as a result of having to service everybody, you don't get the choice of saying: "I'm Coca-Cola and I got 33% of the market appropriately good. Maybe we'll shoot to get 31%, but we're not going to be able to afford to get 40.". But as a government entity or as a public sector university, you have to service everybody that comes to you. And so, the throughput to get through the registration process, the ease of which you can change your classes, the ability for the faculty member to realize that one of their star students is starting to stagger because he or she is having home problems, and they never could work with that before because they didn't even know it, they didn't realize that the person was not only missing their class, but missing others. 

And the ability to take some of that back office work... I mean, you get small organizations that support that university all in the local town. Instead of having to have a person on the phone with might from the meat shop, the services, the cafeteria, they might can't just go in and check the status of anything we ordered.
Have you made my payment? And so the people that they have, that are servicing 4000 to 1 now, that the university is compared to what it used to be, they have the ability to do more important human things. And my personal example is the ATM, right? I honestly don't mind that when I was growing up and until I was probably 25 or almost 30, if I needed money, I didn't go to the bank.

And even today, the kids at McDonald's look at me strange when I pay them in cash because I just can't bring myself to put a $5 credit card charge when they leave off the phone if there's cash. And so the ability for us to start using all of this digital labour, to make it work for us, that was the promise of all of it before. Filling in forms and having a word processor to add stuff to a big Excel spreadsheet was what we needed technology for. We needed technology to give us the outputs that we could sit down and make the student experience easier and better for everyone across the campus, and the return on investment for that could save some money, but it could also avoid. The state of California avoided 2 million dollars in spending just because they did not mail out the annual renewal for driver's license. They realized that they could electronically send all of that out. 2 million dollars, they said: "Well, the state's got a 22 billion dollar deficit.". Okay. But some organization somewhere in the state of California could use that 2 million dollars that was saved, was avoided being spent, so let's just go ahead and acknowledge it. Avoiding spending money is as good as saving money.

And so, yes, you can get better throughput, you can get faster outcomes, you can do more surge work, you can do all of those kinds of things. Oh, and you can save money. 

Kieran Gilmurray: Where do I sign?

Jim Walker: Yeah, it is... And we're past the point where you need to put your toe in the water. Honestly, if you're talking about process mining, document understanding, RPA, or conversational AI, there's no need to say: "Hey, we need to do a little small thing real quick to see if it works.".

There are hundreds of examples of all of that working out there. You need to do something small on your network to just get used to, "Hey, how do we do the credentialing for this automation?"; "How do we get it on the network?" that would make sure it doesn’t cause a problem with denial of services and that kind of thing.

But don't start out with a 2$ savings. Find yourself what is our biggest student problem. Let's put conversational AI on that. Let students ask questions all of the time, and we'll learn from the students what they're interested in. It will improve those. Don't improve the one weird question that comes every three semesters, and pretty soon, through to self-teaching, the system will teach us what the students in the faculty members care about, and we will create not only more conversations, but we will automate the things we realized: "Hey, we create a conversation for that. Well, let's go one step further and automate the rest of that process.". And by putting in all of the here's questions and answers on a single platform, you can start saying: "Could we automate that? Yes, we could"; "Should we automate that? No, we shouldn't, because it's only ask once a quarter", you know?; "Should we automate that? Yes, we should, because it could save us half a million dollars, and we can move that half million dollars over to something that matters to the students."

Kieran Gilmurray: I love that. And at the same time, there's, what I describe, is real-world research happening there. So if that intelligence is connected to the system and the education institutes are learning from the students, learning from the partners as to what's truly important, we can't end today without getting an opinion on the impact of ChatGPT on the higher education space. 

Do you see it as integrating with existing AI technologies to improve student experiences, or do you see this as something else that sits by itself?

Jim Walker: Yeah, I, I personally think it's going to be part of the toolbox, right? But I've got a feeling that if you were in the early 1900s in the US, out in the West, and you still had a lot of horses and a blacksmith in town, the first time you saw an old rickety car drive by, you thought: "No one’s gonna use that.", right? or "Why would you want that when you can have a horse?". I think the same thing is there. We've seen a crazy adoption over the last year, couple of weeks of the whole AI thing. And we've been afraid of it for many, many years: "Oh, it's going to be biased. It's going to introduce bad things. Worse things in the world are going to happen. It's going to just be bad.". Now so they made it free, and everybody is an expert. 50 to 100 different uses come out for it, and different tools come out for it every day. 

Here's what I honestly believe. I'm a good fan and a fanatic fan of "Star Trek" and "Star Wars" and "Battlestar Galactica" and those types of movies. And what I've never done is I've never gone to a movie where it's all about the ship. It takes Star Trek; it can make your tea Earl Grey for you if they still have a bartender. And when they need that dilithium crystal pushed in, for some reason, it always has to be Scotty. It's never sent the bot down to it.

So I'm a firm believer we're all going to have a need to be there in the next hundred years. But what we're all going to have to do is adjust ourselves. But the tasks are going to change, the work may change, but you take that blacksmith, it was one of those in a town, you got an auto mechanic, you got a gas station attendant, you've got an upholstery person, you've got a tire person, you've got a new car salesperson. There are several jobs that have been created. When the blacksmith got displaced, not everybody could see that, they didn't want to give up that horse that they've been riding on for three years. I think the ChatGPT, of all the new technology, is going to have its place, and it's hard to envision right now that one does at all.

And so I'm all for sitting back and saying: We know that RPA works, we know the documentation, and it works, process mining works. We know there's great conversational AI out there with the DRUID platform and all that's out there now. Let's get the value from it now. And if, in three or four years, ChatGPT takes over the world, it will be smart enough to figure out how to take the best from that bot and start doing it. I don't have to retrain it. And if it's not, I'm still making value whether that's time saved, whether that's money saved, whether that's improvement in services. I'm making those today on the technology that was available yesterday. 

Kieran Gilmurray: I love all that. Jim, thank you so much for your time today. I'm going to remember a couple of different things, you know? Conversational AI and ChatGPT is not just another horse. That's the key piece here. And I do remember that Scotty was in every scene and still is. So even if we don't need conversational AI, RPA, and GPT will always need a Scotty in the business. So that's great ideas! And Jim, they say never meet your heroes, but I've read that NASA story so many times that I could probably repeat it, and I'm delighted I got to meet you today, and I'm delighted that the audience got to hear from you as well.

That is one of the most insightful and best interviews that I've done. I love your knowledge of all of the tech, and I love the practical way you built out the examples across higher education and business and everything else. I really appreciate it, and thank you so much indeed!

Jim Walker: No, I appreciate what you're doing, this... having a platform like you've gotten and spreading the word, I think, is an important thing that we all are going to use technology. Let's all talk about it with each other, and we truly democratize it when we do it that way.

Kieran Gilmurray: I love it!