Artificial Intelligence for Better University Operations

SU_AI_1600 x 695

The rise of artificial intelligence (AI) is now empowering universities to deliver a better student experience and more effectively manage their operations. I recently sat down with the Founder and CEO of Capacity, David Karandish, to discuss ways you can leverage AI to bring more flow to your work and ultimately help improve your organization.  



If you’d like, you can listen to our conversation in podcast form here:

What follows is a lightly-edited transcript of our conversation.

Ian Evenstar: From what I understand, Capacity AI is a platform, a solution, if you will, that can reduce the endless stream of support tickets, all of those annoying rote tasks, and ultimately help you deliver better services and products, is that correct?

David Karandish: That's right. So we would describe Capacity as a support automation platform, on the support side, we are here to answer as many repetitive questions as possible, on the automation side, we can take on some of the workflows that happen behind the scenes of higher ed, and then the last part, as a platform, you can build new solutions on top of Capacity, and so we're here to help just make the support process as easy as possible within education.

IE: I love that. but just to be clear, I guess, for anyone... any non-higher ed listener that might stumble upon this, this could also be applied to other business models, other organizations, other brands as well, correct?

DK: Yeah. So the platform is horizontal, and then we've got vertical integrations in the major categories we're in. Our big categories are financial services, educations, software, and healthcare.

IE: Excellent. And I know you have quite a background. You were one of the co-founders of, which I think was the number one Google search result for many years after Google launched, but aside from that illustrious background, what has motivated you at this stage in your career to start an AI company and really help solve some of the challenges facing higher ed today?

DK: So I got into internet advertising in the early 2000s, and started a comparison shopping engine, ended up buying, took it private, and then grew it up to a productive firm, but after that, one of the things I realized was that when you get into the early stage of a market that's just about to take off, there's so much fun to be had, there's a lot to learn and there's an opportunity to go make your mark, and so when I was thinking about what I wanted to do next after my last 10-year gig, I wanted to get into something where I could learn, something where I could grow, something where I was part of a market that was just starting to take off, and so the AI space checked all those boxes.

IE: So it's the AI space that you feel is a market ready to really explode or grow exponentially, not necessarily the EdTech space or the higher education space that's also, I guess, primed for growth?

DK: I think EdTech is also primed for growth, primed for major disruption, we can talk about that in a little bit, but I don't know of another market that's growing as fast as artificial intelligence. When we started the company just a couple of years ago, it was projected to go from a couple of billion dollars to $60 billion by 2025, I looked at it fairly recently and the same people, put on the same projections said, "Oh guys, sorry, we got that wrong, it's actually going to be $126 billion by 2025," and I would bet that if we looked at it a couple of years out, they'll be like, "Oh, sorry, again, it's actually $180 billion-plus by that timeframe." I think artificial intelligence is... it's not a page turn or chapter turn, it's like a throw out the book, rewrite the rules approach to how everyone is going to work and live, and I think the EdTech revolution's happening alongside it, is going to intersect very well.

IE: So let's talk about, I guess, the application or the use case of AI and what Capacity's unique, I guess, flavor of that is. How can Capacity AI or AI generally speaking, how can it actually help schools, universities, academies, how can it help those institutions really deliver high-quality service, or maybe improve the content and the product... the educational content that they're also delivering?

DK: When I think of AI, I like to first take a half step back and try my best to define intelligence itself. I'd say everyone intuitively knows what intelligence is, but it's actually difficult to define. My definition of intelligence would be that intelligence is the ability to recognize and apply a pattern, so you could have business intelligence, you could have mathematical intelligence, musical intelligence, vocal intelligence. There's a story about LeBron James, a few years back, where he was calling the plays of the other team because the guy on the other team wasn't standing in the right place in the middle of the play, so obviously he has a great basketball IQ.

But if intelligence is the ability to pick up a pattern and go apply that pattern, then artificial intelligence is the ability for software to do the same kind of thing. And so if we think about where this, specifically, can intersect with the world of EdTech, one is just answering the myriad of questions that students and parents and faculty have, anything from, "How do I apply for a scholarship," to, "I want to transfer this class," to, "When does the club soccer team meet?" All these types of questions are questions that typically would have to go to a person, in an organization like a university, oftentimes that is very decentralized, you don't really know who to go to, and so what we are doing with Capacity is we're saying, "Okay, time out, what if you could place an AI-powered bot on your website, it can answer your questions, and then any question it can't answer, it will intelligently escalate that up to the appropriate person based on the category that the question is sent?"

And so when we start, I'm going to say January 2017, we had about a 55% match rate, so about half the questions that you asked, Capacity could match to it, today, we're at just over 90%, so 9 out of 10 questions you ask to Capacity, we're going to give you an instant response to, and we have a 98% positive feedback rate in terms of folks liking the answers that come back. And so our whole idea is that the first place we can go make a big dent in improving the educational experience is in being able to answer student, parent, and faculty member questions.

The second area is that if you think about all of the workflows that happen behind the scenes, within a school, anything from transferring a student in, to setting up all of the effective internal marketing campaigns around registration, classes, and all the things that happened there, all the way up to, "How do we help our most at-risk students go achieve their educational goals? How do we figure out who is in that bucket?" So we've got some workflows that we've done across a couple of our different schools, where we're doing things like automating the onboarding experience, helping student advisors and/or life coaches do a better job of targeting the students that they want to go reach out to. So very high level, you could think of Capacity as answering as many questions as we can, automating as many workflows as possible, and then ultimately giving you the tools to go build whatever new solutions you'd like on top of the platform.

IE: That's great. And I love the fact that you simplified, I guess, the two main primary buckets, one being around workflows and automating that, and then the other around Q and A or intake in terms of questions or prompts that are coming into an organization or an office, so we'll dig into each of those, I think, individually.

Coincidentally, I was just on a call this morning with what is referred to as a restart team, and this is a dean's cabinet, so senior administration level at a major university talking of about the number of times they have to send the same email out, manually draft and send the same email out because it's this pattern that you mentioned in terms of intelligence, this pattern of questions that continue to come in at different times of the year, different semesters, different peaks, ebbs and flows, and if we could solve that, if we could add Capacity... now I see where the name comes from, if we add Capacity to the areas that are actually more meaningful work and decrease the burden of this Q and A process, that's a big win across the board. I guess I'm curious how this might differ from maybe a knowledge base or an FAQ on your website, how is this different than search?

DK: So search has its place, and we're big fans of search, but there are a couple of major limitations to the way most knowledge bases and most search platforms are set up. Issue number one is that people tend to think of knowledge base as information goes in and it never really comes out, so it wasn't designed to have an expiration date, it goes in and stuff gets stale and it doesn't get cleaned up well. So one of our guiding principles is, when we built out our knowledge base, we said, every goal question should be able to have a start and end date, so you can tee up content that will kick in in the new semester, and you can also have content that automatically expires when that time is over, so that's just general-purpose hygiene that we like to start with.

DK: The second thing though is that search does a really bad job when you've got contextual answers, when you've got follow-up questions, you need to ask to clarify or to help winnow through the various options. You might have a very different answer if you're a part-time student or a full-time student, or depending on... if you're on, say maybe you're in the military or maybe you're an adult, undergraduate learner, is very different than on-campus... adult online versus campus in-person, so all of these types of contextual pieces are fed into what we would call guided conversation, which makes up a good chunk of the experience that you have on the Q&A side. And then lastly, what I would say is search typically ends up restricting to a set of knowledge within that search system, so you have a knowledge base, we can search the knowledge base, with Capacity, you can think of it as a system of engagement that sits across all of your key systems, so we can search the knowledge base and website and Canvas and Salesforce and pick whichever systems of record you end up using.

IE: So there's a higher level of engagement, the information and the access to the information remains current, so there's a hygiene thing that you mentioned that is part of this, but you're all also inferring that it doesn't necessarily replace search, it maybe, ties together all of the other platforms where knowledge or information exists and then layers in or works, I guess, in a more collaborative way with other processes that you might already have in place.

DK: Yeah, let it say it another way, search is a component of the experience, but if it's just search, then you're missing out on a richer, more engaging experience that you can have. Another example that we didn't talk about is how if you do a search and you don't find your answer, usually pretty SLL, with Capacity, if you ask your question, Capacity doesn't know the answer, you can immediate thumb that down and we can bring someone in via live chat handoff and take you directly into a conversation with the human, so it's just a graceful escalation from level zero support up to level one.

Alternatively, if you don't have someone available on a live chat, we can still go create a ticket in our help desk, but when someone on that end goes in and respond to that, they're not just responding to that ticket once, every response gets stored back in the knowledge base, so the next time anyone else in the school has that question, it doesn't have to go back into somebody's email inbox, it's now instantly responded to. So it's this learning component and the escalation component that differentiates what we're doing from just pure-play search.

IE: I see. Yeah, and that was going to be a follow-on question, is how this might support or differentiate from universities and schools that are using now a student ambassador to be on the other end of a chatbot, and I guess in a way it could escalate to point where you're actually engaging with a real-life human, but do you think that this is superior in some way than having, say, a group of student ambassadors that are just on standby to chat and engage with prospective students?

DK: So we're not trying to replace student ambassadors, I think there will always be a place for someone to want to talk with a real person, but there is a subset of use cases, whether it's the time of day, whether it's the simplicity of the question, whether it's someone who's on mobile, maybe they don't have time to go do a full live chat, there's always going to be a scenario where all you want is just a quick answer to your question, where bringing someone and getting them involved, it's just not the ideal experience. At the same time, we don't think of it as either/or, we think of it as both/and, you want to have your great human experiences for things that humans are good at, and you want the bot to be there to answer as many questions instantly as possible, so it's really both/and not an either/or.

IE: Yeah, that's good. I was grateful to hear that response because I don't think people are ready to throw out their student ambassador initiatives or that tactic altogether, it really helps with nurturing, engagement, conversion when you have those advocates really speaking on behalf of the school's brand. Does the bot itself... I'm just curious, does it feel like you're talking to someone else? You mentioned that intelligence is being able to recognize a pattern, and if you are sophisticated enough at recognizing those patterns and the inflection in those patterns, then it starts to feel like empathy, it starts to feel like a human interaction, does it feel human?

DK: It feels as human as anyone's technology is today in this space, and obviously there are limitations, but that said, we want to actually call out that you're talking to a bot, we don't want you to think you're talking with a human and you're not talking with a human. But we do a lot of human-like things, so if we don't know the answer, we'll tell you, "Hey, I don't know the answer to that, let me get someone else involved." If we think we know the question, but we're not quite sure, the bot will actually come back and clarify for you and say, "Did you mean this, this or this?" And when you make a selection, it'll go back and remember that for next time. We recommend the bot being as colloquial as your university would normally communicate, so if you're on maybe a friendly large state school, it should reflect your values, and if you're a small liberal arts college with very much better grammar than I have, then it should reflect that. So you want the bot to reflect the personality of the institution itself.

IE: That's great. So in a way, you could think of the way that you're branding your school, the way that your school is positioned in terms of mission, vision, values, brand persona, brand personality, you can actually have Capacity as an extension of that brand position, if you will?

DK: Yeah, it's almost a digital brand ambassador.

IE: That's great. Sign me up. it sounds like a perfect addition to add to any major college or university. Does it take a long time to onboard and get it installed, and connect your Salesforce to... I don't know, maybe it's a SharePoint-Salesforce integration that you're working on?

DK: And so if we sign with the university today, we can get Capacity in your hands, testing, answering questions within 30 days or less, so it's not this big six-month rollout sort of thing. Now we've got over 85 different applications that we've integrated with, anything from SharePoint to Salesforce to you name it. If there's an app we haven't connected to, we have a developer platform, and either we can go build that integration out, or if you've got developers to build that out, that works too. So we might not have that on day 30, but within the knowledge base and the apps that we work with, we try to get this rolled out again, first 30 days, 45 days max.

IE: And I'm assuming this lives on the front end of the experience on a website typically, right?

DK: It's literally one little line of JavaScript you can go put on your website, your mobile app, it's responsive, so if you have a responsive design site, it can adjust for that, but we can also plug in... for faculty members, we have a plugin for Slack, we've got a plugin for Microsoft Teams, we are working on an SMS version, so you can text the bot and it'll text you back. So multiple modalities, we would call those interfaces, wherever you are, the bot could be.

Like what you're reading? Sign up for the Higher Education News Brief, and receive the top stories from across the landscape of Higher Education every Monday.

Yeah, let's speak a little bit on the faculty need and the faculty's experience. It's pretty clear that from an administrative standpoint, this could be really helpful in terms of recruitment, so think about recruitment and advancing the endowment of a university, that could also play a role, but for the faculty member who may, I don't know, let's say have a long archive or a large library of knowledge that they're trying to include in their course, is this something that it would actually aid in the educational experience as well?

DK: Yeah, so we've had Capacity be a TA helper for various classes, so you can have it answer questions on your behalf. You could have Capacity take on some of the workflows and processes behind the scenes. So we were working with one university where if you wanted to drop a class, it took about nine weeks to process, which is a long time, but that was for their campus environment, for their online environment, it also took nine weeks to a drop class, but their online classes were only eight weeks in length, and so you live would not know if your class was successfully dropped until the class was over, which just did not-

And so we're taking processes like these and helping to just make it as efficient as possible, by coordinating between all the people that need to be coordinated and then taking the people out who don't need to be sending emails and then checking inboxes for this sort of thing.

IE: Wow, it sounds like a promised land, like a Mecca, if you will, an Oasis of just staying in your flow whether you're a faculty or an administrator, staying in that workflow of where you really add a lot of value to your day to day and removing more of that redundancy that you see, that can be automated and also built into this AI. I'm thinking about the fact that you might have the chance to differentiate from other universities, do you feel that Capacity in some ways is actually helping universities reposition themselves or differentiate from their competition as well?

DK: Oh, absolutely. We have universities that use us extensively on their recruiting, where they're like, "Look, we have a digital bond that helps us make this place more efficient, more effective." We've had some organizations we work with that haven't hired Capacity out of their tech budget, they actually hired Capacity out of their hiring team budget, and they said, "Hey, this is like a new team member we've brought on," so a differentiating position of bringing this technology in, is something that we definitely help universities have the benefit for.

IE: Absolutely. And in your experience so far with, I guess, onboarding and recruiting new universities to sign on, do you need to be a certain shape and size? Is it made for, as you mentioned, that small liberal arts college, or do you need to be a major player in the university industry?

DK: No, and we've worked with everything from top 20 undergrad programs to one of the fastest-growing community colleges, to the second fastest-growing hybrid university in the United States. We've got a good mix of sizes. We've also done work going down even outside of higher ed, like going into things like school districts. We've got a school district bot with one of our clients where we're answering K-12 questions for parents and faculty members. We're also working with a couple of the OPMs in terms of the folks that power the experiences for these online programs. So we've got a good mix of different schools and higher ed adjacent organizations.

IE: Yeah, that makes sense, and that also makes me think, does this fit within a toolset for an agency, or are you more like B2B going directly to the university and to those key decision-makers, or do you find that you're leveraging the agency-university or the OPM-university relationship as a way to move this tool into that ecosystem?

DK: Honestly, we're finding opportunities on both ends, so wherever we can get the fastest introduction to the appropriate folks is a good place for us, both directions are good.

IE: I just wanted to come back to workflows because you mentioned that earlier, what are some other use cases in terms of workflows? Adding and dropping a class, I think you also mentioned maybe registration or possibly some onboarding experience, whether it's on the faculty side or maybe a new student coming into a new school or department, what are some other workflows that you've seen are really good cases where you could bring this in and add Capacity?

DK: Yeah, a good example... I'll double click on the onboarding piece, one of the universities we work with, they recognize that if they could get a bigger proportion of their online undergrad students to get all the way through their orientation, those students tend to have better academic outcomes down the line, and so what you find is that sending them an email the day before might not be the best way of engaging with some of these students. And again, different populations, you can't just all go down to the quad and play hacky sack and tell people when the next orientation meeting is, these are working adults, these are folks who are trying to go back and get their undergrad degree, sometimes at 8:00, 9:00, 10:00 at night, and so what we did is we helped build a working workflow that would send an automated response, all the way through an SMS, to these students and then track each week who's gone through which parts of the orientation, and if you've gone through it, great, and if not, we can send you other followup marketing messages to get you through that process.

We've had phenomenal success with just that one particular workflow, we're now using a lot of the same underlying ingredients that would be used for the onboarding workflow to go help some of the life coaches, where they're like, "Hey, we want to know... We've got X many thousand students that we need to think about, there might be a smaller subset that are most at risk, most likely to drop out, most likely to do not get the academic success that they were hoping for, let's figure out how to separate the chaff from the wheat and then send the right message."

So if Susie is an A student, gets a C- in physics, maybe it was just a hard physics class or something like that, but you got Bob over here who's got a 2.9 GPA, his grades are dropping, he might have gotten the same grade, but man, his overall GPA... he's got C- but he's got two Ds and an F, and all of a sudden, you're going to have a very different level of intervention. Honestly, it's very analogous to that personalized medicine concept that we're starting to drive toward, where we're being more interventional rather than reactive.

IE: And you mentioned, we could think of this as a digital TA, but in a way, it's a digital guidance counselor or career advisor, something that is really going to bat for the student in identifying some of those risks, so we could play a more preventive role or in some ways maybe empower the student by coaching them through the places where they're stuck. I really love that because I think that can help not only with student retention, but help and support learning outcomes and career placement and graduation rates, this AI from what it feels like, Capacity has its "capacity" to really help improve the standards across the board, and if you believe that education is the means of ascension for society, dare I say that Capacity is helping society ascend. Can we go that bold?

DK: I'd go that bold. Absolutely.

IE: You mentioned this business strategy called variable elimination, what is that? Can you tell us about that?

DK: I think when you start an organization, it's very different starting a company than your typical educational experience. And so I was good at math, I like math, but at math, you usually know what your variables are, and you're starting out with a fixed number of variables, when you go to start an organization or start a company, sometimes you don't even know what the variables are, and even if you do know what the variables are, you don't know which ones are important and which ones aren't important initially, and so one of the most important salient traits for entrepreneurs is the ability to go in and eliminate all of the excess variables that don't matter and focus on, what are the two or three or four things that are going to be most impactful for you hitting your goals?

And I think that same entrepreneurial mindset that you see great startups employing, I think that applies out really to any great IT project, any major change that someone's starting to roll out, it's just so much harder with a startup because literally, no one's going to tell you what you need to figure out when you're starting a company.

IE: Yeah, and so if we apply that to higher ed, I've been in that situation where a school is trying to launch a new degree program and they may have a preconceived idea of who their target audience is, or who's going to respond really well to this new degree offering, let's say it's a new master's in science, so could this work to our favor to help eliminate some of those hypotheses we make about our target audience, or maybe our hypotheses or the variables that we're introducing anytime we launch a new program at a school? I know it's a bit of a reach from what you just described, but in a way, does that relate at all or not?

DK: Well, I think whenever we demo our technology and people give us their feedback, one of the things they oftentimes miss, and then when we explain it, they're like, "Oh, that makes total sense," is that not only is Capacity here to answer questions for students and parents and faculty, but Capacity is also here to help identify what are the questions that people are asking because they weren't able to get an answer in the other mediums. So school goes through a new website redesign, and everyone's all excited about it, put Capacity on there and see which questions did we not do a good job of servicing, which areas did we bury in the menu that we really should have risen up to the top, otherwise somebody could have found that information.

We had one organization do a big web redesign and they missed one section that they thought they were going to port over, and somehow between the big... coordinating with the outside firm and the inside firm, our inside people who are working on it, they didn't have it, and the next thing you know Capacity was up and they started to get all these questions on a section that didn't get ported over, and they wouldn't have found that if they just looked at their Google analytics or just looked at some of their logs. And so these are the types of things where Capacity can help not only answer the questions but give you that key insights and intelligence around, "How do we create a better experience?"

IE: Yeah, well, it comes, again, back to this notion of empathy and really understanding we can try to predict through maybe keyword, searches or Google analytics, the types of questions people are asking to find content on your site, we can also try to predict or assume or synthesize traffic patterns throughout a website, "Oh, these are the best traffic pages, for example, so we should maybe look raise them in the primary navigation, raise their visibility," but it's actually validating that because you're taking in real data from the human themselves and in a way, using your design to meet that empathy where it's at, matching a redesign or matching a new initiative or matching your brand messaging to the types of questions and the things that Capacity is pulling in for you.

DK: Absolutely. The whole idea is that you just want to remove the friction from the experience and also remove the friction from the experience of the people on the other end trying to understand why people aren't getting the information that they need.

IE: Wow. So do you want to talk price at all, or should we just go to the website to figure out the price and put in a demo question? I'm curious if a university or a school is going through a big redesign, how much does it cost to plug this in?

DK: Yeah, the way our pricing model works is that we've got a bunch of different parts of the product that you can use, so it's one price if you use just the knowledge base and the help desk, but another the price, if you add workflows, you've got conversations with the app integrations, or we've got an upcoming release that we're doing around some of our database work, and so we can work within just about any school budget based on the scope that they want to try to roll out. But check us out, go to capacity com, I'd be happy to talk with you.

IE: Now what's on the horizon? Obviously, you're not going to stop with this particular set of tools and you're going to add more as we continue to go, but what are some of the trends that you're seeing in the EdTech space or within the AI space? Where do you see evolving? What's next?

DK: So if you look at the history of education, you end up getting more and more personal over time. So my kids love to watch these reruns of Little House on the Prairie, and you watch these old shows, still in old grainy colors, I think they find that to be funny in a Netflix generation, but one of the things you look at, you'll see the old, one-room schoolhouse, and that was amazingly advanced for that time rather than people having to learn at home and be at the best of how good or bad their parents were as teachers, and then you see that model move toward more of the classroom model where you could try to target content more at one group's individual level, but we've seen a lot of cracks in that model now, where you might have an eighth-grade math brain and a fourth reading level, what grade should that person be in?

If you ask me to jump on my DeLorean and go into the future and see what is education going to be like, I think the concept of fixed grades is going to be viewed very much as an anachronism going forward, where it's like, "Wow, that feels like something that should have happened out of a different time." So I think if you... the whole thing is you're moving toward more and more personalized learning. Look at the rise of places like, it's the largest online tutoring marketplace, happy to be an investor there and was on the board for many years, they grew out of this desire for people to get one-on-one, direct education, and so I see that personalization need applying within the classroom, but also applying outside of the classroom. Anything from applying for scholarships, applying for college, to automating the workflows behind the scenes, I see the future of education being so personalized, it will make today's education feel as archaic as dialing up would today.

IE: Yeah, we need to move kids off of the conveyor belt and into their own personalized roadmap, into their own personalized mission, if you will. I could see that also applying to the way that we structure majors, and the way, possibly, we have this hierarchy between bachelor's, master's, and postdoc work, so it's quite possible that you're declaring a mission or a concentration or a domain of interest, and then just pursuing knowledge within that domain at your own pace, it'll probably come back to credentialing and making sure that employers or I guess society at large can say, "Oh yeah, well, they have an equal amount of education here." So having some standards, I guess, in place would be still probably appropriate, but the idea of personalization, I think, is one that I haven't heard. A lot of people are talking about the disruption of education and disruption with intent or interest around personalization, I think that's a very compelling idea.

DK: Yeah, I think a tale of two online education experiences here, so my kids were home during part of the pandemic as many kids were, and they hated Zoom at home. Zoom is a four-letter word to my eldest daughter, and I mean that in every way possible, conversely, they love Duolingo, and they use Duolingo as an adaptive learning platform, it doesn't feel like they're in Spanish class, they're having fun, and so I think we saw the limits of one giant class for all people, all day, for all time, and we're seeing the benefits of adaptive learning, back to what we were seeing over at Nerdy. I got really excited when they rolled out their SAT and ACT prep, where instead of having to study for the entire thing, you go and you take an adaptive learning assessment, and it's going to personalize the specific areas you need to go focus on and thus save you a ton of time, I think you're going to see that pattern or that concept apply out across all of education.

IE: Anything to maybe leave the listeners with? I know you've shared a lot and many people will have to go back and refer to the notes and see exactly how they're going to pick this up and apply this to their organization, but any parting words or any continued advice that you might want to offer our listeners?

DK: Now is the time to get started. You don't have to clean up before you take the bath, you don't have to have everything ready and perfect, now's the time to go bring this technology into your organization.



Let's Work Together

Looking to use AI to improve your operations? Let's take things to the next level.



Older Post The Case for Digital Credentials (Pt. 2) Newer Post Three Strategies to Buck the Trend in College Enrollment

Higher Ed News Brief

Sign up to get the top headlines in higher education every week.