#25 Freeing up people through conversational AI

KRISTINA PODNAR, HOST: All right, welcome to the Power of digital policy podcast. Today's guest is Justin Schmidt, who's the VP of marketing at Capacity. Capacity is an enterprise artificial intelligence SaaS company focused on helping teams do their best work. And that sounds awesome to me! Justin, welcome and thanks for taking the time to hang out today.

JUSTIN: Absolutely, thanks Kristina. Thanks for having me.

KRISTINA: Justin, let's jump in. Tell us a bit about what you're up to at Capacity. Your company genuinely believes the workplace can be better, making people productive, making sure that we have the type of information we need. I have colleagues who are envisioning the future of digital workplace using virtual reality to feel like we're sitting next to each other or maybe at a master class, and then I have other colleagues that dream about working opposite hours of most colleagues. So they never have to see each other in person again and having access to information. So, what is the future of work look like from your and your company perspective?

JUSTIN SCHMIDT, GUEST: Yeah. So this is this is a wonderful question at probably the most opportune inflection point in history to ask it, right, with everything going on in the pandemic and the great experiment sort of force distributed work, right, you read about people leaving San Francisco and New York and companies kind of waking up to the fact like, I don't really need to pay for 20,000 square feet in Soma, San Francisco, right? Like I can, we can work from anywhere. So I think the future of work is really going to be it's we're never going to completely get away from in-person contact and being in the same room together. That said, the distributed work environment, especially for an office job, white-collar, information type work, is definitely here to stay. I think your friends are both right. I am a little cold on virtual reality being, you know, I haven't seen the technology get to the point where we don't have giant masks goggles of covering her eyes yet, right, if it gets to the point where Warby Parker's making VR headsets that look stylish, then I can get a little more behind it. But that said, video conferencing is quick, easy, Slack, Microsoft Teams, quick communication, we take it for granted. It's so simple now, right, and then you have artificial intelligence coming on top of that, and this is where I think things can get a little interesting because you don't want A.I.to completely take over all human interaction. But a lot of the transactional stuff AI absolutely can, so give you a great example this if I were to say to an AI. Hey, can you pull XYZ report? It's a very transactional question. And that's also the kind of thing that artificial intelligence really excels at because the data is standardized and packaged in an easy way to where it doesn't take this massive amount of context kind of figure out what the query is, right? That's a very transactional type of interaction. I don't see those being the primary topic of conversation amongst humans in the workplace in the future. I see, and this is kind our dream at Capacity here. It's AI and tools and software helps the transactional stuff frees people up for the really high value, ROI important conversations. So instead of pulled me to report its, let's talk strategy about XYZ and come up with our go-forward point.

KRISTINA: So you've said, or I think this is actually a quote I pulled directly from Capacity, but one of the things that were interesting to me that I read on your site was while the upcoming transition into the future of work might prove to be a bit bumpy, it will still be decidedly people-powered which is what you're pointing to, and so I keep thinking about this, you know, if we're talking about still being people-powered and having some of this basic kind of background processes what are or are there like categories of things of work that simply can't or won't be replaced by AI and machine learning and robotic processing, you know, are there categories or things we should be thinking about?

JUSTIN: Yeah, great question. So the escalated customer service issues, I think, are always going to be people-focused. I just yesterday had; I will save the brand the tarnish, but I had an issue with a television and had to go to their customer service department is actually a very paradigmatic customer service experience. I went to the website, went to the troubleshooting that it suggested work got on a chat, chat with someone, they needed a picture of the issue. So then they send them an email then they called me back. Back, then we talked didn't quite get the resolution and help for, but nonetheless, I had that sort of classic escalated issue. AI could cover the first 75 percent of that in terms of I go to the website. The information that I need to do to myself troubleshooting is going to be a lot easier to find. I can pop on a chat, and you know quite frankly, as long as I get the right answer, whether I'm chatting to a person or a bot doesn't really make a difference, and I think the technology has come along far enough now where I think all of us when we get on website chat in the back of our mind wondering this person or not. And I mean ultimately, if you get the answer, you don't care, right? You just want to be able to talk to someone when you need to do so that the last 25% of the customer service experience, I think, is something that's always can be human. The other thing that I'm very confident in is AI in terms of like creativity and decision-making, and strategic thinking AI is an incredible good tool to augment that stuff but the we are quite far away from AI being the only thing driving that stuff. Like we've got a while before we turn them into the radio or load up Spotify and hear a song completely generated by artificial intelligence. That's emotional people, least I hope.

KRISTINA: Well, I was wondering about that actually so, you know, what's holding us back from that last mile and I'll tell you as a consumer. I guess I tend to be a little bit more forward-leaning in the sense that I want just more. I always want more, and one of the things, for example, that really bothers me, I'm not sure why bothers me so much, but it does which is my neighbor and I both use Amazon a lot like just crazy amounts of Amazon ordering and Amazon has this really awesome feature where they take a photo of your package on your doorstep, and a few weeks ago a package was delivered to my neighbor's house by accident. They took a photo of it it was uploaded, and I was annoyed because it shows that my package was delivered the images of her door and I'm wondering look, you know, you must have thousands of images now of her door versus my door when you place the package at the door, why doesn't artificial intelligence sort of run an algorithm and alert the driver that is the wrong door. You know, what's keeping us from like getting to that last mile. It's like that last 10%. Why can't we get that?

JUSTIN: Yeah. This is interesting. This is an interesting thought experiment. So think about all the last models we've already covered. So, Spotify, go back to them, Spotify is new music recommendations. I'm not sure you are a Spotify customer. But there they are really good, and Netflix is another example that's their recommendation. What you should watch is really good, and that is the result of machine learning being fed millions, millions of data points. It's getting to the point where it is. So but that was the last, you know, a quote Last-Mile not too long ago with something like what amp sounds doing as we think about how long they've been in the logistics business and it's been what like maybe a couple of years that we've come back just for them, and they're getting better at it. I have noticed just anecdotally that when I give I gave bad feedback once, I like the exact same issue happen; by the way, they've delivered to a neighbor's house, and it gave bad feedback on it, and it hasn't happened again. So part of me wonders if they are fighting as the human turn in a sort of temp worker type job that those that the delivery driver is right? Where maybe their systems are doing everything they can but she still like, you know, the person still just chuck the package onto the porch from the steps because they've got a million other houses to do and maybe their phone did alert them to the wrong spot, but they, you know just didn't notice it, right, the finer details of what AI can produce the focus and the sharpness of that is constantly evolving, and I don't think we're that far from that last mile being covered and then the new last mile is going to be we already see it where there's image recognition AI that's better at diagnosing cancer than real doctors like the next frontier that's a thing and beware the last mile is and then we're going to be eventually at the point where you know, we're going to come into work sit down. And as you go throughout the day, you're going to have artificial intelligence saying well, Kristina, Justin emailed about the schedule for the interview, but I already took care of it for you, right or it's three o'clock. You've got XYZ meeting coming up here is everything that was flaming regarding whatever is changed between last time, just sort of proactive and predictive artificial intelligence. This is an interesting question because it's always moving forward. So like your misuse this term a little bit, but the overton window kind of like what really consists of the last mile is this constantly moving.

KRISTINA: You actually touched upon an interesting aspect with temp workers and sort of the skill level of workers. And so I continually read that upskilling of workers needs to happen to meet the demands of the coming digital workplace. And you're talking about a lot of these processes may be automated to support us or more productive. But how do you see that upskilling taking place? Who's responsible for that? Is that the employer, the employee? Like, how does that work?

JUSTIN: Yeah, that's a great question. Personally, I don't feel that there needs to be a public service that helps with this because left to their own devices. The enterprise is going to be the profit motive is ultimately going to be what drives when that investment is made and investments going to be made a sort of at the last possible second. Just that's just the nature of the way things work. Right? So I think organizations that want to get ahead of it. I feel that like they need to get involved in that as early as possible as often as possible, but it's also incumbent and on the individual to figure that out, and I think about this just in my own line of work, right like marketing is this interesting combination of quantitative analysis with sort of qualitative thinking and creativity, but the quantitative analysis part of it like even just in the last few years, look at something like buying Facebook ads or Google AdWords. A lot of that is machine learning-driven from an optimization perspective and used to tell Google. Hey, this isn't a widget. I don't want to pay any more than best for sale. You think about it? Right and then, they send in sales for so I would like to see personally I would like to see some public effort to provide upskilling. I think there's a lot of really great programs out there with things like Udemy or you know, Khan Academy, Skillshare that kind of thing where people can start teaching themselves new skills, but with employers, I think they're if you're cognizant of what it is, you're trying to automate your cognizant of the roles that are going to change because of that in the investment you need to make to put those people into higher-order thinking and higher-order level positions, right? That's the whole ethos of helping you do your best work is when you have the simple stuff taken care of for you can focus on the more important tasks right in you just assume that out to the org level, and that's how you get your opportunity for upskilling.

KRISTINA: That makes me wonder about the educational space, and you mentioned public service and also some of these other private entities as ways to upskill, but I'm thinking about the educational sector. I have a 13-year-old, and I'm watching the school district struggle with a very basic like they can't format an email properly. It can't go content in; you know PDFs in a way to make it easy for people with disabilities to access. So we're talking about the really basic functions that have been around for like 20 years. Nothing really sophisticated, and yet I've read that you believe that educational institutions need to embrace an innovative and revolutionary approach to student learning. So is that happening? I mean, you know who's doing it? Right? Am I just missing who's doing it?  

JUSTIN: Well, yeah, this is a topic that's front and center for every parent in the United States right now, and I can say from my perch. It's interesting because, obviously, we sell into the education market, but my wife is an academic advisor at a State College in Illinois...

KRISTINA: ..front row seat..

JUSTIN: ..that is front row seat, and specifically, she's the academic advisor for student-athletes. So I'm watching her go through a lot of the student room boarding in new freshman orientation stuff. Not only digitally and virtually, but also for kids whose sports seasons have been canceled, right? So the who's doing it, right and what you can do to this is an interesting thing to unpack. The truth of the matter is, college students are, the students between 18 to 22 years old, of today, are kind of the ultimate combination of digital. They are the first real all-digital generation, right? Like they're born in 1992 or whatever it is. And the internet was a thing since they could remember they were on CompuServe or another bulletin board, AOL, whatever it was and we continually six of the four-year-old and for they are so the digital-first convenient access information for them is key in this is where I see students are college is winning being having messaging apps like a text message Saint Louis University here in said in St. Louis provided every student with an Alexa. And they have an Amazon skill, where all of the class information is accessible through Alexa. In their dorms, they can shout and Alexa and getting the right answers, and then the other way that colleges are going to our education's can win on be clearer. I think most of the winning right now in higher ed. So that's where that's where the dollars are to invest and this stuff, and it's going to matriculate into K12 eventually, but the other place that I see really fascinating things happening is using these technologies to help predict and manage summer melt or student dropout. So if you have said a student who's skipped X number of a particular class by y date. Well, we know that that's the leading indicator of drop out or failing that class so we can alert the student and we can alert the advisor the adviser can set up a time with the student help them out the teacher involved appropriately, and you see these companies like Blackboard and the other LMS is starting to adopt some of those principles. We have a product that does exactly that as well. So that's the that's where I kind of see this is how we can leverage the data that we have to offer just in time correction and information to students and faculty to help nudge these things in the right way in terms of like formatting email and sending a PDF. My son goes to a charter school. That's a non-profit Montessori school, and they are amazing. They are not technically adept, so I know exactly what you're feeling there, but I think it's going to be a while before that gets commonplace in the K-12 world.

KRISTINA: And that might be actually good because I'm thinking, you know, having it Alexa in every corner of every classroom can be really great. But also a little bit scary, especially from that ethics perspective. I know that as a company, you've published guiding principles around ethics, but what are the core principles and you know for most companies out there that are adopting things like mass constructs of information. And do you see those evolving like what should they be?

JUSTIN: That's a great question. And it's a really great question. So this is interesting. So we just recently went through our SOC 2 certification and as and as I'm sure you're very well aware of, and given your line of work like the amount of like policies that have to be sort of pedicured to get that as is not light and we have made of a very strong effort to. So if I give you an example from Capacity. At Capacity for information, we use various analytics components and stuff within the platform, right? And you know, we've been very cognizant about making sure that customer data is well protected that you know, our privacy policy is and all that stuff is well available and easy to find we also have taken all of our security recommendations and like during the sales process we make sure that we are very upfront with the prospect about what those are how we view these things. So I think for to your point the way these are going to evolve is especially in a world where like, I mean, it's kind of interesting like the in a distributed workforce if I'm on a Zoom in my dining room, which is where I'm working these days because my kids are going to use my office for a classroom. So I'm in I'm in the dining room, you know, my camera is on, and my home is in the background, right? So we've got this in this interesting world now, we're like everybody has basically seen coworkers houses when you previously that just that's not something happens all that often and I think just as long as companies are have very open and easy to find policies on like what data they keep what data they need and that that's easily accessible for employees to see that in that means the conversation about how those involved when you continue to change right, I haven't seen yet any malicious or nefarious data behavior with it within the enterprise within its own employees at least in at from AI perspective not to say that it hasn't really been, but I haven't seen it yet. And I think part of that at least in the enterprise; big business has traditionally been pretty good about it is, like somewhere within the legal entity organization those policies are written down somewhere. Maybe people don't read them when they just sort of like signed a license agreement or the car the contract or whatever without fully read it over. But all that stuff is usually written down, and I think for as we evolve in AI and then you have more of this sort of I don't want to say introduces a bad word. But as AI becomes more socially accepted as just part of something, we deal with that. We're just going to make me make sure that the policies and procedures we have in place on the data that collects understood by everybody who's are involved. How was kind of a rambling answer to a question, but it's like I'm it's I'm just thinking out loud here because it's such a fascinating topic? I'm actually curious like this is right in your in your wheelhouse. What where do you see this going?

KRISTINA: So, you know, you can't see me. I'm actually grinning here, and my heart's going pitter-patter because you've touched upon an area that always triggers me, right which is thinking about policies and having things written down. So when we have policies that are written, especially anything that's like longer than two pages. I call it shelfware, and I call it shelfware because what happens is usually you have, you know, pretty written paragraphs and you know, they go through with this crazy review internally legal compliance signs off on them and maybe you're one of those really large life sciences pharma company, so not only do they go through sign-offs in the workflow through SharePoint. They also give a physical signature on the cover page of the policy, and they get scanned and it gets put on SharePoint or your intranet, and it sits there, and nobody ever goes to read like, I mean, I've never and I mean I've been around the block a few times. I don't want to date myself. Here, but I've been in the industry for a while, and I've never ever had anybody say to me, Kristina, my biggest delight is going out there and reading 32-40-page documents on policies. And what we should we do like nobody gets jazz by this, right? And so I actually see AI and machine learning and chatbots at all of this new technology as an awesome way to empower people and deliver what I call guardrails, right, policies to me are guardrails. They keep us from crashing and doing bad things because they tell us what we should always do or never do, and I think there's an opportunity really to use different tools and different mechanisms to deliver that information. So I'll short you know, allow you to interview me on this, but I'll share one of my recent examples with you here. I actually worked with an organization where we built a chatbot for marketers, and yeah, it was really cool…

JUSTIN: Actually, I'm familiar with the practice.

KRISTINA: Yes, you are, exactly, that's what are the things they do at Capacity. So we actually built a chatbot to deliver guidance on what marketers should do or should never do when they're rolling up campaigns, and the idea was to stand up a global way of distributing information in context to what marketers were trying to do. So, for example, this is a pharma company. It used to take them like 21 days to get content approved, which, if it's a new product, is fine because products take years to get approved and get out the door in the pharma environment, so you have plenty of lead time. But you know if you have COVID or if you have the Ebola outbreak, you want to wait twenty-one days to get content out there to consumers. And so we basically created a mechanism or a framework in which marketers could ask things of the chatbot. Like I'm going to do a marketing campaign. It's related to it's not related to the product. It's in the EU. I'm collecting information. It is personally identifiable information or sensitive. Of information and I'm going to be using a website. I'm going to be too, so, you know, doing some social media and we're going to launch a mobile app. What do I need to be thinking about, right? Yep, and so to try it back basically to turn around and say like hey, you know, what? Here's a checklist, right? You need to actually worry about brand policies. And here's what we want you to do in terms of mobile. Here's what we want you to follow for the policies in terms of privacy and security and branding and navigation and SEO and handoffs from the mobile app so that consumers could actually call customer care if they needed to how that happened and how the information was captured. So when a consumer jumps from the mobile app to Consumer Care, they don't have to repeat everything yet again to a human being and so what's really great about this is we were able to deliver a lot of information on-demand in context because once somebody's logged into the network, we actually know who the marketer is where they're located we can tell if they have the budget for what they're doing or do we need to alert finance that they're going to need budgeting. We could also alert procurement that marketers are going to be engaging with a vendor. And so all of these requirements around their campaigns need to actually be integrated into a statement of work and so lots of really cool efficiencies. And so I see that as being sort of the next really really cool area and yes, you know, you have to worry about, you know, privacy protection and security and ethics and all of these things, but I think there's a much dare to say sexier way of delivering this in context of our daily work so that we are not going to go and read 35-40 pages of documents that nobody ever wants to read.

JUSTIN: Right, and that's exactly right. Yeah, and that's I mean the sort of human scale of ethics and AI you have things like making sure that you're not introducing bias and that you have diverse data and mature, you know, got guardrails are going to that's one thing but just even kind of what long lines what you're talking about. One of the things that we do we have a lot of mortgage companies in your customer roster in selling at the act of originating loan has a lot there's a lot of policies around than what you can can't do. You know what the limits are certain things are, and chatbots are perfect for this because if you're on the phone with the prospect, you're not going to sit there and start reading through the four hundred page document remain said that has all this stuff in and you can just you know, ask a chatbot. What's the USDA income elements, you know our business expenses considered when determining DTI all these inside baseball mortgage-related regulation questions, right? You can just sort of asking it's exactly the kind of what what you're talking about in the example you gave where You can provide just really quick, easy guardrails in the stuff. That's infinitely easier than pulling out a PDF going command F and like looking for whatever it is. What's really cool to me is where you'll see this evolved is the predictive like imagine being a phone call with a client you're chatting, and then you get a little notification says you guys are getting ready to start talking about income limits here in the current. In combat, right or the customer just said X they're likely going to ask about why later here's some information. You can just kill me. You can sort of in real-time very easily guide people through the guardrails of that stuff. But again, like get back to your original question like the data that makes that possible is important and sacred in a lot of ways, and when you have customer data included in that like you really need to have airtight controls over to who has access to that like the logs of the machine of the sort of decision engine kind of how it arrives certain things. You can debug. It's the fascinating opportunity for the industry to get this right and interesting to see how it evolves over the next few years, and it's rapidly evolving right and getting more amazing by the minute in terms of what you can do in the type of information you can either create or retrieve from this these kind of systems. I'm just excited to watch it, you know.

KRISTINA: Absolutely, and I think you're right. I think this is an opportunity to get it right, and I always talk about balancing risks and opportunities, but that's all really boils down to getting it right, and I think a lot of enterprises need to be thinking a bit more about that because you know, it's so easy in the day-to-day to get lost and not focus on the right things. So, you know, that's what it makes me kind of go back to this original productivity point that we were talking about, and it seems like we're really moving to a hyper-productive world which you know obviously is great for businesses, there's opportunities, there's risks, efficiencies, after all, translate into savings more opportunities. Is that all at the cost of worker mental health. Should we be worried about the ethics of making people more efficient and stressed out at all costs or how do you see that playing out in the industry right now?

JUSTIN: Yeah, you know 2020 is a very interesting year to ask somebody about mental burnout and mental health needs and stress but kidding aside, this is a great question. A good augmentation through AI should, if you're doing if you're if you're focusing more on high-level, creative, and strategic thinking and tasks and less on bookkeeping and you know distractions, then my hope is that productivity doesn't necessarily come with an increased pace that makes any sense. And therefore you can be, you know, the augmented information worker, but you're not going faster than what's biologically tenable right when you think about when I think about how technology is created stress and issues. The number one example is social media because it's it's just it just says a person on, you know in involved in human communication and her current day and age It's very clear that social media has outstripped on her biological Capacity to communicate properly with each other. Right and in the dopamine drip of getting likes and retweets, and you know communication at the cost of engagement as that were clearly as an issue. I think in the workplace, what's nice about work relative to just social lives or outside of work lives is that if an organization is run well, all the people involved share a common goal and you know, everyone has their sort of area of responsibility and lanes that they spend a lot of time and I don't go over to our engineers and bark at him for something that I don't know what I'm talking about, right but on social media, that kind of behavior seems to happen. So I don't think the AI-augmented and enhanced workforce is necessarily going to come with the same risk of mental damage that that's some other technology in our lives is created because if you plan it and you execute what you're going to allow your people to do is focus more on the things that they love, and that drives higher value. And I think we're just wired to really enjoy that a lot more than constantly being interrupted or dealing with the same issue for the 15 thousand time and a week or whatever.

KRISTINA: That makes me a much happier question.

JUSTIN: No, you should be like we should all be very lean forward and excited about what AI can bring to our jobs. It's like it's easy to be scared of it because of just you've seen any good science fiction, it's always, you know, AI becoming self-aware and taking over the world, right? Like I think we have a chance to make that in a case here and I'm with you it there's a lot to look forward to.

KRISTINA: On that note, you know, thinking about a lot of things to look forward to and be happier about being more efficient and more productive. You recently had a had given a talk re had actually hosted a webinar where you shared 21 productivity hacks on the way that we work. I'm kind of curious. We don't have time to review all 21, but you know, what are the top two or three that you want to see everybody who's listening today start using, like what are the two or three things we all should be doing right now?

JUSTIN: There's three out of those 21 that I think like is absolutely must-tries for everybody, two of them are going to be much more obvious, one of them you're going to be, but trust me on this. Okay? So the first one is I think everyone should make an effort to adopt the Eisenhower decision matrix at some point in their sort of self-optimization. So basically, the Eisenhower Matrix is he would get up every morning. It's sit down, and he would pick all of his tasks. And he would bucket in them and to do first do later delegate and eliminate, right? So you basically you're in any case if you imagine those on a two-axis matrix and the top right you're sort of rather your top left is sort of the most important most urgent things what you should do if it's urgent, but it's not important. We delegate it if it's not urgent. But it's important. You just make the decision right and what you're going to either do it tomorrow or whatever and then if it's not urgent and not important, just literally just don't bother with the tasks. So after you go through that exercise, you then have your sort of list of things that you're going to do today, right which phrase in my second tip, which is the Pomodoro Technique, which if those of you who maybe have never heard of this it's effectively a time management technique where you spend like 25 minutes of focused work on a particular task and then a five minute break another 25 minute run, five minute breaks, 25 run, final break and every two or three little sprint's you take a longer sort of 15-minute break. Pomodoro, because the person who invented it his timer that he used to do is 25 minutes print and the 5-minute break was a little Pomodoro tomato. So therefore so I'll get your task at the Eisenhower Matrix and then execute with the Pomodoro technique and then the third thing, which sounds weird, but I like everyone I tell this to I promise that if you do it, it will really help and that is on email whether you use Gmail or Outlook or a third-party app, learn all the keyboard shortcuts and process through your email without taking to the point where you can go through your email without taking your hands off the keyboard and what this does is it does a couple of things one it puts you in this mindset where like it's time to tackle email now, right? Like I have been here. I've got my hands on the keyboard. It's time to go and because you're never distracting yourself from liking, you know, go open up a browser and like, find the answer to the email. But then while you're doing that you notice some other thing and then the next thing, you know, you're on a different web page and then someone Slacks you or sends you a IM, and I are huge completely distracted you get through your email by just plowing through, you know archive and send next email reply. All these things have keyboard shortcuts, and you just get into this zone where you just kind of crank through your entire inbox and it's done and you can, you know, not until the next morning or later that afternoon whenever you schedule your email to them again to do you really even worry about it or think? It's like it's like this way to force yourself into a flow state on the task that I don't think anybody enjoys, which is getting through their email and it like it's like a life-changing thing for me.

KRISTINA: This is great. Well, Justin, thank you for giving us really great insights into productivity space the opportunities. He obviously brings for organization are tremendous, and the digital workplace will never be the same again, I think. We're going to evolve into a much more interesting, more productive space, and certainly looking forward to that. So thanks for sharing all of your insights today, appreciated.