The First Years Are to Prove You Can Exist, Then You Can Grow until You're Hard to Kill

David Karandish, CEO of Capacity, explains how AI automates customer support across 70+ apps. He breaks down the startup lifecycle: first prove you can exist, then grow until you're hard to kill. Founded January 2017, Capacity scaled fast by nailing hiring and onboarding early.
Questions this video answers
January 2017.
About 70 employees.
Five or six people.
Over $6 million (annual revenue last year).
87% (with the remaining 13% escalating to a human).
Welcome to this episode of the Million Dollar Mastermind. I'm Larry Weddell. And before we get started, if you want to know exactly how to win again and again, go to Wydell on winning.com webinar now to watch something I've put together for you. Now let's get going into this episode of Million Dollar Mastermind.
I am here with David Karandish from St. Louis and he is the CEO of Capacity and is a new kind of help desk powered by artificial intelligence, automating support for customers and employees. I'm going to find out how long that company's been going, but they were up over 6 million and annual revenue last year, named America's best startup by Forbes in 2020. And so lots of great things happening over there. David, tell me about this capacity and why you're excited about it and why it solves problems that if you're starting a business, you're going to need solved.
Yeah. So every business has support needs, whether it's internal support, like IT and hr, whether it's external support for sales and customer success. And so when we started Capacity just a few years ago, we recognized that there are a lot of people in the mix within your support organization. And we thought, what could it be like if instead of having to ask a person every time you wanted to update your vacation balance or you wanted to return a product, or you wanted to find the hours or any question you have internal or external, what if instead of having to talk to the person, what if we could automate that support function? So we started it just a couple years ago, initially internally focused on automating internal support. Then we started doing external support. And then over time, we're now plugged into over 70 different applications. We're working with some of the biggest financial services firms, work with large universities, working with some big healthcare practices as well. But our whole idea is that we can use artificial intelligence to understand what you mean by what you say and to build out this support automation platform where we currently answer 87% of the questions that get asked and then the other 13% gracefully escalate up to someone on your team.
Yeah, that's the idea of, you know, that's the advantage of a system and all. I have this thing about great generals, and the great generals all have great systems and they usually have more systems than most people, which means if a given situation comes up, they've got a game plan for dealing with it. They know click, click, click, and they've thought about it. And as a result, they can handle more things. But what if your system was automated? And as you, you know, there's no escaping it, folks. You start a business, you're going to be a one man paper hanger. You got to do everything yourself because you're starting and you only understand. But the more energy and you put into collapsing timeframes where you can get some success, add on some staff, then you can start to get a picture of what are the repetitive functions? What are the repetitive, what are the washing the clothes and folding the towels type activities involved that are always going to be there and how can I train someone to do that? How can I, you know, what kind of questions are going to come in on over the phone and how can I automate that, get a system for that, to offload that off of me so I could focus on the more unique things? You know, it's crazy for, you know, it's just like David, we used to say in the early, early days of our company. I'm talking about 70, you know, the 70s, early 80s is like, why, why spend your time on $5 an hour jobs? Now those jobs are $15 an hour. Why spend your time on $5 anhour jobs if you want to make $500 an hour? You know, and today if you want to be somebody who is your time is valuable. You got to look at what you're spending your time on. And the only way the things have to get done by somebody or something. And so you've taken that a step from figuring it out how to put a system in place in your office with your people with training manuals and things like that. And you've taken it to artificial intelligence. So automatically now you're far beyond me and figuring that out. So I would have to come to you. And now I can come to you. So talk about how an example of how that would work.
Yeah. So a good example would be you have your website, your website gets questions and right now people are emailing you these questions. What are your hours? What are your policies? How can I find this product or that product or another example. We just hired some new people and they have all sorts of internal questions. From the mundane, what's the office door code? To the where should I park? To the how do we enter new leads into salesforce questions.
Right.
Imagine if all of those questions could go into a knowledge base that uses artificial intelligence so you don't have to type it in. Exactly right. And you can do all the matching and all the searching and you get an instant response. And even in the times where it doesn't know the answer, it will intelligently route it to the right person on the team that can get you the answer that you need. We think that that's not some pipe dream, futuristic scenario for 20 or 30 years from now. We think that, we know we're seeing this, that our clients, this is fundamentally changing the way people do support. And so, you know, we're not just here to save companies money, although we can do that. We do do that. We're here to free up your team so your team can focus on higher order, higher dollar value, more important tasks, and then let computers do what computers do well.
And at what level should someone consider getting involved with something like this? How big do they have to be?
From an internal side, usually kind of tips at around 100 team members. From an external side, if you have customers to support, we can help you out no matter what your size.
And how big is that market? In the United States alone in North America, of companies 100 employees and up.
I mean, there's, there are thousands, thousands of companies that fit that description. When we looked at the AI software market, it was anticipated to go from a couple billion dollars when we started to 60 billion by 2025. Then I looked at this about, about 18 months ago. They said, actually, guys, sorry, we got the projections wrong. We said it was going to be 60 billion, it's going to be 126 billion by 2025. And I'm going to go on this podcast and say, I think they're going to turn around in two years from now. They'll say it's 180 billion, 2025. Because there's so much opportunity to create better both internal and external experiences. Companies are just pouring money into this space.
Yeah. And how did you. I mean, how, how big of a company do you have? Because people are listening that have companies that are growing companies, they're wanting to scale up how many. It's helpful to know how many people are involved in a business that has this kind of outreach. So how many employees are you running now?
We've got about 70.
And how long has this company been going?
So we started in January of 2017, and we've basically been tripling the business. I mean, we started out very, very small with just a few, just a few people. But we have been basically tripling the business every single year since then.
And so you started 2017. How many people did you start with?
We had five or six.
And now you've got.
70.
Talk about how you went through that.
Process, hey, listen, there's a lot of information online, but there aren't a lot of people who have actually done something. In my case, I've actually built a successful business, has accrued over $5 billion in assets under management, and has done well even during trying times. Now, if you want to know exactly how I've done this, go to whiteellawinning.com webinar Now I've compressed a decade of learning into five short weeks. Just for those of you who want to give yourself an incredible advantage and are tired of waiting and watching others move up.
You know, one of the things that I'm a, I'm a big believer in, it's that startups are not like most companies. And what I mean by that is that in a startup you're, you are trying to prove that you should exist. Google doesn't have to prove that it exists. Microsoft doesn't have to prove that it exists. Apple doesn't have to prove that it exists. You launch a tech startup, nobody is going to hand you their time, their attention, their dollars, their subscriptions. You have to go prove that you should exist in the marketplace. And so those first early years are all about going from getting that first customer to getting the requisite size where you pick up the patterns that exist in the marketplace. Then once you get to where you've proven you should exist, the next big hurdle is getting to where you're hard to kill, where it doesn't mean you're guaranteed to be the next salesforce.com, but you get to a certain size. And particularly with the subscription business model on, it's really hard to, really hard to unseat you. And so those are the stages you get your first customer preview should exist. You get to hard to kill and then eventually you get to a size where you could start to either go public or do some buy side acquisitions and go through that whole process. And so in that initial phase, all we were trying to do is look for or separating the signal from the noise in terms of what's working with these customers. How can we continue to improve their support? What can we do to build the best frictionless support experience out there?
Now talk about how that actually worked out in the real world in your situation over the last three or four years.
Yeah, so for us, it's first started out by finding great people. For us, we really believe that not everybody is a good fit for a startup. And that doesn't mean that you're a bad person or you're not a good Fit for other types of work. But we really want people who can do well with ambiguity, people who can work independently. And in a team, we really like team members who have a proactivity to them because, you know, in a big company, something doesn't get done, you know, you're probably not going to sink that company's quarter here. In a startup like every single person counts. So we really like people who are proactive and I in particular like people who have a high talent to low ego ratio. I know, would you say that again? High talent to low ego ratio. Okay. I know people that are high talent, high ego, they don't tend to play well in the sandbox. I know people that are low ego, low talent, they don't tend to get things done. Occasionally you'll find high ego, low talent, that's the worst, that's the worst they exist combination. But it's the high talent to low ego ratio where you know, you got team players, folks who are willing to go the extra mile, but you also have, you also have a level of competence and willing to figure things out. No one's going to hand you the exact playbook on how to make your startup grow. But if you can work with people that are willing to try things, willing to do the experimentation it takes to get that product market fit, then you're in great shape.
Now how do you go about doing that?
So we screen all of our team members on the way in. So we put people through standard personality type tests, we do multiple rounds of interviewing and we try to ask questions, assessing how they do in ambiguous situations. How do you deal with a lack of requirements? How do you deal with problem selection? We use that. We also use another kind of heuristic. So a buddy of mine came up with this, Andy Ketter. We've adapted it. He said, david, in a startup you got really four types of people. You've got your prospectors that have just a natural, intuitive sense of where the gold is on the land. And those folks, you don't need a whole team of prospectors, but you need enough folks to be able to understand where the opportunity is. Then you've got folks who are effectively the map makers. So they don't, they don't have that intuitive sense of the land. But if you show them where, where the gold is, they will build you the best and most efficient route to get there. Then you got the guides. And your guides can't build a map. They can't prospect. But if you hand them the map, they will execute on it. They will take people in and out all day long. And then lastly you got the folks who kind of starve in the wilderness and end up getting, getting eaten by bears. That last group is usually somebody who didn't have a map. And so what we try to do is we try to make sure for each role how much of this role is in prospecting, how much of this role is in map making and how much of this is more guide type work and making sure that you have the right requisite group there. Because sometimes at a startup you have to do all three of those roles and you need people that can, that can shift between them depending on the.
Scenario and what you said we put people through the standard test and this and that. What particular ones out there have you found to be the most useful?
So we've used a couple.
In other words, David, if I had to put it to you another way, like the first sort of, you know, you got people coming in, you've got to write your ads to get a certain kind of response, but that doesn't even mean that they're going to even you're going to get what you asked for. But so how do you do your first sort on people coming in and who does that?
Yeah, so we do a lot of, we do a lot of multichannel interviewing. So we've got our HR team that's out both finding candidates, doing the initial recruiting and vetting. We'll do skills based filtration for things like coding tests, but we'll also do some baseline initial interviews there. I think one of the things that we like to really drill in on specifically is why do you want to work at a startup? You can work at any company, big companies, medium sized companies, small companies. Why do you want to work for a company that still has not proven that, that it should exist? Why do you want to sign up to work for a place that isn't a 40 hour work week? It's totally fine if that's what you want to do. We're not that, we're not that kind of place at this stage in our, in our career trajectory. And so we really try to actually disqualify people as much as possible early on. And then after we've gone through all these kind of disqualifying questions in terms of motivation, then we start to say, okay, well what have you done in your past job? What are you most interested in? Where do you see your own career heading? That sort of thing. We really try to dequeue people out early on and we do that through multiple rounds of interviewing with multiple people to end up with the best spot.
Yeah. Now talk about starting new people. You know, you have 70 people. That's a whole lot of people that have had to be located and integrated into your system. Now if you continue on this tripling path, you're going to be at 1,000 people here in the near future. So how are you starting? What have you learned about the best way to start new people?
So the first thing that we like to do is really walk people through what our mission is as a company. And we have it painted on the wall when you can't get out of the lobby at capacity without seeing our mission to help teams do their best work. So we want people to understand the why behind what we do from the very, very beginning. But beyond that, we want all of our team members to use our products. So the best way to sell, the best way to be a customer success person, the best way to be a software engineer. Even on the finance team, we've got finance team members building out workflows around automating different parts of their own, their own day. Like, we want to be a company where it's not just that the engineers and the product team people work on the software. Like, what? Part of, part of our thesis is that software is becoming more and more abstracted, from the zeros and ones to where more and more people can be programming rather than just a few elites in an ivory tower somewhere. And so we kind of view it as getting everyone bought into the product right away, building out their first guided conversations, their first workflows, their first chatbot interactions. And from there people get a sense of, oh, wow, this is a canvas in which I could, I could go build things on top of. It's a really great way to get buy in. Now, it doesn't work for all businesses, obviously we're a software company, but for us that's definitely been key to our success.
And I really appreciate these insights and I think you said it very cleverly that of going through the process of the first years are to prove you can exist and then you grow until you're hard to kill. Yeah, I appreciate those insights.
If you enjoyed what you've heard and are dead serious about finding out for yourself exactly how this works in the real world. I've taken the most valuable business lessons I've learned over 40 years and put them into something for you to watch. Go to whiteellawinning.com forward/webinar now, in order to move up as fast as possible, I'M Larry Weidell, and I run the Million Dollar Mastermind. Go, go.
More from Larry

Rick Jordan on Self-Engineered Proximity, Cyber Security Moats, and Truth-Based Leadership

Fire Fast or Fail Slow: Kevin Brittingham Reveals the One Rule That Built a $100M Empire








