Do Your Research

Do Your Research

Colby Tunick

Devin Miller

The Inventive Journey

Podcast for Entrepreneurs

11/6/2020

Do Your Research

Do your research and take the time to understand the different segments within your market so that you don't have pointless conversations. Your not talking to the wrong people. Sometimes you have to talk to the wrong people to know that you're building the right thing. You want to keep those wrong conversations to as few as possible.


The Inventive Journey

Starting and growing a business is a journey. On The Inventive Journey, your host, Devin Miller walks with startups along their different journeys startups take to success (or failure). You also get to hear from featured guests, such as venture firms and angel investors, that provide insight on the paths to a successful inventive journey.

Get New Episodes

Get 2 brand-new podcast episodes sent to you every week!

ai generated transcription

is to do your research and take the time to understand the different segments within your market so that you don't have pointless conversations you're not talking to the wrong people and sometimes you have to talk to the wrong people to know that you're building the right thing but you want to keep those wrong conversations to as few as possible [Music] hey everyone this is devin miller here with another episode of the inventive journey i'm your host devin miller the serial entrepreneur that has built several businesses and startups and small businesses to seven and eight-figure companies as well as the founder and ceo of miller ip law where we help startups and small businesses but their patents and trademarks and today we have another great guest on the episode it's a colby tunic i think is how we pronounce it and uh kobe is uh started out in about 2015 graduated from san diego state and political degree and looking into trade and then he uh really or after that found a passion for ai and machine learning and was kind of a self-taught coder and then and i think i don't know if i you transferred to san diego state in 2015. i almost got that wrong graduated in 2017 and then started working on some defense contracts you went from there decided he was going to get into more of how you apply machine learning to insurance and started uh to do as the economy hit with or started out to look into that more and then the economy hit with covet had to make some pivots and adjustments and now he's built it to where they're in the marketplace have some pilot pilot clients and are being uh growing from there so i'm sure i slaughtered it to some degree but i did my best but welcome to the podcast colby devin great to be here it sounds uh more of a journey when i hear you say it than when it's in my own head all right well that's what i'm here for so with that i gave a very brief introduction but maybe going back to kind of when you did we're at san diego state we're studying politics and kind of pick it up from there absolutely so san diego state university had the opportunity to work with the center for information convergence and strategy focusing on international trade my degree was international security and conflict resolution and had the ability to travel all over the world and talk and deal with you know different individuals tackle different trade problems that really just came back to the better implementation of uh during that time uh i began to look into the use of machine learning or as you know the buzzword today's artificial intelligence i think anybody who's firmly in the machine learning space will say that artificial intelligence the long way off i myself would love a sentient being to do my cooking and cleaning for me and you know drive me from point a to point b but don't think we're there yet uh and i began looking into this use of artificial intelligence machine learning to really help combat hate speech online that's a really intractable problem there's a lot of different mediums that are used for this sort of hate speech and the interesting thing is it's not none of it's uniform a lot of it uses coded language it's not obvious and those are real places that machine learning shines and have been validated finding single points of data among millions of points of data noticing trends over time and then using this machine learning to help target education dollars to combat hate speech a lot of a lot of what hate speeches is just lack of education there's lots of great nonprofits out there that spend millions of dollars a year putting money into schools to try to do diversity and education training which is even more relevant in 2020 than i think it was necessarily in 2015 or at least more prevalent in 2020 than it was in 2015 when i was starting to work on this anyway graduated um worked as an international defense consultant and projects all over the world hold on just diving really quick into that so you talked about hey you kind of got interested in using ai and just a 30 second aside from my information is there a difference between ai versus machine learning is machine learning kind of more we we have a whole bunch of workflows and you know decision making trees and just make those really good or is there a difference between machine learning and ai yeah so when we think about ai oftentimes we think about what we see in the movies again we think about computers that are sentient uh computers that have the ability to make novel decisions without a precedent right they're able to just see a situation understand the situation and then act on that situation and of course that's not how machine learning works today machine learning is very much in the statistical field you're looking at trends over time anomaly detection and because of that even though machine learning falls under artificial intelligence uh as kind of an umbrella master category they are two different things uh and in terms of where the field is going or where we're seeing real innovation most innovation we're seeing is on the machine learning side uh just because sentient beings are probably on the horizon but just a long ways off interestingly though how you train a machine is the same way you train a three-year-old or your dog you showed lots of examples repeated patterns showed good and bad and over time it has the ability to start finding those uh patterns statistically so just think about that next time you hear about ai it's very much like teaching a kid how to catch a ball or i'll give i'll give one more side or i have two questions i'll give one site so if you ever want to learn about machine learning and you probably know you already know a whole lot more than i do but one was interesting so i love podcasts i do my own podcast i like listening to them if you ever go to uh american innovations is another podcast it's put down by wondering and they have two of my top favorite podcasts one is called business wars and the other one is called american innovations but they actually just went through a whole long series on the early all the way up starting from the beginning of ai to where it's at today and talking about how it started out with doing it on a lot with trying to figure out how to play chess and how to be grand masters and that was a good measure for is it smart or is it just doing you know how to do it so if just as a complete assignment since you're interested you may want to check out that uh podcast series um now jumping back to it so you did it for hate speech and before we jump on to graduation working for defense contractors did that ever go anywhere did you ever do anything did you just learn a whole bunch or whatever happened with your interests with competing combating hate speech and that while you were in uh school yeah so that interest is still there but unfortunately it didn't get a lot of progress and that's because um oftentimes the biggest struggle obstacle to any sort of public good initiative is trying to make it financially self-sustainable and i know as a serial entrepreneur you know that all too well nonprofits do not operate and what we would consider as entrepreneurs a financially sustainable way because they don't produce anything that people will pay for they request in donation dollars and then allocate them out to causes that their patrons care about um and that was the real reason that uh it didn't get as much traction at the time as it liked and i was also still learning a lot about the field machine learning in 2015 through 2017 is far different than machine learning in 2020 we've really seen a renaissance in the application and usefulness of the technology because of how open source development has progressed over the last five years um okay so and that makes that makes good sense so now you do that so you decide okay i'd love to find this interesting can't quite figure out how to make a career out of it or how to make a profit on it so i've got to find something that will pay the bills once i graduate and he moved into a bit of defense contracts is that right yeah so i initially started as a grant writer working on different brands for companies that were looking to get government money most of them were in the defense field it was really exciting because it was an application of my background that a lot of people need and not a lot of people want to do grant writing is very difficult but it's also very educational i was writing grants for both the us department of defense as well as the uk ministry of defense and then supporting my project manager on a whole bunch of other projects around the world and you learn a lot about how companies operate and also how governments work with those companies um and really where their synergy between the two and where there's opportunity okay so you did that you worked for grant for working with grants and other you know there's parallel things for a period of time and it did that and then it sounds like you know remind me so you did that for a period of time and then you said hey i want to do something different i want you moved into the insurance kind of industry and looking into earthquake authority so remind me how you did that or how did you move or made that transition from defense contractors to earthquakes and insurance yeah so i think it's great to say or just just acknowledge the fact that i was very lucky getting that job right out of college i and i recognized that that set me on a pathway that a lot of recent grads don't have i think when most people graduate there's that initial confusion period unless you know you want to go and get your medical degree or you want to go and get your juris doctorate or you want to pursue it you know a harvard degree in microbiology for instance most people don't have that sort of direction out of their undergrad after working at uh as that with the defense consultant as a defense consultant um i wanted something that could help people and also do government work work in the government um and i applied to and uh got a position at the california earthquake authority which is the second largest single line natural catastrophe insurance company in the world california is unique because we are so earthquake-prone uh there are of course government-owned insurance companies for a lot of things i think floods uh the national flood insurance program is probably the one that comes to mind for most people but other sovereign nations also have programs like this so new zealand and japan both have at their you know national level earthquake insurance programs so this was i never thought i'd end up working in a for an insurance company in insurance uh that was never something on the radar and i think that's most people in insurance if you talk to them they were doing something else and you know you don't just grow up as a little kid say i'm going to work in insurance i'm going to be an insurance agent i you know i haven't met anybody i've also only met one person in my life that wanted to grow up and be a dentist so i think there are just some career fields that lend themselves to more adult versioning i guess you could say sure so so you jumped in big second biggest insurance agent so i'm sure california both they have wildfires they have earthquakes and they have a whole bunch it's a good place to live on some sense and you better have good insurance on the other side so you worked for that for a period of time and then how did you then transition over to kind of what you're doing now i think you said you were now going you kind of saw the opportunity to apply machine learning to insurance yeah so insurance is one of the few industries that have too much data most other industries struggle from a lack of data and it's interesting that you talked about the american innovations podcast because when we think about the history of artificial intelligence or machine learning the same algorithms have been in use since the 60s the big limiter in the space has been the amount of data to train these models so a great example of this is gpt3 this is google's new natural language processing algorithm and it used something like billions of uh data points and took months to train um that's not data that was available in the 60s so the great limiter for machine learning has always been the amount of data to train these models and insurance is one of the few fields that has too much data and also doesn't necessarily know what to do with it um accenture estimated uh i believe in 2016 that insurance companies only use 15 10 to 15 of the data they have and this is all the data they have um where i work is uh you know it's a it's an insurance carrier it falls directly under this you know 10 to 15 of data and it was difficult for us to try to answer internal business questions because we have so much data how do you turn that into actionable items machine learning is great because it you can look at that at the macro level and come up with you know micro tactics really is where machine learning shines and you can automate it in such a way where you don't need to have a data science team looking at the same data day in and day out to try to come up with just that new version of the data you can hook into the data source automate the whole process and come up with really actual business intelligence at the end okay so so now you you kind of have that idea that realization hey we've got a lot of data insurance companies have more data than they know what to do with they have they're not utilizing it or leveraging it as well as they otherwise could so you know kind of the same questions before how did you what did you do to start forming a business or making a strategy or to fix that problem what was the next step for you yeah so to be fair i didn't actually start off using machine learning for insurance i was i was still in the vein of machine learning for social media and what i did is i went out and i talked to 10 people and that's the minimum number of people that any person is interested in starting a business of any kind whether it's a high growth tech startup or it's you know a local service company should speak with the start to understand the market so i did that spoke to 10 people in the marketing space and i got lukewarm feedback at the same time uh my business partner was my business partner then but uh now he's our chief operations officer elijah chang uh he and i were going to the uc davis big bang innovation challenge and he was also talking to people about my idea without me knowing which was really interesting and just giving me feedback on it from what he was hearing from people in the marketing field and uh the feedback for that was not very good uh to be fair and i think that's one of the hard things in having this idea you believe is amazing and then people that you would ultimately sell up to say i don't need it don't want it too expensive not interested so i called up an advisor who i knew from a previous startup i'd worked at and i said you know we're running into this obstacle what do you think and he said i really think you should evaluate the industry you're in which is insurance do some market discovery there start understanding the numbers and there there may be a better application of your technology there and he was absolutely right it was interesting because that was the piece of the puzzle that was sitting right in front of my eyes but i just was too close to the forest to see the trees so to speak so then a couple weeks after that we spoke to 10 more people in insurance and these were very preliminary calls these were calls where you asked three questions you know what are the problems you're facing uh on a better way to solve those problems right those are the three standard business development questions anyone should always ask um and the feedback we got was really positive and they said yeah there's lots of need in insurance for us to use our data we're not necessarily that we're struggling but we're kind of churning our wheels um and we were like okay there's something here so then what we did was we took to another 40 people brought us to 50 and we began to invalidate some of our early assumptions early on we thought we'd be on the underwriting side of the house and which is just the process of a company taking in a a potential policy holder and saying is it worth us taking on this risk is this person going to fit into our risk tolerance into our risk portfolio and invalidated that had a couple more conversations uh never stopped talking to people and um started then building a product after these initial conversations to just show uh two interested parties and i think reid hoffman the ceo of linkedin founder of linkedin did really well and he said you should always be embarrassed in your first product i know i'm paraphrasing it but it's entirely true you always when you put together something small for the first time that it's functional it works but it's not it's not your vision of course most of the time you show it to someone you kind of like grin you bear it kind of maybe suck on your teeth a little bit and say what do you think and grace for impact um so that's really how we got started and so now see so you get started and then you know i think it was you said it was what it was january february they get started or when was that when you really made the launch or got started we got really started in jam midway through january so you make you get started you said okay we figured out finally our fit we figured out you know who will necessarily pay where it's going to fit in where the data is going to be and you kind of get all that information together and then you mentioned kind of covet hit and that had a bit of a downturn you know that or you had to make some pivots or adjust is that correct yeah so um code was the unexpected uh wrench in the planet early on so covet really took off in march of 2020. of course we we were starting to hear rumors of it in february but 2020 is when things really started to change rapidly for everybody not just for us as early co-founders not just us in the insurance market everybody in the united states and around the world went through a really dramatic change overnight um and we were talking with someone who we wanted to be an advisor to the company ultimately to not become an advisor to the company um and he sent us an email in in march telling us you know this is literally the stupidest time in american history and american history was in caps in this email to go and start a company that's got to make you encouraged hey this is the absolute worst dumbest or time you could ever start a business why would you even think about it so what did you do with that um i think i probably made myself another cup of coffee and kept working so we always starting a company people always talk about there's going to be obstacles and there's going to be trials and tribulations and i guess covert was the like the trump card the ace card the ace in the hole whatever you want to call it right the the maximum amount of disruption possible uh but we just kept working we were still early on enough where we didn't have the product in the market we were still doing early customer discovery so it didn't interrupt necessarily our viability what it did do though is push us to really focus in on the problems our customers have and that comes into using their data to make more sales and we were kind of skirting around this issue and we were speaking to a company that's right in our sweet spot and they said we love what you're doing but at this time we cannot buy anything that's not going to help us make more sales and after that we really took a hard look and said okay how can we fit into this new normal where everyone in this entire industry is now sales focused staying alive focused if you will so koba definitely helped uh but at the same time my co-founder went on employment which was quite stressful for him i went through a divorce which was also uh definitely a bit of a game change uh so covet unemployment divorce you know these are all things that you never think will happen and you also never think will happen when you're in the middle of you know 16-hour days trying to build a company from nothing [Music] a little bit makes it a little bit harder to to get your get yourself excited about the 16 hour days when you're going through divorce you're going through kovid you're having a hard time getting or making payroll or getting any money coming in so i'm sure that's stressful now let's fast forward so you made covet started still going today although things are at least opening up a bit more getting a bit more of a normalcy so how how did you go from that point in life to now today where you've got a couple pilot clients and you're you know having some revenue source and building the company yeah i think it was just a matter of continuing to talk to individuals that was really the um the the factor that got us over that you know initial three-month come from i guess you can say april through june uh was just continuing to talk to people continuing to push forward on product development continue to get closer and closer to the pain our customers have that we're ultimately trying to solve and just talking just continuing to talk to people um the paid pilots that we have today were not to say totally unexpected but these were situations where we sent someone an email and they responded and said let's jump on a phone call today let's just use your product today and we're very excited to continue on those paid pilots and continue the development of the product and one of the interesting things is for some of the paid pilots we have in the works we had people who just asked to buy the product and they're like we just want to give a test today implement it for hundreds of people uh and it was actually really frustrating that i had to say no to that and we had to go to play pilot ralph because we just don't have the team we're team of four myself my cto uh alexander pearson goulart uh elijah our ceo and then uh botswal garwal who is our machine learning engineer and they've all been fantastic partners in this you know they've all pushed day in and day out and realizing that among the four of us right now we don't have the resources to just bring on these hundreds of uh users uh has been a learning experience um and that's also something that you don't think about when you start a company you'll think once people want to once we get to the point where people want to pay for it of course we'll just be able to have them pay for it yeah it's one of those you're like man we'd love you to pay for it but we can't quite do it yet so so all right so so and congratulations that's exciting for be able to take it all the way from idea conception make it through covid get told it's the worst time to start a business actually having both customers that are pilot customers as well as ones that want to buy it so well there there are always more things to talk about than trying to talk about but what i would like to do is now kind of as we jump or get towards the end of the podcast i always have my two questions so we'll jump to those now so the first question i always ask is so within your journey within everything you did what was the worst business decision you made and what did you learn from it yeah um so the worst business decision that we've made uh has not was not necessarily the worst business decision for this startup but as i mentioned uh myself and both my co-founders have been part of other startups and the mistake that we learned there that we didn't repeat the worst business mistake that i've ever seen made is not listening to what your customers want thinking that you know more than your customer and then designing a product that doesn't actually solve their pain point and then you know realizing that that's not why you're making not making sales no i think that is one that you have to learn hey if i i can have the world's best product if nobody's willing to pay for it or they're not paying what we need for it it doesn't matter how good of a product or idea it is it's still not going to work so i think that's one certainly a good lesson to learn from so okay second one question i always ask is so now if you're talking to someone that's just getting to start up just getting into a small business what would be the one piece of advice you'd give them yeah um i guess this is something that you would normally learn by going through a free startup program or an entrepreneurial course and of course those are all you know highly recommended most communities have a startup or an entrepreneurial focused organization that just free education for anybody in the community who's thinking about starting a business but the one piece of advice that i would give to someone who's looking to start a business is to do your research and take the time to understand the different segments within your market so that you don't have pointless conversations you're not talking to the wrong people and sometimes you have to talk to the wrong people to know that you're building the right thing but you want to keep those wrong conversations to as few as possible no i i think that's great advice and certainly something to take to heart so now people want to use your product they want to be a pilot or pilot customer they want to work for you they want to donate to you they want to any any or just know more about how you're applying machine learning to insurance industry or any or all of the above what's the best way to connect up with you and find out more absolutely so uh you can go to our website which is refocusai.com and contact us that you can reach out to me on linkedin it's just colby tunic on linkedin or send me an email at colby 3focusai.com all right well i definitely encourage everybody to reach out connect up find out more so well thank you again for coming on the podcast it's been a pleasure now everybody else if you have your own inventive journey to tell feel free to apply to be a guest on the podcast by going to inventivejourneyguest.com if you are a listener make sure to click subscribe so you get a notification of all the new episodes as they come out and lastly if you never need any help with patents and trademarks for your startup and small business feel free to check out us at millerip law we're always happy to help thank you again colby it's been fun to have you on and wish you the next leg of your journey even better than the last thanks devin it's been a pleasure stay safe and take care thank you English (auto-generated) All Recently uploaded

Download This Episode & More  on the Following Platforms


Podcast for Entrepreneurs on Apple Podcasts
Podcast for Entrepreneurs on Spotify
Podcasts for Entrepreneurs on Google Podcasts
Podcast for Entrepreneurs on Simplecast
Podcasts for Entrepreneurs on Pocket Casts
Podcasts for Entrepreneurs on Stitcher
Podcasts for Entrepreneurs on Tune In
Podcast for Entrepreneurs on Deezer
Podcast for Entrepreneurs on Radio Public

JOIN US ON SOCIAL MEDIA


← Another Awesome Article Another Awesome Article →



Love to here your Comments/Feedback (If you want to chat with us directly grab a strategy meeting)

Please note, comments must be approved before they are published