Silicon Valley’s Top Investors Bet on This 22-Year-Old Founder | Sola, Jessica Wu
Jessica, co-founder and CEO of Sola, shares her journey from MIT and hedge funds to raising $21M by age 22 from top Silicon Valley investors, including Andreessen Horowitz (a16z).
Watch on YouTube →Transcript
Intro
- 00:00I was actually the youngest quant
- 00:01researcher at a pretty big hedge fund
- 00:02today and so had started my career in
- 00:05finance but had seen a lot of different
- 00:07areas very thesis driven to just pure
- 00:09numbers and numbers crunching. You know
- 00:11when I was working in corporate finance
- 00:13I would work uh I don't know 9 10 hours
- 00:15a day and uh 5 days a week and then if I
- 00:18was called in on a weekend I would
- 00:20honestly be very miserable. It just felt
- 00:22like it it wasn't my own thing. Now I
- 00:24work well every waking second uh and 7
- 00:27days a week, but I've never been
- 00:28happier. And I think I think I'm young,
- 00:30so that's one piece that gives me a
- 00:32little bit more optionality. When it
- 00:33comes to what you want to do, though,
- 00:35this sounds very simple, but it's really
- 00:37just what feels right. The team feels
- 00:38that way, too. It's really like when you
- 00:40own a piece of something and you're
- 00:41you're really growing it from ground up,
- 00:43it's easy to feel excited. And so my
- 00:45advice for like stable career versus
- 00:48startup is if you have the optionality
- 00:49to I think you know whatever makes you
- 00:51feel happy, whatever makes you feel like
- 00:53you're excited to go to work every
- 00:54single day and that you feel very
- 00:55fulfilled in what you're doing. Hi, I'm
What is Sola?
- 00:58Jessica, co-founder and CEO of Sol. Sol
- 01:01is an agentic process automation
- 01:03platform. We help businesses automate
- 01:05their most critical and operational
- 01:06workflows using AI in a much easier and
- 01:09faster way than traditional RPA. Solus
- 01:11started out of YC about 2 years ago.
- 01:14Since then, we have raised a seed round
- 01:16led by Sarah at Conviction. Most
- 01:18recently, we raised a series A led by
- 01:20A16Z following an enormous amount of
- 01:22traction from our enterprise customers.
- 01:25We have 5xed our revenue this year.
- 01:27Execution volume on the platform is
- 01:28doubling month over month since the
- 01:30beginning of this year. And we're proud
- 01:32to partner with some of the largest
- 01:33companies in the world, including the
- 01:35Fortune 100, Amlaw 100, and some of the
- 01:38largest private businesses in logistics
- 01:40and healthcare.
From Youngest Quant to Founder
- 01:50Uh, I grew up doing a lot of very
- 01:51competitive things. Played competitive
- 01:53piano and I did competitive math and
- 01:55things like that. I think that gives you
- 01:57a pretty hard will. It probably makes
- 01:58you very disciplined. I think I'm good
- 02:00at taking risks and putting myself out
- 02:02in places that I don't feel comfortable.
- 02:05And I think that builds up a lot of
- 02:06resilience. Like doing something even if
- 02:08you don't feel like you're 100% ready. I
- 02:10do that a lot. And just jumping into
- 02:12things and really going for it. I think
- 02:13that builds up a lot of character.
- 02:17When I was in high school, I had visited
- 02:18MIT a couple of times. I think it does a
- 02:21great job of the whole, you know, you're
- 02:23the dumbest person in the room kind of
- 02:25thing. It's always the most fun when
- 02:26you're surrounded by a lot of people you
- 02:27can learn from. I think MIT is a very
- 02:29unique place. It's the only place where
- 02:32it's really exactly as it portrayed in
- 02:34the movies. You have people building
- 02:35roller coasters in the front lawn and
- 02:37you have people training models in the
- 02:39dorm basement and things like that and
- 02:41it's every bit as real as as it's
- 02:43described. I think that MIT really puts
- 02:46a ceiling on how technically difficult
- 02:48things get. I've been working on this
- 02:50company for about two years now and have
- 02:52experienced a lot of different things
- 02:54since college, but never have I had to
- 02:56use like so many brain cells to just
- 02:59think about a really hard problem. I
- 03:01think the most important thing you
- 03:02should do is jump into everything as
- 03:03quickly as you can. And MIT does a great
- 03:05job of showing you exactly what's on the
- 03:07forefront of what you should be paying
- 03:09attention to. You should absolutely go
- 03:11learn about all the new things that
- 03:12people are doing research about. You
- 03:14should absolutely spend as much time as
- 03:15you can with people that are doing said
- 03:17research. And I think that if you create
- 03:20a habit around trying to learn as much
- 03:22as you can and intake as much what's on
- 03:24the frontier as possible, then you're
- 03:26better equipped, especially now to to
- 03:28learn and to work in an environment like
- 03:30today where things change so rapidly and
- 03:32there's so many things happening in the
- 03:34tech world. I think the thing that you
- 03:36get from just a very technical
- 03:37background is you can break things down
- 03:40very easily. If you can communicate what
- 03:42you're trying to solve in a very simple
- 03:43way, it makes it a lot easier to solve
- 03:45and that just gives you a really good
- 03:46framework around solving things,
- 03:48especially in a landscape like today
- 03:50where new model basically comes out
- 03:51every single week and things are
- 03:53constantly changing. I think there's a
- 03:54lot of value in always centering back to
- 03:57like what are you trying to solve for
- 03:58your customer or for your user, whoever
- 04:00that is, and then just building on top
- 04:01of that. And I think once you look and
- 04:03sort of narrow things down to the root
- 04:05problem, not just the problem that
- 04:06people are saying, then you can deliver
- 04:08an enormous amount of value to the
- 04:10people that are interfacing with what
- 04:11you're doing.
- 04:14Before I worked on Solola as a startup,
- 04:16I had no idea that I wanted to do a
- 04:17startup. I had tried a lot of different
- 04:19things. I had worked in venture. I had
- 04:22worked at a couple hedge funds. I was
- 04:24actually the youngest quant researcher
- 04:25at a pretty big hedge fund today.
- 04:27Finance was an interesting choice. I had
- 04:29an adviser who recommended to me that
- 04:31it's actually pretty good to spend some
- 04:32time in finance because it teaches you a
- 04:34way of thinking that's pretty special,
- 04:36especially if you're doing trading.
- 04:38Teaches you to be very objective, that
- 04:39you should always calculate odds for
- 04:41things. I think I am someone who's
- 04:43guilty of not thinking in a very
- 04:45numerical or like standardized way all
- 04:47the time. And so I think working in
- 04:48finance for a couple years gave me a lot
- 04:50of perspective on being objective, being
- 04:52able to calculate and just think through
- 04:54things in a very rational way. And
- 04:56that's something that stuck with me.
- 04:57Doing a startup can be very emotional
- 04:59and I think always going back to first
- 05:00principles thinking and trying to think
- 05:02in a more like statistical way can be
- 05:04pretty helpful in terms of making
- 05:05decisions even like in a startup lens.
- 05:07Maybe you're looking at a super huge
- 05:09deployment and that's something you have
- 05:11to really weigh or someone is really
- 05:13looking at features like all of those
- 05:14things can be mapped out in a way that's
- 05:16more rational. You can say, you know,
- 05:18what are the odds that this particular,
- 05:20you know, this isn't perfect math. This
- 05:21particular customer converts by this
- 05:23particular day. What are the things that
- 05:24I'm trading off? And it gives you a very
- 05:26statistical framework to think about
- 05:27these decisions without just I really
- 05:29like this customer, so I want to work
- 05:30with them. People tell you that you can
- 05:32learn a lot from playing poker and that
- 05:34applies a lot to the startup world. I
- 05:35think that's true. A lot of it is just
- 05:37odds, being rational, and then, you
- 05:39know, making decisions in the best
- 05:40direction. And it's actually very useful
- 05:42when you work on a startup.
The Oldest Inconvenience, The Biggest Opportunity
- 05:48At one of the hedge funds we worked at,
- 05:49they had a lot of manual work around
- 05:51interacting with this really old
- 05:53brokerage software that they built on
- 05:55top of. At the time, I had been using
- 05:57RPA tools to try to automate that. I
- 05:59didn't really know how big the space of
- 06:01RPA was. So, RPA stands for robotic
- 06:04process automation. RPA tools basically
- 06:07automate manual work by replicating
- 06:09workflows the way humans do. So moving
- 06:11the mouse, typing things in, clicking,
- 06:13and interfacing with browser and desktop
- 06:15applications the way humans do. What I
- 06:17did know was that there is a lot of
- 06:19manual work at these larger companies.
- 06:21There's no good tools out there that are
- 06:23really easy to use. And even though I
- 06:24have a computer science background, it
- 06:26was still really hard for me to just
- 06:27build a simple browser or desktop
- 06:29automation. That was sort of where the
- 06:31the seat of the problem happened. My
- 06:33co-founder actually had a really similar
- 06:35experience. He was building out hospital
- 06:37systems at MGH. they had used some very
- 06:39old tooling and he was also surprised by
- 06:42how brittle it was and how hard it was
- 06:44to implement. I think those were really
- 06:46good glimpses for us into what real
- 06:48world work looks like. Um, when you're
- 06:50at MIT, you live in a very tech-forward
- 06:52world and you're surrounded by tools
- 06:54that are very easy to use, that
- 06:56everything has APIs that plug into each
- 06:58other, you can build automations very
- 06:59easily. But in the real world, when
- 07:01you're looking at most companies out
- 07:03there, they're doing really manual work.
- 07:04People operate across a ton of different
- 07:06systems that don't connect. As you can
- 07:08imagine, people doing operations will
- 07:10touch spreadsheets, internal portals,
- 07:12external systems, they will touch files,
- 07:15and just about everything in between.
- 07:16Those were two really lucky glimpses I
- 07:18think that we had into what that
- 07:20actually looked like. From there, when
- 07:22we went through YC, like I mentioned, we
- 07:24kind of came in without too much of an
- 07:26idea. We spent about the first month or
- 07:28so just figuring out what we wanted to
- 07:30build. I think the RPA space was very
- 07:32exciting to us because we understood
- 07:34very well what that problem looked like
- 07:36and how we wanted to solve it. Obviously
- 07:38RPA is very big and I think that piece
- 07:40is exciting to us. The ambition of
- 07:42automating all digital work is a
- 07:44perennial and obvious one and I think
- 07:46the aspects of like being able to marry
- 07:49AI in the real world for enterprise
- 07:51companies and all the technical and
- 07:53model improvements on the flip side of
- 07:54things was a very exciting sort of
- 07:56combination for us.
- 07:59Solo's MVP was very simple. We had a
- 08:01recorder just like today. That part's
- 08:03the same that lets you take a recording
- 08:04of your workflow. You could upload it,
- 08:06but uh different from today. Didn't
- 08:08really do a great job of laying out what
- 08:10your workflow did. You could also run it
- 08:12and it would just run on your own
- 08:13computer and repeat back the same steps.
- 08:15Not quite as intelligent or easy to use
- 08:17as we had hoped. And since then, we've
- 08:18kept a lot of things the same. So, you
- 08:20know, you can still record a workflow,
- 08:22you can still upload it to the platform,
- 08:23but now you can run it hundreds of VMs
- 08:25at scale. You can edit it and change
- 08:28things in very fine granularity. You can
- 08:30add logic. You can add information on
- 08:32top. And you have a lot of optionality
- 08:34on how you want to orchestrate and run
- 08:36these things at scale. One thing we had
- 08:38probably underestimated was just there
- 08:40were a lot of trade-offs that we had to
- 08:41make. But it it is a big question of,
- 08:43you know, existing tools offer
- 08:44everything under the sun and they've
- 08:46been building this tool for 20 years.
- 08:48Like how do you actually get up to
- 08:49speed? One thing that we found to be a
- 08:50pretty big balancing act is just getting
- 08:52to the point where we can be helpful for
- 08:54customers but also, you know, being able
- 08:56to do so in a pretty timely fashion.
- 08:59There are a couple hard moments that I
- 09:00can think of in the early days where we
- 09:02had to give up really exciting early
- 09:04revenue. It was just a point where we
- 09:07had to really focus on building this
- 09:09core product, but we had to turn down
- 09:11customers the very beginning. Turn down
- 09:13people when we were, you know, almost
- 09:14there, but not quite. And that was that
- 09:16was very difficult. I think in the early
- 09:18months you get pulled in a lot of
- 09:19different directions. There were some
- 09:21hard conversations where we had to say
- 09:22no to some huge names that we were very
- 09:24excited about. It didn't feel aligned
- 09:26with the direction that we wanted to go
- 09:27down. We were capacity constrained and
- 09:29it would mean giving up or sacrificing
- 09:31too many other things and other areas of
- 09:33the business.
The Counterintuitive YC Playbook
- 09:38I'm pretty big on advisors. I'm someone
- 09:40that asks for a lot of advice. I like to
- 09:42surround myself with people that I think
- 09:44have a lot more experience. So, whenever
- 09:46I get panicked or I'm under pressure,
- 09:48the first thing I'll do is just go to my
- 09:50all of my resources and get people's
- 09:52takes on things. I think usually that
- 09:53puts me in a pretty good place. We chose
- 09:55YC because we had heard really good
- 09:57things about it and that was also the
- 09:58advice that we got early on, which is
- 10:00that you're going to be surrounded by a
- 10:02lot of really amazing people and you're
- 10:03going to have 3 months to lock in. YC
- 10:05tells you to try and they really push
- 10:07you to try and sell something. This
- 10:08means that even if you don't have a
- 10:10working product, you can put up like a,
- 10:12you know, use lovable or something to
- 10:13put up a fake fake front end and then
- 10:15you should try and go sell it. And the
- 10:16good thing about that is a lot of the
- 10:18times when you build something, I don't
- 10:20know if this applies as much to solo,
- 10:21but in general when you build something,
- 10:22you don't know if you're going down the
Stop overbuilding, Start selling
- 10:24right direction. Most clear way you know
- 10:25you're building something that people
- 10:27want is if they'll pay for it, right?
- 10:28And if they'll keep paying for it. YC
- 10:30does a good job in really pushing you to
- 10:32sell even if you may not be there. You
- 10:34know, don't spend 6 months to build a
- 10:36product. First try and sell something.
- 10:37See if it works. You might burn a couple
- 10:39bridges doing that, which is why I'm
- 10:40saying your mileage may vary, but at
- 10:42least you have very clear picture of I
- 10:44know this person would pay for this. I
- 10:46know a lot of people would pay for this
- 10:47and they would depend on it a lot or I'm
- 10:49solving some really meaningful problem.
- 10:51First cell is not good when you get a
- 10:53little bigger like you know don't sell a
- 10:55fake feature. It still is a pretty good
- 10:56mindset to operate behind which is even
- 10:58once you're working on the right product
- 11:00there's a lot of directions you should
- 11:01go try and experiment build different
- 11:03mockups of things try and talk to
- 11:05customers get something out there and
- 11:07then go from there and iterate and
- 11:08polish it afterwards I think we met our
- 11:10first customer actually very early on it
- 11:12was around YC we were in the phase of
- 11:14doing the thing where we sold the
- 11:16product that didn't really exist the
- 11:18good thing was we were pretty honest
- 11:19with them and we told them this thing
- 11:20doesn't exist but it will exist in a
- 11:22couple months and they were like okay
- 11:23well come talk to us when it does I
- 11:25think it's really hard advice. It's very
- 11:26unintuitive. You know, you're taught
- 11:28your whole life to finish something and
- 11:30then go sell it or then go present it or
- 11:32whatever it is. And this is almost the
- 11:33entire opposite. It's like put up the
- 11:35most minimal fake version you probably
- 11:37can of something and then go sell it and
- 11:39then if it's good then go back and build
- 11:40it. I think it's not for the faint of
- 11:42heart for sure. And sometimes, you know,
- 11:44maybe you really succeeded and you
- 11:45actually sold the thing and now you have
- 11:47to deliver. And YC's advice there is
- 11:49probably like spend a week in the
- 11:50basement coding so that you can get it
- 11:52out. It's not easy advice to take and
- 11:54selling without actually having the
- 11:55thing is not very intuitive, but I think
- 11:57it it at least puts you on the right
- 11:59direction. It gives you a very clear yes
- 12:00or no for if this is something you
- 12:02should be doing at all.
Delivering Happiness
- 12:07One of my favorite books is called
- 12:08Delivering Happiness. It's quite good.
- 12:10It's about like sort of how the little
- 12:12things matter and how customer delight
- 12:14is kind of a north north star, but it
- 12:15doesn't just come in one form factor of
- 12:17like I I did this thing on time. There's
- 12:19a million other ways you can build that
- 12:20up. If your customers are happy, most
- 12:22other things fall in place is something
- 12:23that we've tend to found. You know,
- 12:25especially when you're building very
- 12:26critical software for a company, you're
- 12:29asking a company that's been around for
- 12:31decades or maybe even centuries to trust
- 12:33you to do their most critical
- 12:34operations. That is by no means easy. To
- 12:37this day, uh most of Sol's customers
- 12:38come through word of mouth. And that's
- 12:40because existing customers have felt so
- 12:42happy with the experience. You're small,
- 12:44you're early, so what can you do to make
- 12:45up for being a much newer tool? And
- 12:48that's usually one build a really great
- 12:49experience and really solve their
- 12:50problems. That's the obvious one. Second
- 12:52is to ship fast and really take in good
- 12:54feedback. This goes both ways. So you
- 12:56know you want to pick customers that
- 12:57will give you good feedback and you also
- 12:59want your customers or you also will
- 13:00want to take in the feedback that they
- 13:02give you and move very quickly on what
- 13:04they're asking. There's a lot of ways
- 13:05that are more traditional than just
- 13:07building good tools where you can win
- 13:09enterprise companies over. You can be
- 13:10very supportive. You can listen to
- 13:12people. Listening is the most important
- 13:14thing you can do for customers. For some
- 13:15of our early design partners, we still
- 13:17do. We hop on a weekly call with them.
- 13:19We basically take in their feedback week
- 13:21over week and then we try to deliver at
- 13:23a pretty fast rate. And that's the sort
- 13:24of thing that will make people pick the
- 13:26startup over the really old legacy
- 13:28incumbent. Aside from just the tool
- 13:30being better,
- 13:32one of our largest customers wanted to
- 13:34do a deployment of Sol last year. They
- 13:36wanted to do it during the most quiet
- 13:38period of the year, which was on
- 13:40Christmas day because they don't get a
- 13:42ton of volume then. I hope that we don't
- 13:44have to ask this of the team again. We
- 13:46all came in, we worked that week. They
- 13:48had been using Solo for a long time and
- 13:49they had been championing it internally
- 13:51and were very excited. We were like,
- 13:52"Okay, if the deployment is going to
- 13:54happen on Christmas, like we're going to
- 13:55do it." There have been moments like
- 13:56that. I think in general, we try to
- 13:58really go above and beyond. People are
- 14:00trusting us to run their operations. So,
- 14:01they'll use sola to send their invoices
- 14:04and do their accounts payable or they'll
- 14:06use it to actually send out shipments or
- 14:08they will enter in patient data. So all
- 14:10these things are very critical like uh
- 14:12one thing that we realized when we built
- 14:13solar is like we can never go down. That
- 14:15actually means that a real business out
- 14:17there who has hundreds or thousands of
- 14:19employees like their their billing is
- 14:20going to be late that day. We've done a
- 14:21lot to basically maintain that. It's one
- 14:24of those things where it's a little bit
- 14:25different from like a consumer tool or
- 14:28something that's AI that's generating
- 14:30copy or you know finding sales leads
- 14:32like these things. If your tool goes
- 14:34down it's a big deal and it disrupts
- 14:36people's work but it's not the end of
- 14:37the world. the nature of our business
- 14:39when we go down is sort of the end of
- 14:40the world. So, we've done a lot for our
- 14:42customers. We stay up and running all
- 14:44the time. We've spent a million weekends
- 14:47and a million different holidays in the
- 14:49office. But, I think along the way, it's
- 14:51it's been good. Like, there's a little
- 14:53bit of, you know, we're extremely
- 14:54thankful to the people who trusted us to
- 14:56take care of these workflows and made
- 14:58the bet on an early new startup. And
- 15:00hopefully, we can continue to deliver on
- 15:02everything that we've promised.
If You Can't Do It for 10 years, Don't Start
- 15:08I think building a startup is really
- 15:10really unimaginably difficult. You're on
- 15:12a roller coaster and you're constantly
- 15:15experiencing very high highs and very
- 15:17low lows and then it's also compressed
- 15:19and so it's happening, you know, morning
- 15:21might be your best day ever and then 2
- 15:23hours later you're suffering from
- 15:25probably the worst news you've ever
- 15:26heard. And this happens uh pretty much
- 15:28every day. um maybe every week, maybe
- 15:31not like an exact moment, but I think as
- 15:33a founder, you really just learn to you
- 15:36learn that things sort of return to the
- 15:38norm. You have this comfort in your head
- 15:39that things will be okay even if they
- 15:41don't feel okay because they always end
- 15:43up being okay provided that you are
- 15:45building in the right market and also
- 15:47that you have the right people around
- 15:48you. And with those two things, you
- 15:50might have some huge incident happens.
- 15:53Really bad news like someone leaves the
- 15:54team. Maybe a customer is your champion
- 15:57who you were selling to left and in the
- 16:00middle of the sales process. Like these
- 16:01things feel really bad. And then once
- 16:03you get a little bit more used to them
- 16:05and they've been happening so often, you
- 16:07learn to take a step back and you're
- 16:08kind of able to process a lot more of
- 16:11the things happening at a startup um
- 16:13than before. And I think the early team
- 16:14also kind of learns that as well and be
- 16:16able to look at the big picture and that
- 16:19tends to be much more helpful in terms
- 16:20of just mental piece when you're doing
- 16:22this.
- 16:24What we're trying to build is a really
- 16:26big mission. So there will be no problem
- 16:28of us having 10 years to spend on it. I
- 16:30think it's an enormous problem space.
- 16:32The other is that I I really do believe
- 16:33in what Solola is trying to accomplish.
- 16:35Um, I think that there's a lot of work
- 16:37that people shouldn't have to do that's
- 16:38very manual and very tedious. And if we
- 16:40can free people up, they can do things
- 16:41that are very fulfilling and creative
- 16:43and very interesting. And I hope that,
- 16:45you know, businesses can start running
- 16:46on sola and relying on sola to do the
- 16:48things that are manual and wrote and
- 16:50that people don't really want to do and
- 16:52instead spend time on things that are
- 16:54exciting and think through leadership
- 16:55and strategy and decision-m and all the
- 16:57more interesting parts of work.
- 17:07[Music]