Transcript

Fitz: Hi everyone and welcome to our initial discussion for the Zeitworks Webinar Series. Today, we’ll be talking about the inception of Zeitworks and the development of the product from its early days all the way to current day. I’ll be speaking with the Founder and CEO of Zeitworks, Jay Bartot, who knows this topic better than anyone. Jay - thanks for taking the time to be here - I’m excited for this conversation. I’d love to start by hearing a little bit about your background prior to joining Madrona Venture Labs in 2016.

‍Jay: I’ve been a serial tech entrepreneur in Seattle for 20+ years, so my background is software engineering and I had a number of software engineering jobs. Pretty early in my career, I started gravitating towards startups and in the late 90’s machine learning. I got interested in machine learning from working on a startup with some University of Washington computer science faculty who were spinning out some of their research out of the university and into the private sector. I got exposed to machine learning there and certainly a lot of data and data analysis. I went on to co-found a number of data and machine learning companies, all venture-backed, and interestingly in different verticals – travel, online advertising, e-commerce and medical informatics.

‍Fitz: One of the things that fascinates me about your background is the amount of time you’ve focused on building products that utilize machine learning and artificial intelligence technology. Even when you joined Madrona Venture Labs in 2016, where you took a different role of incubating companies and spinning them out. So, let’s just talk about one of the ideas you spun out which you are still focused on today, which is Zeitworks. I’d love to dig into the details of how you came up with that idea as a starting point.

‍Jay: Yeah sure, so as you mentioned, I joined Madrona Venture Labs (“MVL”) which is a startup studio in Seattle and is funded largely by Madrona Venture Group (“MVG”), the largest venture capital group in the Pacific Northwest. I joined MVL as a Managing Director and CTO in the fall of 2016, about a year after I sold my previous startup to Hulu. The idea with MVL is that there is a small group of experienced entrepreneurs and operators in the lab, and our job was to think up interesting new, venture-backable technology startup ideas. More importantly, we had to research those ideas, do our homework on those ideas, and try to figure out as soon as we can if they were good ideas or not so good ideas. Having been an entrepreneur for so long and working with entrepreneurs, I know from my own personal experience and pain that finding a good idea really requires research and honesty with yourself and teammates. We worked on a number of different ideas during the 5 years I was at MVL – a lot of them were data-centric, ML, enterprise B2B playbook and Zeitworks was one of those.
‍

It turns out that humans are humans – they’re imperfect. A process may be written down at some point but then humans stray from it. Then, some people do the process better than other people. Over time, organically, the execution of the process can become ill-defined.”
‍
Jay Bartot, CEO of Zeitworks

‍Fitz: What went into vetting the Zeitworks idea and at what point did you realize it was one worth pursuing.

Jay: As they often did, Managing Directors at MVG pointed us at the RPA space. This is early 2019, and I can never remember if MVG had done their investment in UI Path yet, or if they were about to, but they, like many other venture capitalists, were very interested in the RPA space. It was booming, growing like crazy. There were a number of big players that had turned into Unicorns overnight. Those kinds of things tend to excite venture capitalists. Basically, we were given the direction to go investigate the RPA space and find out what’s going on and if there are other opportunities in that space now or in the future. So in our usual fashion at MVL, we rolled up our sleeves and started doing our homework which in the early days of an idea typically means talking to a number of people. We talked to RPA vendors about their platforms and some customers as well. We didn’t get very far before we realized that management consulting firms and systems integrators play a large role in the RPA space and frankly in the business process space. Management consulting firms like McKinsey and Deloitte, and small ones as well, are oftentimes the designers of these business processes that a lot of companies run. When I say business processes, I mean claims processes in insurance or loan processing in finance – basically, business processes that are executed by humans but involve moving information around from place to place. We realized that management consulting groups played a large role in not only designing business processes but also the implementation of the RPA technology. A large insurance firm, for example, might hire Deloitte or Ernst & Young to implement RPA bots for them. The epiphany came when we discovered this upstream step that needs to happen around process understanding. This means that the consulting firms will come in and help the organization deconstruct the business processes themselves that they’re interested in automating. We thought this was unusual – an insurance company doesn’t know a process that 50 or 100 people are running throughout the course of a day. Same thing with a bank or healthcare or other industries where these processes are executed. It turns out that humans are humans – they’re imperfect. A process may be written down at some point but then humans stray from it. Then, some people do the process better than other people. So over time, organically, the execution of the process can become ill-defined. In order to implement an RPA bot for a process that a human does, the process needs to be really well understood. This is when we learned that management consultants bring in a gaggle of fresh college grads who will stand over the shoulders of the human workforce executing the business process and write down manually everything they do in the course of the process execution. What applications they use and what they do inside those applications. They may monitor the executions of these operations for a few weeks and then at the end of that period of time, they will bring together their notes together and come up with a canonical, point in time description and process map of the business process that can then be taken to an RPA engineer who can develop an RPA bot that does that process. So, that was the big epiphany for us, which was – that sounds slow, time-consuming, expensive and even error prone. Why can’t you have technology do this? What if you had some kind of application you could install on the machines of the human workforce that would listen to all of the events of the human workforce as they are executing their business processes. That information can then be sent back to a server or cloud and machine learning technology can be used to make sense of it all, form analytics and determine what the business process actually is. That was the “AHA” moment at Zeitworks, where we figured let’s use technology to automate this analysis which is being done manually now.

Fitz: Makes sense and it is a fascinating problem to be solving – one that could make a lot of business process enhancements more effective for companies all over. The natural next step is to talk about what’s changed since those early days. We’ve seen the pandemic hit and many companies have remote workers now. So maybe take us through what’s changed in the product and vision since you took over as CEO.

“That’s when we realized that we were less of an upstream tool for RPA and more a tool that helped managers and workers execute their business processes every day.

Jay Bartot, CEO of Zeitworks

Jay: A number of things changed since the original vision of this problem being solved with technology and data and machine learning. We started talking with more regular customers – customers using RPA or companies with large scale repetitive business process execution groups and teams. This is what we did at MVL – we did a lot of customer discovery to make sure we were thinking about the problem in the right way, that it was a real pain point, and that a solution can be brought to the table. What we found from talking to customers was that many people were interested in automation. When we focused on discovery and analysis problem, that was what they wanted in the near term – help me understand my group, help me understand how I help my team members do better, help me understand where there are inefficiencies in my business processes. Right now, I have no insight into the work that gets done, how efficient it is, how inefficient it is, who is good on my team, who on my team needs training, etc. They were interested in which processes were ripe for automation and the ROI behind that, but first shining a light on their group so that they could manage them better. Ultimately this would make a difference in their operations and bottom line. That’s when we realized that we were less of an upstream tool for RPA and more of a tool that helped managers and workers execute their business processes every day. Since I took over as CEO in June 2021 – first, we are obviously in the midst of a pandemic. A lot of interesting workforce dynamics have come into play with the pandemic, particularly with remote work and this so-called Great Resignation. With remote work, no matter what kind of worker you are, it is more difficult to understand and keep in touch with your employees. It is also more difficult for employees to collaborate with managers and ultimately improve the operations of their business. That is a meaningful change that is driving people to think about the future of work and what kinds of tools and technologies, tools and data are needed in this world where people are working remotely, not working in the same space anymore, not having those serendipitous hallway conversations or in person meetings. If anything, they are probably more productive because they are not sitting in traffic during their commutes. But also, there’s the risk of lack of clear separation between life and work, which can lead to worker burnout. All of these phenomena we are thinking hard about at Zeitworks as we grow and evolve our product. The other thing that dovetails into this is the Great Resignation – a lot of the workforce that is executing repetitive business processes are vulnerable from complications working from home. It has always been a challenge for businesses to staff and retain these employees and in a modern era it’s more challenging to keep people on board and keep people happy.

Fitz: It seems to be a timely product given the surge in remote work and the great resignation, and a lot of these dynamics that companies are facing. Jay, it’s been great to learn more about your background and get the full rundown on Zeitworks. I’m so excited about the Zeitworks product and look forward to hearing more about it in the future.

Jay: Thanks Fitz, great talking with you.

‍