Why DCM Ventures is investing in x.ai
What’s the most you’ve ever done to schedule a meeting? Well for me, it took 43 e-mails, 2 months, and 20 apologies on each side. When I first joined DCM Ventures, I connected with an associate at another firm over e-mail, and we were all set for lunch in downtown Palo Alto on a Monday. On Sunday night, I get an e-mail from him, “Sorry, dude. Can we reschedule?” No problem. We go back and forth some more, and settle on Wednesday afternoon that week for coffee. My turn. It’s Tuesday afternoon, and now I have to reschedule — I’m in San Francisco all day as opposed to the South Bay (one of the many nuances of working the Bay Area). We decide to raincheck and reconvene once we both have a better grasp of our upcoming schedules. Two weeks go by, and I shoot him a note, “How are things looking? Want to do that lunch?” Blank rectangles as far as the eye can see on both of our calendars, so we put something down for lunch. The meeting lives in e-mail, but neither one of us had bothered to send a calendar invite. On the day that we’re scheduled to meet, I end up being in Menlo Park, but he’s in the San Francisco. Both of us had forgotten about the meeting since there was no invite. So there we were, back to square one. Living the dream. Fortunately, we did end up meeting. Coffee was great, getting there was a pain.
I’ve gotten better at scheduling meetings, and I’m sure there’s a world of expert meeting schedulers that make my skills look amateur in comparison; however, at the end of the day we’re all just humans.
Venture capitalists like to tell people that our world is driven by relationships. Our relationship graphs drive everything in this business -market research, deal sourcing, due diligence, deal closing, and exit creation. We create and nurture those relationships through an endless slew of coffee chats, lunches, and phone calls. Needless to say, a large part of our day is spent scheduling and re-scheduling those aforementioned meetings. We cancel and raincheck. Then, we try and schedule again. Sometimes, if we’re lucky, we meet and promise to put something more substantial on the calendar for the future.
Scheduling meetings is a lot of work, and sometimes, I screw it up. I’m human after all.
Enter x.ai, which makes an Artificially Intelligent personal assistant who schedules meetings for you. All you have to do is cc’ Amy Ingram (or her brother Andrew), and she’ll take over the job of scheduling your meeting, start to finish. The company is founded by Dennis R. Mortensen, a serial entrepreneur who had counted that he had, in one year, set 1,019 meetings with 670 reschedules. In his mind, going through one’s inbox and scheduling meetings was a task best left for the machines.
In 2015, as a VC Associate, I did an average of 53 meetings per month based on the number of Outlook invites on my calendar. That’s around 630 meetings in 2015 (you win Dennis, clearly you’re meeting with more VCs than I am entrepreneurs). If I conservatively estimate that it took me on average 15 minutes to schedule each meeting, then that’s 9540 minutes or 159 hours. That’s a lot of time that I could have spent doing something else (like reading Medium posts about habits of highly successful people).
Are Dennis and I outliers because our jobs are more meeting-centric than others? Probably not. Every knowledge worker takes part in meetings, even if they’re not necessarily called meetings:
— Engineers have daily scrum meetings and code reviews
— Sales/BD folks have sales forecasting, pipeline reviews, and client meetings
— Finance/BizOps teams meet constantly to go over company metrics and quarterly reports
— Product Managers/Product Marketing Managers meet with all of the above folks constantly
Outside of the tech industry, almost every office in corporate America is filled with meeting rooms. From big board rooms to small phone conferencing booths, the offices of the world are filled with rooms designated for some form of meeting. If you are a knowledge worker, then you are either constantly pushing meetings or being pulled into them.
We are extremely excited to be partnering with x.ai and investing in their $23M Series B along with the amazing folks at Two Sigma Ventures, Work-Bench Ventures, IA Ventures, Firstmark Capital, Softbank Capital, Lerer Hippeau Ventures, Crunchfund, and Pritzker Group Venture Capital.
- They are solving a big freaking problem — scheduling meetings. 87 million knowledge workers in the United States schedule ~10 billion meetings a year. The average meeting takes 17 minutes to schedule, and the average salary of a knowledge worker is between $31,000 and $52,000. That’s over $70 billion in wasted resources. Big market.
- Dennis, Alex, Matt and Marcos have built a world class team of data scientists, engineers, AI trainers, and human computer interaction experts to execute on this vision and make it a reality. From inception, Dennis and his team decided that x.ai would be a data science-centric company. Nearly half of their 64 employees are in data science or AI training. Because of that choice, they are the best equipped team to solve this insanely tough problem.
- They have built a true AI solution from a proprietary data source that they themselves painstakingly annotated. We first met Dennis and Alex in November of 2014. We asked them what they were working on, and they said — “we’ve locked ourselves in the basement and we’re not coming out until we build this thing.” They weren’t kidding. The x.ai team never made any compromises with regards to what its intelligent agent would do. It would never involve a human typing responses because: 1. humans don’t scale and 2. humans screw up more than machines. To do this, x.ai needed to build its own data models from scratch and annotate millions of e-mails in order to train its machine to schedule meetings. No open source AI engines and certainly no off-the-shelf data they could buy.
- They have built a naturally and inescapably viral product with a strong network effect. Guests of meetings will always be exposed to the x.ai product, and guests must use it since it is the gatekeeper to meet with the host. The more users in one’s social graph, the better the product experience becomes. It won’t be a world where Amy speaks to Andrew, but rather a world in which Amy/Andrew is talking to itself. As the user base grows, the experience gets better.
- We fell in love with the product ourselves. Our portfolio company CEOs rave about it, and our executive assistants can’t tell that it’s not a real person. At DCM Ventures, we tug and pull at every single product/service that we invest in to find all the edge cases (yes, even Eaze). Everyone at the firm has been warned to not fall in love with Amy. She’ll break your heart :)
We could not be more excited to start the journey towards democratizing personal assistants with the x.ai team.
If you want to bring Amy or Andrew into your life, you can join the waitlist here!
Thanks for taking the time to read!
-The DCM Ventures team