We welcome the first segment of Rana el Kaliouby’s great origin story on Affectiva, the Waltham, MA company that she co-founded with Roz Picard. Affectiva has developed technology that enables digital interpretation of emotions in order to gain insight into consumers’ engagement.
Affectiva Origins — Part 1
Life does not always work out the way you plan it. When I graduated from the American University in Cairo (AUC) with a degree in Computer Science, I had a Grand Plan: Get my doctorate degree abroad, then return to Egypt and join AUC’s faculty. Never in my wildest dreams did I think I’d start a company 20 minutes outside of MIT, and on the frontiers of digital technology.
Our emotions influence our health and well-being, how we learn and how we make decisions, the big and small. In fact, our desire to connect with one another, empathize and express ourselves by showing how we’re feeling is at the very core of what makes us human.
Our emotions are an inseparable part of who we are, yet they are missing from our increasingly digital worlds, hampering how we communicate and interact. Our communication today is quick and efficient, but it is also very superficial. It lacks the visceral, lockstep emotion expression that is so critical to fostering empathy and creating a deep connection. How we communicate today is out of sync with how we naturally communicate best — through and with emotion.
By bringing emotional intelligence to our devices and digital experiences, I believe we can close the empathy gap — fostering better ways for us to connect with each other as humans, and building more likeable and persuasive machines that can bring positive change to our lives.
So that got me thinking — what if our devices could read and respond to our emotions? What if our digital experiences were emotion-aware?
Cambridge University and a Serendipitous Meeting
Building an emotionally intelligent system was the thesis of my research at Cambridge University, where I studied from 2000 to 2005. My PhD advisor Prof. Peter Robinson empowered me to chart my own path and own ideas. His only request: I had to build a working prototype: an algorithm that read and responded to its user’s emotion in real time. Mind you, this was way back when our webcams were HUGE rounded things that captured blurry images at a maximum rate of 5 FPS and smartphones were not even on the horizon.
During the summer of 2004, something happened that changed the course of my career. Ordinarily I would have been back in Egypt, but that summer I was writing my thesis and had decided to skip my trip back home. We were told that MIT Professor Rosalind W. Picard was visiting our lab in Cambridge. She asked to spend 10 minutes with each student just to get a flavor of what we were doing. I was very excited; I had read her book “Affective Computing” in 1998 and got so inspired that I decided to make this my career. Now I was going to meet Roz in person.
Roz and I ended up meeting for the good part of an hour and we hit it off right away. My real-time demo worked like a charm, and we chatted about the many applications of this technology. One in particular excited us both: how this technology could help individuals on the autism spectrum. We applied to the National Science Foundation to build wearable glasses that were social-emotional aids for those with autism . The proposal scored high on potential impact but the reviewers felt it was too ambitious and thus turned it down. With a little bit of persistence and persuasion, we eventually got awarded the grant. That grant provided the funding that brought me over to the U.S. in 2006 as a postdoc at MIT Media Lab.
MIT Media Lab — Demo or Die
The Media Lab is an amazing melting pot of people from different disciplines with one common interest: make the world a better place through technology innovation. It is unique as an academic department in that most of its funding comes from industry, not from the government. These “sponsors” or “members” are invited twice a year to visit the lab and spend a few days exploring what teams of students have been up to. We called it “Demo or Die” week. You had to have something to show.
At the Media Lab, my research was mostly focused on applications of affective computing to autism. For three years, we would demo our autism work, and the sponsors would say “Well,have you considered applying this technology in product testing, or cars, or phones…?” I started compiling a list of these different use cases: Toyota wanted our technology to detect distracted and drowsy drivers, Procter and Gamble wanted to test how people reacted to new shower gel scents, Fox wanted to understand how viewers engage with TV content. The list went on and on.
To Spin Out or Not? The Tipping Point
By October 2008, there were more than 20 sponsors who were interested in our technology. The problem is we did not have enough research students to respond to these requests.
So one winter day in December, Roz and I barged into Frank Moss’ office (He was Director of the Media Lab at the time). Armed with my list, I exclaimed: “We need more research students to meet this interest.” Frank looked at both of us and very quietly said “Nope. You do not need more students. You need to spin out.”
My knee-jerk reaction was “No wait! I am an academic. I am going to be faculty somewhere.” But the more we talked, the more we realized we had a unique opportunity to bring our technology out to the world, in a way that potentially really changes how we do things on a daily basis. For me, that was the tipping point.
As a result of Frank’s encouragement, I was introduced to Andy Palmer. I remembered clearly that we immediately clicked around our “dual bottom line” vision. We at Affectiva really wanted to be a sustainable, thriving business that also did social good, especially around helping individuals who struggled with communication or regulating emotion, and Andy loved that vision. He became our first advisor.
Spinning Out — Steep Learning Curve
Once we made the decision to spin out and do it right, it all happened fast. We signed up for MIT’s Venture Mentoring Service (VMS), an amazing network of mentors and other MIT spin-offs–some like us, very early in their journey, others several years ahead of us.
At VMS, we got assigned a team of 11 mentors who spanned a variety of expertise. Some were venture capitalists, some had a background in finance, intellectual property, legal, marketing, product, or manufacturing. For several months, we’d meet them once a week. We’d pitch to them as if they were a team of investors, and in return, they’d tear our pitch apart. They had answers to all our questions, no matter how mundane. One time, we got an email from an investor who was courting us (who ended up being our first investor). He asked for our “BS”. The only BS we knew of didn’t really make sense. We shot the email over to one of our VMS mentors. Aha: Balance Sheet!
I also discovered that as a foreign researcher I was not allowed to work for a company (not easily anyhow), let alone start one. These issues were very real. Again, through MIT connections, I ended up working with an immigration firm that quickly applied to get me “Extraordinary Alien” status (picture a one-eyed green Monster from Mars.)
It was a steep learning curve, but these early days were the most fun. We articulated our vision, we dreamed big. Roz and I personally invested our own money, which helped us hire our first two full-time employees. Neither of us got paid until the second year.
Because of our MIT Media Lab roots we were very fortunate in that we got a lot of interest from several VCs. It also really helped that we had working prototypes of our technology, and a list of companies who had expressed interest in paying for our technology.
It was now winter 2009 and we were in full swing, pitching to investors. In February 2009, I had my second child. A few weeks later I was back pitching to investors. Often, I’d have Adam with me and I’d leave him in his car seat with an assistant while I walked into the investor’s meeting room to pitch. I’d pray that he didn’t wake up or scream.
We must have pitched to over 20 VCs. The feedback we kept getting over and over was that we needed to focus. We had two core technologies (one hardware, one software) that both had a play in several verticals. Everyone we presented to wanted to know which was our first market. In all truth, we had no idea where our initial market traction was going to come from.
We ended up with multiple term sheets, some had better terms than others. Ultimately we went with one of the more favorable ones, but also one where we felt the investors shared our vision and our values as a team. We used the proceeds to hire our first CEO and build the team.
Now we were rolling. Our next milestone was finding a product market fit.
Which vertical market where we bringing emotional intelligence to first?
More about that in a next blog post.
Originally published at koablog.wordpress.com on October 20, 2015.