How Spiky.ai optimizes employee experiences
Boston investment partner Alex Marley recently led Dorm Room Fund’s investment in Spiky.ai, a machine learning startup on a mission to make online meetings more effective. Founded by Furkan Eris and Burak Aksar, Spiky offers a daily pulse into the engagement and emotion levels of companies by analyzing videos collected from video conferencing platforms. Leveraging artificial intelligence and machine learning, Spiky provides companies with personalized and actionable insights that can help them become more engaging and efficient. With Spiky, Burak and Furkan are building algorithms that can help optimize employee interactions and enhance employee experience.
We sat down with Co-Founder Furkan Eris to dive into what Spiky does, how it works, and the team behind it.
How did the two of you meet?
When we met, I was doing a Ph.D. at Boston University and Burak was doing a research internship at Boston University. We were both passionate about startups and started meeting for weekly brainstorming sessions where we would discuss projects focused on our interests: human empowerment and potential actualization. We are actually both Turkish, but met in Boston, believe it or not.
I know nothing about machine learning. Explain to me how Spiky works.
We use 19 different machine learning models to track metrics during a meeting. For example, if someone is talking, were they talking in an objective (facts-based) or subjective (emotions-based) manner? Vocally, were they energetic or monotonic? Were they asking questions? What did their visual emotions look like? Did they look happy or frustrated?
We also have a variety of detention and interaction metrics. The detention features can tell us if someone is looking at the same place on the screen the whole time, at different places, or away from the screen entirely. The interaction features provide insight into who is interacting with whom. For instance, when someone is talking, is it the same person replying or a variety of people? Is it always the same person talking or more collaborative? We put all these metrics together to correlate the different interactions between people. From there, we try to figure out the ideal way we can communicate with each other.
How did you come up with the concept of Spiky?
We thought the EdTech industry was ripe for disruption and saw that one-size-fits-all models were not working. Covid began at the same time we started building something, and suddenly professors and teachers were having major issues understanding what was going on in their meetings. Beyond five people, research has shown that people cannot effectively track other people, and typically they tend to look at themselves when they are talking. We discovered this problem and tried to address it using artificial intelligence. Around this point, as we began pitching at different competitions, companies asked if we could apply the tech to training programs. After some initial success with this, companies then asked if they could get metrics on all their meetings. Over time, we started expanding our market and becoming the engagement analytics infrastructure for these companies, and that is where we are now: providing a range of analytics for a range of meetings.
Who is using Spiky right now?
We have a wide variety of companies. Currently, around 100 companies have signed up for our pilot, meaning they want to try the tool out. We launched our platform about a month and a half ago. Before that, we were granting API access to some of our customers. This meant that they could use the algorithms but that there was no interface. We were able to validate the algorithms in doing so. Right now, two of our customers include a 200-school chain in Turkey and a large corporation in Holland building LED fixtures and lights.
Over time, we decided that the partnership model providing API access is not the model we want to go with because it is not very saleable. This is why we built the platform. For the platform, we have companies that have around 50 employees and companies that have 10,000+.
Why would someone choose to use Spiky?
The fundamental first question we need to answer is are online meetings here to stay. We believe the answer is ‘yes.’ Right now, surveys are the primary tool being used in lieu of Spiky. These are typically done every six to 12 months. This is the first major issue with them. Our A.I. and the emotional metrics that we’re tracking are much more timely and immediate.
Most tools tend to track much more rigorous mathematical metrics. Things such as hours of work, lines of code written, and emails sent. These metrics start pushing for longer hours and make people miserable. We’re pushing for the opposite: tracking emotional metrics and helping people better integrate with their teams and communicate more inclusively with their teams will increase efficiency. We are trying to get companies to work smarter — not necessarily harder.