Karan Kashyap
Co-Founder and CEO, Posh Technologies
Age: 26
Industry experience: 3 years
While studying for master’s degrees at MIT, long-time friends Karan Kashyap and Matt McEachern decided to earn some money doing consulting work. As students focusing on artificial intelligence, natural language processing (NLP) and speech recognition, they offered to assist large companies with projects involving this technology. Their efforts paid off. In less than two years, the pair had brought in over $1 million in revenues, including one contract for almost $250,000.
After completing their degrees, Kashyap and McEachern used their earnings to hire some friends and start Posh Technologies in 2018. Rather than working with a variety of companies like they had as consultants, Kashyap and McEachern decided to focus on building an AI and NLP platform for community financial institutions. The joined the DCU FinTech Innovation Center in Boston – the Marlborough-based credit union’s financial technology accelerator program – and now have 40 employees and about 50 banks and credit unions using their product.
Q: What does Posh do?
A: We use conversational AI to try, on one hand, to automate routine customer-service banking interactions that are typically done through calling the call center or going to a branch teller. On the other hand, we’re also collecting a lot of data on how people are asking questions. When a customer in a bank comes into the branch or calls the call center, they’re basically bringing their complaints, their frustrations, what they wish the bank could do better. We’re collecting all this information and giving it back to the bank with insights about how to make operations better and make things easier for customers. We’re doing that across digital and voice channels. On the digital side, it’s through online chat, mobile chat, SMS [texting]. On the voice side, people calling into the call center can interact with our AI before they get escalated to an agent, if needed.
Q: How did you come up with the idea for Posh?
A: While doing consulting work, there was an “ah-ha” moment for us: This technology is so new, but companies and enterprises are so excited about it. That led us to ultimately double down on building our own platform. We realized that to build good AI, you have to be really focused. It’s hard enough to build AI for banking; we can’t build AI for 20 different industries at the same time. Matt and I were actually both members of MIT Federal Credit Union, and so we were exposed to community financial institutions. Instead of going after large banks off the bat, we decided to work with some community FIs and sell them on the same thing that Bank of America and Capital One had been doing. Bank of America had released Erica, which is their conversational agent. They have millions of people using that now, and at the time it was pretty cutting-edge. But think about the community banks, community credit unions, mid-market institutions – there’s no way they can do that in-house. They have to rely on vendors, and who’s the vendor in the market that can do it? That’s where Posh would come in.
Q: Why did you name the company Posh?
A: We’re linguists by background – we did computational linguistics and NLP – and the name Posh was really interesting to us. First of all, it’s a short word, and it’s memorable and catchy. Also, this is probably not true a true etymology of the word, but myth has it that the word “posh” stands for “port out, starboard home,” which is a reference to when ships were sailing between England and India in the 1800s for the spice trade. The posh cabins got the most sunlight and had the more desirable locations on the boat – on the way out, it was the port side of the ship, and the way back it was the starboard side. I really like the fact that we’re a company doing language AI, and here’s a word that has this really cool backstory. Also, starting a company is like a voyage, and so we liked that the word had a reference to going on this big journey.
Q: How are you improving on previous products involving chatbots and speech recognition?
A: We’re in a really interesting period with AI and, in particular, with natural language and speech, because a lot of the major breakthroughs that have happened in this industry have really only happened since 2019. As long as I have WiFi and a laptop, I can now procure these very powerful machines in the cloud and then deploy very powerful AI models that have only really come out in the last couple years.
A lot of the platforms we’ve seen in the past have attempted to do AI technology like this, but they were never really focusing on just picking a small subset of use cases and doing those really well. We are very specifically targeting banking and financial services, and it gives us the ability to collect a lot of very specific data that we can use to build our technology model. I think that’s what’s led to us succeeding with our customers so far. We’re not trying to take data from a health care customer and then try to use that data to make our banking product better.
Q: This year you established a credit union service organization (CUSO) with several credit unions. How did that come about?
A: We focus on serving both credit and banks. As we were working with a number of credit unions, they were very interested in having some involvement in helping us build a roadmap of our products for their market. And at the same time, we were looking at raise some funding. It wasn’t a very large amount of money. I think we raised just shy of $4 million doing this. It was a very strategic move for us in that sense, and we’re doing the same thing on the bank side.
Q: What role did the DCU FinTech Innovation Center have in Posh’s development?
A: We joined the center in fall 2018. That’s been really key for us because we had a chance to really get to know not only the team at DCU – and they’ve gone on to become both a customer and an investor in our business – but through DCU’s relationships with other credit unions and other banks, we had a lot of exposure to the industry. I think that fast-tracked a lot of our learnings and fast-tracked access to pilots and customer engagements. It allowed us to just be plugged into the fintech scene much faster. We been using the center as our headquarters ever since.
Q: What are your long-term goals for Posh?
A: The reason we started this company in the first place is that we really want to make an impact on the world of AI and NLP and speech. Getting customers and growing the business – I think that’s really awesome. But the thing that makes my heart happier, that I hope will happen from just a life-goal standpoint, is to really move the needle on AI and enable breakthroughs and enable really cool things that didn’t exist before in this field.
Kashyap’s Five Favorite Foreign Cities
- Rio de Janeiro, Brazil
- Agra, India
- Tokyo, Japan
- Rome, Italy
- Versailles, France
