interviews

Nat Cartwright, Finn.ai

Elyssa Macfarlane
Nat Cartwright, Finn.ai

I’ve found that it isn’t unusual for people to underestimate the complexity behind building a chat bot. But designing the flow of a conversation, regardless of whether it’s one that is b2b or b2c, is convoluted in terms of design and the need to personify the technology. In other words, it’s difficult from a UX, development, and marketing standpoint.

This is why, when I initially saw Nat speak as the co-founder and COO of Finn AI at a Machine Learning forum at science world, I was immediately intrigued by both her and the fintech’s business model. Finn works with banks around the world to advance their technological capabilities by allowing their customers to communicate more directly with a chat bot. Customers can simply text their questions to the bot and receive an immediate response without having to deal with a customer service line.

The business has caught on with banks in a number of countries, which led to it being named the top start up on the most recent season of BNN’s Disruptors.

Nat has a calmness to her that makes her very easy to talk to. She is approachable, yet direct. After running into each other a couple more times and discussing the ins and outs of Finn AI, I asked her to do an interview on MoT to showcase how the business has developed to where it is today.

Interview

Tell me a bit about yourself and you found yourself working in the tech industry:

Ha – good questions! I never thought that I would be in tech industry. My first career was in international development working for the UN in Geneva where I managed a 250-million- dollar investment portfolio for HIV, TB and Malaria programs in low income countries. After my MBA, I wanted to move back to my home town of Vancouver - where you can’t really work in international development. My now business partner asked me if I’d do something in FinTech with him and, given that tech is the future, I figured I’d give it a shot. That was three years ago and we’re still going strong.

When you and Finn ai cofounder Jake Tyler initially got together to start a business, you started out by founding the P2P payments platform Payso, which then transformed to Finn ai. Can you describe the story of how you decided to pivot the business model?

Jake, Guru – our third business partner - and I did really well with our consumer business. We had 10,000 users and transacted over a million dollars in small payments for Canadians. Unfortunately, because of the payment rails that we were using, we were limited to the Canadian market and the margins that we had in a market the size of Canada just didn’t add up for venture investors. ATB Financial, our first partner, was interested in white-labelling our product and that’s how we first moved to B2B. The pivot was really just based on following demand.

We’re living in an age where bots are becoming evermore prevalent, yet many struggle to master their execution. For example, Facebook recently scaled back their bot-chat efforts due to a 70% failure rate. In your opinion, what are some of the major considerations one must take in order to design something as complex as a chatbot?

It’s true. You have to meet the technology where it is at and, for the moment, it’s an early manifestation of bots in the North American market. We do things like design with lots of quick reply and menu options to compliment the Natural Language Processing when it is still learning. We also integrate directly with our client’s customer support so that the user can always connect to a human if needed. Last, we do small, controlled roll outs to make sure that we’re happy with the user experience. It’s early days for bots but just like the web and apps before them, they’re going to become ubiquitous in no time. The banks that we work with know that they have to be learning and getting experience with bots now to not fall behind.

Have you faced any challenges in approaching banks across the globe with new technologies?

I’m speaking to you after just getting back from a whirl wind trip to Africa and Asia where we had really successful meetings so maybe my answer is a bit skewed but, generally speaking, we’ve seen way more demand than we originally anticipated. We estimated that it would take an average 9 months to build partnerships with banks and it is currently taking on average 6 – 8 weeks. Bank-FinTech partnerships can be a great win-win and we’ve found that there’s a strong alignment with our bank partners so - so far so good.

In some of our past conversations you’ve mentioned how Finn ai is starting to experience rapid growth. What does this look like for the company and how is the team planning to scale over the next year?

The last year has been extremely hectic in the best way possible. We’re currently working with partners across 4 continents and are starting to scale with some of them. We’re raising our Series A and plan to double the team by the end of 2017. The growth will mostly come in the engineering and data science side.

Finally, what are some of the discrepancies that you’ve noticed in Canada’s banking system versus those of other countries that you’ve worked with when it comes to adopting innovations like A.I.?

Hmm – that’s an interesting one. The most innovative countries are the ones that have leap frogged some technologies. For example, the first time that I texted money was when I lived in Kenya in 2006. Generally speaking, we often find that mid-sized banks see technology as their opportunity to gain market share and tend to move fast with us. We also find that Asia, Africa and Latin America are faster adopters of our technology. That being said, we’ve yet to speak to a bank who doesn’t have bots in some form on their roadmap.