Interview: Anthony Tassone, CEO, GreenKey Technologies

Anthony Tassone is the Founder and CEO of GreenKey Technologies (GK), which develops natural language processing technology that structures voice and chat data for banks and trading firms. GK is based in Chicago with offices in New York and London.  

What are you most looking forward to at the AI & Data Science in Trading conference?

I’m curious to hear how others are applying NLP, where it is making the most impact and where it can improve. We’ve found a growing appetite within OTC sales and trading for using NLP to automate tedious tasks and unlock client insights to generate more revenue. Banks can’t keep up with the amount of unstructured data bombarding their desks and leveraging NLP is the best solution. I’m also interested in meeting like-minded entrepreneurs and data scientists who are innovating new solutions, and listening to their points of view. If I can offer any guidance to help them navigate the world of finance, I will.


What do you think are the biggest challenges facing banks’ adoption of data science in 2019? 

As far as challenges go, we’re reaching an inflection point where the industry at large is deciding what technology is best developed in a proprietary fashion inside individual banks, and what is best outsourced as a common utility. Vendor-developed solutions like GK’s are increasingly appealing not only given reasonable costs and short time frames for implementation, but also because they are recognized as best-in-class products that only dedicated specialization can offer.  

In my opinion, though the low cost and ease of implementing and operating vendor-provided solutions in training environments is obvious, the legacy mindset of legal and compliance teams remains the single biggest hurdle to implementing data science within financial markets. Data privacy and data mining are seen as at odds with one another and they needn’t be. Nothing short of an entire overhaul of legal and compliance processes is needed in order for banks to be able to leverage the technologies of the future.   

Looking ahead a year from now, how do you see the structure of your market changing?

First, automated trade and quote reconstruction capabilities will progress quickly. OTC traders want to know why and how a completed trade happened, and why and how a missed one didn’t. Quotes are just as important as trades to understanding desk performance, so their capture and analysis is critical to understanding execution quality and market structure efficiencies. Second, analyzing client interactions and touchpoints will separate the best-performing sales traders from the rest. Maximizing trade opportunities for each client could mean changes in trading behavior such as being able to recommend alternative products with similar characteristics to the original request for quote, but that are trading at a discount.   

I also see finance firms moving away from purely transaction-based revenue models. Customer data will be monetized to a new degree, which means cost savings for end users but an increased focus on privacy and security concerns. 

What are the most important factors in selecting a new technology vendor? 

Go into the selection process prepared to ask a lot of questions. You’re looking for a partner and a team that can grow and evolve to solve many of your problems. However, be careful because few vendors can deliver more than one high quality solution and service, and too many entrepreneurs say they can do too much. You’ll want to know with laser precision exactly what your vendor does that differentiates it from the rest of the market, and then you’ll want to deploy that acumen across as many problems as possible.  

Additionally, try to understand the potential partner’s culture as much as possible. People often confuse lots of activity with producing value, but it doesn’t always line up that way. Look for partners who don’t waste your time and are prepared for meetings and deadlines. They will be excellent project managers for you. To that end, I recommend getting to know solutions provider teams out of the office. In a social setting you can better understand their perspectives and team values, and most importantly identify if any miscommunication exists. 

Can you share an example of how your system has been used by a new customer?

Our customers come to us with a need to “voice-enable” a task or to leverage the same NLP engine to “unlock” their audio and chat data, i.e., look for client insights during the trade process. To understand why a trade did or did not happen requires accounting for quotes noted in phone conversations, chat windows and email. For example, our customers often want to extract and define information shared on their Bloomberg terminals, including quote sizes and response times. We help them build a plan to improve their data resources and then automate the process. After implementing GK technology tailored specifically for each client, we help them learn how to work with the data in real time - which usually requires us to integrate into a Bot. 

What can be done about the talent war in AI and machine learning and how do you handle this in your organisation?

In my view, the differentiating factors aren’t dollars and cents, but culture and purpose. We have a huge focus on culture at GK that includes expectations of working hard and accountability, but also a priority on life balance like the ability to work from home. We also require excellent communication discipline. We believe most meetings are a waste of time and that a lot of problems can be solved over Slack or email. If a phone call or meeting is absolutely required, then we do a lot of work prior to identify and frame gaps so meetings run efficiently with follow-ups and deliverable tracking. Data scientists and other high performers do not want their time wasted in meetings when they could be producing product. But above all, people want to work with other nice, smart people on a problem that matters. In addition to our trading solutions, GK develops technology for emergency services and public safety including the digitization of 911 calls and police field needs. Tackling these challenges and their life-or-death implications brings meaning and purpose to what is otherwise an extremely technical project.

What are the top qualities or skills a data scientist should be able to exhibit?

There are many capable data scientists out there with expert technical skills, but the main qualities I look for go beyond intellect. Honing the ability to communicate simply and effectively, and without ego, cannot be overstated. If you can’t describe your ideas in layman’s terms, it will be hard to make their full potential a reality. Similarly, I look for candidates who aren’t afraid to highlight a bad idea and replace it with a better one. Respectfully disagreeing on important issues is an absolutely essential quality in a team member. I believe having decisions regularly tested makes the entire company stronger and I remind our employees to challenge assumptions. If your manager and you are not getting this level of feedback, you may be operating in an environment of fear, or worse, apathy. 

What is your biggest professional achievement to date?

I’m most proud of my management team at GK and their rapid professional and personal growth. I am really fortunate to have them. They’ve scouted and hired our world class data science and engineering groups and constantly aligned them with our strategic goals and mission. The GK company culture we’ve built together is an immense source of pride. Over time, our GK brand has become the product we sell. We are excited for the continued acceleration of NLP in financial services and public safety, and look forward to what’s ahead for GK and these industries.

To view when Anthony is speaking at AI & Data Science in Trading, go to: