GreenKey and the Art of the Bot

Tom Schady

October 19, 2020


The art of bot creation is to hide complexity from the user – presenting a simple, natural interface for knowledge or action.  This is difficult.

“The Law of Conservation of Complexity: Every application must have an inherent amount of irreducible complexity. The only question is who will have to deal with it.” — Tesler’s Law

I was confronted with 3 forms of this law when building our latest Natural Language Processing (NLP) chatbot for Symphony Innovate last week:

  1. the complexity of understanding the user’s command
  2. the complexity of bot security & messaging mechanics
  3. the complexity of NLP on terabytes of sales & trading chat data

Each of these has an elegant solution for the bot creator to push complexity away from their users, and even make their own life easier.

1.  Natural User Commands

Structured slash commands like /help or /getquote are easy for computers to understand but remembering all these puts the cognitive load onto the user.  This is backwards – software is supposed to serve us, not the other way around. Humans have been using natural languages to communicate for at least thousands of years, but only recently has technology caught up.  GreenKey’s focus is on financial language NLP to power workflow automation. By using our SDKs and the GreenKey NLP engine, we are able to build interpreters that not only understand your intent, but also the complex financial instruments you’re referencing. Let GreenKey’s SDKs do the work for you.


2.  Bot Framework

Now that we have our NLP chat interface established, how do we securely and efficiently pass messages to each other? GreenKey’s preferred framework is Symphony – a leader in communication, sharing, and workflow automation for financial markets. Wiring all this up by hand would be an exhaustive process, but Symphony has abstracted all that away behind a rich set of APIs.  GreenKey has been building Symphony bots for years now, but after going through the updated Symphony Developer Certification process, I was struck by how far they’ve advanced in the last year. By creating a true platform, and embracing open-source contributions from major banks (see recent news of Deutche Bank’s contribution, they are managing more and more complexity for the developer.


3.  Sales & Trading chat NLP

We are all inundated by a neverending stream of email, chats, notifications, and alerts. You have a choice: either miss out on some important messages, or spend an inordinate amount of time essentially working for your inboxes. Processing massive amounts of data are what computers were designed to do – but until relatively recently, they were not smart enough to be effective agents. To understand and distill important information is a complex job, one that GreenKey’s NLP engine was designed specifically to do. We ingest massive amounts of financial conversational data, structure this information (including product identification, quote extraction, trade construction), and provide insights and workflow automation. Placing a chatbot interface overtop this rich structured financial data is the perfect solution to capture missed opportunities with minimal effort.


Let GreenKey do the Work

There’s a final way to hide complexity: let the GreenKey team do the work. You can avoid the learning curve and get to market faster by engaging our skilled Professional Services team. We work hand-in-hand with you to understand your users and craft the perfect chatbot powered by GreenKey NLP.

Focusing on your needs and that of your users is hard enough – let us do the rest.

Interested in learning more about GK NLP or our professional services?  Contact