Highlighting critical insights in a deluge of unstructured trading data

Chris Winberg

March 31, 2020

Blog

Given the current unprecedented levels of market volatility, we explain how our NLP Insight Reports can help you navigate the coming months of uncertainty. But first, let me explain how we got here…

GK started out as a trader-voice application, so unusually we also began our journey as a natural language processing (NLP) provider by tackling audio first. This meant that GK had a key differentiator – we could voice-enable your manual workflows.

This was the hardest way to build models – it required thousands of hours of quotes, audio files and SME input, ingesting 250+ accents in the process, but the nett result was a richer directory of GK interpreter models which could be leveraged to voice enable virtually any Financial Markets application. GK had bridged the gap between an “off-the-shelf” general voice transcription service and an NLP provider.

IM chat and email NLP always seemed like a secondary requirement – lots of vendors offer a transcription product which they claim uses “NLP”.  We assumed our expertise in voice would always keep us in front of the market.

However, what was, and still is, a key differentiator in a lot of ways is now table stakes in a GK engagement – our customers all know we can voice enable their apps, but what they really want are our market leading NLP models to process written word. It turns out Siri and Alexa aren’t too great at finding the mid on 2’s 5’s 10’s Euro!

Why? Because we trained our models the hard way, and now they are supremely capable when using them to process text inputs both with and without audio.

This eclectic mix of inputs is what gives us the edge and generates the most insightful reports currently available in Financial Markets.

So, across our text-based NLP workflows, which are most prevalent?

  • Insight Reports – both buy and sell side are finally able to unlock the value in their IM, email and message data, adding value to an archiving process which has for so long just been an expensive box to tick for compliance. Everyone stores it, very few are using it. Customer rankings, sales efficiencies, even quantifying what makes a strong customer relationship outside of just binary liquidity metrics are all now possible.

 

  • CRM dictation – Sales and Trading do NOT do a good job of updating their CRM. These are hugely expensive investments for a firm, even an investment bank – but often they lay empty as shelf-ware. Enabling a GK call bot, or even adding a direct CRM update toggle to an existing bank-owned app can exponentially increase the volume of data added to your CRM. Most importantly, once it’s in there, GK can use its NLP models to extract the important bits and report back to the business.

 

  • Service Desk and Support – these teams are often product-driven silos of activity, inundated with requests and escalations from the customer base. In the same way that we can extract an Equity Derivative price or a Corp Bond from a chat conversation pre-trade though, we can identify it on an inbound enquiry about a settlement which has failed and help you funnel it to the best team to resolve the issue. Call centre technology is finally arriving in Finance, it just needed GK’s NLP engines to filter the jargon to complete the puzzle.

 

  • Best Execution, Trade Reconstruction – the post trade environment is also struggling with a problem which NLP can solve. They receive huge volumes of predictable, structured data across trade confirmations, settlements and payments – the problems arise when they are asked to dive into the unstructured data which surrounds a trade. Regulators obligate firms to present a reconstructed trade in 72 hours, but when a trade can span multiple input channels, often starting on email, continuing on IM and executed by voice, the trail is far from easy to follow. We can recognise product mentions right from their genesis as an expression of interest, through an official RFQ, to execution and settlement.

So how do you quantify a “good” relationship?   Which customer should you call first when the market opens with these unprecedented moves?

GK can help your sales team focus on the highest value decisions in a sea of choices.