From Chat to Collaboration: Trust, Automation and Intelligence
Table of Contents
In our role as market analysts, we are lucky enough to attend some of the most interesting industry events in capital markets and fintech. In that vein, Symphony asked us if we could attend their Innovate conference in New York this past fall and give our unbiased feedback on what was interesting and new, as well as what we thought these innovative ideas and solutions could mean for the market going forward. Rather than a sales pitch, the event was more a gathering of several fintech startups and incumbents that were there to think about how collaboration, automation and smart digital communication could change the market. It felt a lot less Wall Street and a lot more Silicon Valley (which meant we didn’t even need to wear ties).
Here’s what we came away thinking:
Efficiency is Driving the Shift
Fee and margin compression across much of the institutional capital markets has brought the importance of efficiency to the forefront, opening the door for technology to enter and provide solutions. In fact, Greenwich data shows that 78% of institutional investors agree that technology has made financial markets “better and more efficient,” compared to 20% of participants feeling that technology has made financial markets "overly complicated”. NatWest’s SCOUT digital trading assistant is a great example, using natural language processing (NLP) to automate traditional trader-to-client chat.
To that point, asset managers and hedge funds have allowed dealers and exchanges to swallow most of the costs related to upgrading systems and implementing new technological solutions. It was NatWest, not their clients, that poured money into developing SCOUT in an effort to service clients more effectively. Furthermore, execution algorithms and front-end trading systems, for instance, are also largely developed and/or supplied by the sell side.
But even with the billions already spent, the industry has yet to update the myriad systems that allow our efficient markets to operate – a change won’t happen overnight. It is still common for finance, accounting, clearing and settlement systems to run on mainframes – something that sounds crazy to Wall Street’s millennial leaders. Case in point, Greenwich data shows that antiquated processes cost financial services firms over $1.5 billion annually.
Regulatory and compliance requirements have also slowed down technology adoption. Some still worry that putting client data in the cloud opens them up to privacy concerns, leaving even more data and systems running on technology from the 90’s.
Cost-Effective Access to Solutions is Key
The challenge, therefore, is less about developing technology to do complex tasks for financial markets professionals and increasingly about creating those solutions in such a way that they are accessible without multimillion dollar implementations and integrations.
Moving away from systems interacting through paperclips and duct tape is a no-brainer, but doing it in such a way that the cost of change doesn’t cancel out the short-term benefits of the upgrade is critical. This is a big reason why middle- and back-office systems often run on outdated technology, whereas the front office gets the latest and greatest. The ideal path forward is via platforms that put the power back in the hands of the user who, whether tech-savvy or not, is left with the ability to create automation tools tailored specifically to their needs.
Symphony’s SPARC platform, demo’d at the event, shows how bespoke, over-the-counter derivatives negotiations can still be done via chat, but in such a way that true OMS integration and downstream processing becomes possible. The idea of improving a trader’s existing workflow rather than complete disrupting it has proven more successful time and time again. MarketAxess and Tradeweb initially got the market hooked on electronic bond trading by digitizing the then phone-based RFQ process, not by throwing out the old rule book and starting again.
The shift in mindset to user-friendly innovation has accelerated as of late, both via the products fintech firms are delivering and users' expectations of the technology they need to do their jobs. Front-, middle- and back-office professionals all expect to utilize technology at work in the same way they do at home. Mobile-friendly, easy-to-understand interfaces are expected today, for instance. This is true not only for the often discussed digital-native Millennials, but for the Gen Xers in senior management positions fed up with system crashes and the inability to keep up with the office from the road.
Machine Learning Enables Smart Notifications
These user demands are driving fintech providers to focus on creating solutions that are easily customizable and increasingly accessible. And as both markets and technology have become more complicated in the past decade, providing the user with the latest technology innovations in an easy, consumable way is critical.
For instance, smart notifications are increasingly commonplace in new applications. Imagine those constant alerts on your phone being smarter, only interrupting you for things that really matter. These smart notifications allow the user to focus on what is essential and best directs limited resources to where they are needed most. Both Citi and UBS demonstrated their new research applications that do just that, alerting the user to only what it has “learned” the user cares about, rather than everything that falls under a much broader umbrella.
Put another way, it’s about cutting through the noise. Having every email, instant message, price update, and trading alert distract from the other things you’re doing is a fact of life on the trading desk. But with everyone being asked to do more with less, those distractions increasingly create the risk of lost opportunities. Alerts intended to help you not miss anything are now having the opposite effect.
This is where machine learning, and, more specifically, NLP are starting to have a real impact. Turning a trader’s thoughts into computer code that never makes a mistake is tremendously difficult. But solving that problem is becoming more necessary and increasingly possible. The idea is not to replace people but to augment them.
AI/ML Bolsters Communications
Automation in this instance isn’t only about executing trades, but also about communicating with colleagues and clients. NLP can allow interactions that still feel natural, while injecting a level of information and efficiency that a human just couldn’t achieve. For instance, a client’s request for a price on a security via an instant message could be replied to automatically, rather than requiring a trader to cut and paste from their pricing spreadsheet.
Synthesizing the meaning of a statement in an earnings call can also mean more robust risk modeling and faster communications with clients about a market event that is set to happen. So not only is AI-driven automation possible today, it is increasingly user-friendly. Barclays' BARX trader desktop, Goldman's RFQ and J.P. Morgan's BETSI all show how this approach and technology are being used today.
A More Efficient Future
Looking forward, the best applications of artificial intelligence and big data will be those that leave the user unaware of just how sophisticated the technology actually is. Do you really think about how many billion web pages are examined and the AI involved every time you do a Google search? Exactly - neither do we. Traders want to interact with “bots” as if they were human. No one really wants to talk to a machine.
Many financial services applications aren’t there yet, but the last few years have seen tremendous progress toward a much more efficient future. The financial crisis caused many people significant hardship, but it also brought the markets to the incredibly efficient state of today.
