Trading Technologies announced yesterday the acquisition of Neurensic, an artificial intelligence regulation and compliance firm (terms were not disclosed). The deal brings Neurensic’s AI technology into TT and provides a new compliance platform for the firm. But the potential for the technology goes well beyond that. TT’s CEO Rick Lane spoke with JLN’s Jim Kharouf about where this technology is going next.
Q: When you look at the AI/machine learning space, where is this industry going and where does it fit in for TT?
A: Now we have another tool in our toolbelt. We’ve been providing a new, unprecedented level of access to data for our customers for a couple of years. This will be an empowering tool for them and for us. For instance, looking at alerting. This compliance and trade surveillance is largely defined by alerting compliance or risk officers when bad behavior is occurring. But a trader who is using one of our APIs or something like our algo design lab, or even manually trading the markets, would really benefit from having a system in place that alerts them to trading opportunities. I think that is one obvious place we’re going to take this next. Alpha generation – automatically and manually identify trading opportunities – is another. But it’s empowering to have this in-house and we’ll see where we go from here.
Q: To me, that’s the crux of this deal. It’s not just that you’ve brought in a compliance tool and can fold it into the current TT platform, but people are really anxious to see where AI and machine learning is going to go from here, not to mention trying to figure out how much AI is going to cost to develop and keep up. Will this help democratize AI somewhat?
A: One thing TT has done really well over 20 plus years is democratize and access otherwise inaccessible technology. Even if it is not there today – although the Neurensic installation is not a terribly heavy lift – this will really be empowering, with access to bleeding-edge technology. The one thing we’ve seen since we launched the TT platform (in 2015) and grew the user base, is that all our customers are struggling with how to manage their data, both from a compliance perspective and a trading perspective. Now that they have access to their data at their fingertips, how do we turn that into an opportunity? This technology is really there to address our partners’ challenges in an ever evolving data landscape. This implementation will address compliance concerns but it will certainly go from there.
Q: Do you have an example of what that might look like – how you manage the data?
A: Our customer base with the TT platform was and still is required to store every trade transaction that happened for seven years. Ten or 15 years ago, that may not have been terribly onerous on them. But today, even a moderately active trader can create tens of millions of orders in a single day. This has become a great burden to our partners. This storage requirement makes it almost an untenable situation. So today they have access to that data on TT and can store it in a way that really puts it at their fingertips. If you are called upon by an auditor to see what an individual trader was doing two and a half years ago on a single day, within a 20-minute time window, the answer to that is yours in a second or two, whereas before it was weeks or months before you found the right box of tapes in the right closet. The eyes are opening for our clients in seeing the level of access to that level of data.
That said, clients are struggling. Yes, it might make it easier to answer a compliance audit request. But being able to manage that data to make informed trade decisions going forward, or finding bad behavior of a trader or trading firm, that’s where they are still struggling. That is because it’s a really, really hard problem. We’re talking about billions of rows of data potentially, and trying to glean some insight out of that is not easy to solve. One technology that this industry and the world sees that can help is machine learning. It can help find anomalies and opportunities. And that is game changing here. The first offering here will be trade surveillance but we do think we can take this in other directions.
Q: How long will it take for TT to integrate this and offer it to customers?
A: The integrated version will be done in six months. That said, Neurensic still has an ongoing business and will continue to remain open for business.
Q: Is there much common architecture or software between TT and Neurensic’s platforms?
A: The Neurensic platform ingests a bunch of audit-level data and marries that with market data for the same period of time, and spits out a report based on that. Where this will fundamentally change, when it’s fully integrated into the TT platform, is simply where does the Neurensic “brain” get the data. It’s no longer downloading and FTPing (File Transfer Protocol) files from the server, putting it in the right format, putting it in the right place. It is marrying the Neurensic brain with our transactional and market data that makes this a seamless and end-to-end experience for the user. So in the same way a user can pull up his or her audit trail from two years ago, they can also go look at that same day’s trading activity, pull in the equivalent market data and spit out a compliance report without any manual file transfer or integration.
Q: Are you looking at more acquisitions and if so, what are you looking for?
A: As we see solutions and platforms that marry well with our clients’ data, and we can offer services on top of that, the door is still is open. The TT platform was not an end in and of itself. We set out to build a platform that was unparalleled in every meaningful metric. But what we really set out to do was build a platform that would offer unprecedented access to data for our partners, the trading community, and capital markets participants as a whole, and in a way that is not just warehousing.
So with solutions like Neurensic, we’re always looking for those types of opportunities, and to build value-add services around those capabilities in-house. One example is MIFID II, which is mandating a certain level of algo approval and algo management capabilities for our partners. We can build out a pretty compelling algo management and approval workflow because our customers have access to all of their data and market data in a way they simply couldn’t get on another trading platform, even outside the trade execution space.