BNP Paribas has unveiled its newly revamped FX algorithmic trading platform following a major upgrade, which includes the launch of interactive algorithms and real-time analytics delivered via an artificial intelligence-based digital trading assistant known as ‘ALiX’.
ALiX, which BNP Paribas said is the FX industry’s first digital trading assistant, will deliver content and running commentary on execution to traders using the bank’s modernised Cortex LIVE platform. As market participants continue to battle with the onslaught of data volumes, ALiX will digest market information, presenting tangible options to improve algo execution mid-flight, effectively giving BNP Paribas’s FX algorithms a voice.
“We’ve taken a leaf out of Silicon Valley’s book with ALiX,” says Asif Razaq, global head of automated client execution (ACE) at BNP Paribas. “Similar to the concept of Alexa or Siri, ALiX was designed to focus on financial markets. These assistants are solving multiple workflow problems. ALiX is designed to be your personal trading assistant, where it will monitor multiple live market data feeds within our real-time analytics portal.
“ALiX will also intelligently inform the client of any key market events and present context relevant options to choose from, ultimately providing the client with an intelligent roadmap of what to do next.”
Alongside ALiX, BNP Paribas has launched its Cortex LIVE platform with a real-time market analytics portal, known as Insight LIVE, which the digital trading assistant will use to deliver market content, data and intelligence to traders.
Insight LIVE builds on the FX pre- and post-trade transaction cost analysis (TCA) that BNP Paribas already offers clients through its Insight portal, however, the Insight LIVE data has been expanded to include real-time information to bridge the gap between the pre- and the post-trade TCA.
Navigating the execution flight path
BNP Paribas launched its FX single-dealer platform, Cortex, in 2013 at a time when FX markets were undergoing a major evolution following the introduction of algorithmic trading and electronic trading platforms. FX markets first adopted the basic TWAP and VWAP algo trading strategies from equities, but as the arrival of multiple new FX execution platforms caused fragmentation in liquidity, second generation algorithms were rapidly developed to aggregate that liquidity across numerous venues.
Although slightly late to the party, BNP Paribas says it pioneered the third generation of FX algorithms. Razaq, who has been part of the FX algorithmic trading team at BNP Paribas for almost a decade, was tasked with developing the bank’s expedition into FX algorithms and soon developed what would become the third generation algo: adaptive algorithms.
Razaq says he deployed basic forms of AI technology and real-time insights so that the algos can self-adjust when working a trade. At that time, BNP Paribas launched two adaptive algorithms, Chameleon and Viper, providing clients with both aggressive and stealthy algo trading strategies when the Cortex platform first went live.
“Execution algorithms have grown in use in the FX space, whereas in the equities space, they’ve existed for a very long time,” Razaq adds. “FX algorithms were more difficult to build due to a lack of transparent data and fragmented liquidity. As we saw the electronification of the FX market take place, we started to develop models to gather market intelligence and were able to build algos that adapt in real-time.
“It was a greenfield project in terms of having a blank slate, and this is where I really wanted to utilise my expertise in artificial intelligence. Can I build something using basic AI techniques that will give us an edge, to improve on what the market is doing with the first and second generation algos? We also wanted to simplify the algo journey, because clients think of algos and see complexity. Sim plicity was key. When BNP Paribas launched its algo platform, one of the key objectives I set myself was not to overcomplicate the algo selection process for the client.”
Simplicity has remained a key part of Cortex following the revamp, and rather than building several different algorithms, BNP Paribas focused on enhancing its existing Chameleon and Viper algo strategies. While clients found value in the self-piloting and adaptive algorithms BNP Paribas had developed, the bank says many wanted to be more involved in the algo execution process. To meet this demand, BNP Paribas has introduced what it is referring to as the ‘fourth generation’ of FX algorithms: interactive algos.
The interactive algorithms provide a view of the data it is processing and reacting to from the Insight LIVE market data hub, allowing clients to consider various options for order execution. But, rather than running the algo on autopilot, as has traditionally been the case, traders are handed the controls to turn off autopilot and navigate the execution path, based on real-time analytics delivered via ALiX.
The four pillars
The FX algo trading team at BNP Paribas focused on four pillars when building Cortex LIVE; user experience, market analytics, integration and intelligence. For user experience and integration, market participants have a multitude of single-dealer platforms to choose from, all of which are competing aggressively for client desktop real-estate. Expert user interface developers were brought in by BNP Paribas to examine the Cortex platform and map out the user’s journey to identify where that workflow could be streamlined or simplified.
Simultaneously on the client side, firms are upgrading or implementing digital infrastructures to automate their own workflow, so BNP Paribas says it is imperative that Cortex LIVE can be integrated within the whole FX ecosystem and with client workflows. In order to achieve this, BNP Paribas teamed up with financial markets operating systems specialist and industry disruptor, OpenFin.
“Cortex Live is the first external platform that BNP Paribas has put on OpenFin, but the bank has been a supporter of OpenFin in terms of internal use cases for many years,” explains Adam Toms, CEO of OpenFin Europe. “With the development of the new major integration process platform, BNP Paribas adopted some of the best components that OpenFin has to offer, including the FDC3 interoperability standards and user experience. It is a big step forward and, frankly, any organisation moving forward with interoperability is going to be a market leading vendor.
“It’s encouraging to see, given my background and that FX has historically been an asset class with a lower level of transparency, that BNP Paribas has made a real push with transparency, with real-time TCA and in-depth venue analysis. That is a powerful differentiation.”
Onboarding clients, particularly the more traditional asset managers, to a new platform can be a headache in terms of legal paperwork and getting the system onto client desktops. To address this, BNP Paribas has ensured that the Cortex LIVE platform is accessible and downloadable via the app store on Bloomberg, which means that clients can download the platform within the Bloomberg ecosystem without having to go through a major integration process. For those that often struggle to keep track of usernames and passwords, Cortex LIVE clients will not even have to log in, as Bloomberg is able to authenticate a client’s system for easy access.
“Cortex LIVE being available via Bloomberg removes that integration barrier, and that’s why we expect the adoption rate of clients consuming Cortex LIVE to be significant,” says Nick Hamilton, head of EMEA eFX sales at BNP Paribas. “It streamlines and simplifies workflow. Clients just have to download the platform from the Bloomberg app store and they’ll have access to all of the enriched functionality.”
With the OpenFin integration, BNP Paribas claims it is ‘future-proofing’ the new platform and, eventually, systems on the client side will be able to communicate with BNP Paribas’ digital trading assistant to create a fully automated and integrated trading platform. Although the FX industry is not quite at this stage yet technologically speaking, BNP Paribas has laid the foundations to take ALiX and Cortex LIVE a step further.
In the early days, ALiX will be fed with basic responses, but the machine will self-learn and be able to respond, building a rapport with the client as they ask the digital trading assistant more questions. As ALiX is given more data its intelligence will grow and broaden its remit from guiding clients not just through their FX algo trading execution, but for all FX activities.
“The vision is that we will broaden the ALiX concept as a holistic solution across the FX execution environment for all of your activities as a user on Cortex,” says Joe Nash, digital FX chief operating officer at BNP Paribas. “Then it becomes more than the execution control point for the algo business, it will be a digital trading assistant for everything you’re doing in the FX space. Pushing content, trade ideas, figuring out when you expect to put a particular trade on and provide pricing to you in that regard, there’s a whole sphere of functionality that we can bring to the FX space through ALiX.”
In the long-term, BNP Paribas aims to take the new technologies in the form of ALiX and the new functionality and analytics on Cortex LIVE to other financial instruments and asset classes. Cortex LIVE, ALiX, as well as interactive algos Chameleon and Viper, could soon embark on a journey into equities or fixed income. BNP Paribas says this expansion is being considered amid a trend towards multi-asset trading in asset management.
“We are now looking to extend our award-winning algo execution service to other asset classes such as fixed income, futures and so on. This is against the backdrop of the buy-side consolidating their activities and forming multi-asset trading desks. Eventually, we want ALiX to become that multi-asset BNP Paribas trading assistant,” Razaq concludes.
This article was originally published in the TradeTECH FX Daily magazine, produced by The TRADE, which was distributed to attendees of this year’s TradeTech FX Europe conference.