Liquidnet, the technology-driven agency execution specialist, has partnered with AI startup Boltzbit to enhance its fixed income primary markets workflow, using generative AI (GenAI) technology to reduce the time required to process unstructured deal data and prepare bonds for trading, by 90%. The collaboration accelerates the processing and display of newly announced bond deals by leveraging Boltzbit’s advanced AI machine learning solutions and custom workflow model.
By integrating Boltzbit’s AI technology, Liquidnet can now offer members and partner syndicate banks faster access to trading and data distribution, processing and displaying bond deals at a rate significantly faster than its previous parsing technology. This ensures that bonds are quickly available through the company’s deal announcement dashboard and new issue order book.
“This partnership improves the speed at which we can process messages, create, and then send structured data directly to our clients, which in turn allows them to quickly populate their OMS and prepare for trading,” says Mark Russell, Head of Fixed Income Strategy at Liquidnet, in conversation with TradingTech Insight. “The quicker we can do this, the better it is for those clients. Beyond this, the clients of our new issue Trading Platform (grey market) benefit as we are able to launch the new bonds on the screen earlier, giving those clients earlier access and more time to trade. More trading time on our visible trading platform means more transparent data points, which is very useful for the syndicates and issuers as they get a view as to what is going on in the market.
“Structuring the bond data is not done in a single step, during the bond creation process we need to interpret the market chat, back and forth messaging, that drives the final structure of the bond,” explains Russell. “Our system needs to be able to capture and update any changes to the meta-data, such as coupons, issuers, benchmark, maturity etc. that describe the bond and feed those changes into the trading platform and other information platforms.”
He continues: “We’ve automated this process extensively with our partners at Boltzbit, creating a tool that handles the heavy lifting of structuring this data into a comprehensible bond format. Our partnership with Boltzbit is focused on speeding up and enhancing accuracy, bypassing traditional parsing tools and leveraging artificial intelligence instead.”
Boltzbit’s GenAI technology utilises the data captured from messages exchanged across various mechanisms and channels to create a large language model (LLM) that transforms the information into a structured and usable format.
“This process might seem simple, but it was actually extremely challenging,” explains Dr Yichuan Zhang, CEO and co-founder at Boltzbit. “Firstly, it involves very complex business processes. It’s not just about parsing one email; understanding the context of the conversations and the associated business processes is essential. Secondly, this is a highly specific solution, requiring the model to be extremely accurate and to follow the precise logic of the business flow around new issues. Finally, the solution needed to be highly secure and deployed in a way that allowed Liquidnet full control.”
Since the launch of its primary markets offering in 2022, Liquidnet has achieved record trading volumes in its new issue order book and increased participation from over 35 European syndicate banks, highlighting the company’s commitment to modernising primary markets and delivering substantial value to clients and the industry.
In addition to partnering with Boltzbit, Liquidnet has previously collaborated with NowCM and BondAuction, reinforcing its dedication to fostering efficiencies and connectivity for investors, banks, and issuers through strategic partnerships.
Subscribe to our newsletter
