Deutsche Bank (DB) is looking to automate parts of the price-discovery process for over-the-counter (OTC) trades.

It has created a bot that is able to monitor chat rooms to track and distribute request-for-quotes (RFQs), collate responses, and obtain price information for accurate, near-real-time information. It also supports post-trade dispute resolution by automatically creating chat rooms with the relevant operation teams to facilitate resolving trades quickly. 

The bot came off the back of a hackathon, which was co-hosted by Deutsche Bank and the messaging platform provider, Symphony, in Singapore in June, ahead of Symphony’s Innovate 2019 conference.

Teams from DB and nine other companies, including clients and competitors, participated in the hackathon. The Deep Waters team from DB won the award for most impactful business automation development at the Singapore hackathon. 

Chris Bezuidenhout, CIO for APAC for the corporate and investment bank at Deutsche Bank, tells WatersTechnology that the Deep Waters team successfully delivered a unique breakthrough during the hackathon, which involved automating new term sheets for total return swaps. In most cases, new term sheets are usually produced with certain information given, manually, to a data delivery system and only then passed on to a client. 

“By virtue of that information, you would generate documentation that otherwise would be sent to an operations team downstream,” Bezuidenhout says. “Currently, there’s a bit of lag between that creation.  There would be manual intervention and the potential for it to go wrong, then the review process of that would need to kick off after that.”

The bot automates the entire workflow and allows RFQs for OTC trades to be pushed out immediately. Bezuidenhout says very little incremental testing is needed before the bot is put into live production within six months. 

“Instead of using email, the solution moves it to an interactive chat,” he adds. “Instead of using different people to do price discovery, we can automate quoting. Essentially, it’s the integration of quoting tools you would use for a specific financial product, and instead of a user having to look at three different systems, you would [see] the basic parameters that you’ve entered into the chat.”

Then, rather than sending that information manually in an email to the operations team, which would create the documentation, the bot will generate the template from the data in the chat. 

Bezuidenhout adds that this will save at least 30 to 45 minutes from the normal manual workflow process.

Working Together

Moving forward, Deutsche Bank wants to work with its clients and even competitors more directly, which could include developing additional bots to streamline transaction processes. 

At the hackathon, DB provided a few use cases that might be useful to explore in the future, such as hyper-connectivity between firms and automation of workflows, to set the table for the next stage. 

Participating teams were then left to decide what use case to apply. “The challenge was to see how far teams could progress a particular idea, with the support we provided on the day on-site,” Bezuidenhout says. “The judging that we did at the end of the day was, essentially, ‘How close to production did they get?’ There was a reward for prizes based on if we can actually move this to production quickly based on a single day’s worth of development.”

Deutsche Bank looks to replicate similar hackathons every few months. “The aim is to connect better when we do transactions with each other in both corporate and institutional client journeys,” he adds.

Luc Meriochaud, global head of innovation at DB, says the proof of usefulness is in how hackathons like this are followed up. “All the winners we had from the Berlin hackathon, we invited to the Berlin Innovation Lab to spend a couple of weeks working on their products to get them to the production stage, one of which Deutsche Bank actually invested in,” he says. 

In 2016, DB hosted the API/Open hackathon in Berlin, in which Frankfurt-based company dwins—a company founded by real twins Alexander and Benjamin Michel—won with their Finanzguru prototype for personal finance. The Finanzguru app, which has more than 200,000 users today, gives financial advice based on transaction history. DB took a 25% stake in the company in 2017.