What’s the catalyst for ISDA digitization

In 2017, when Eigen was only 15 people strong, we won our first-ever production customer and use case. The customer was Goldman Sachs, and the use case was to digitize all of its QFCs (including ISDAs, GMRAs, etc.) to support Dodd-Frank compliance. The implementation was so successful that Goldman Sachs decided to invest in Eigen’s Series A. At the time, we thought that every bank was going to jump on the bandwagon – to use AI to digitize their ISDAs and other Capital Markets documentation, and by 2020 all banks would have complete data coverage of all of their Capital Market documents.

We were wrong. In fact, we were too early to the market. To our surprise, the vast majority of the market was capturing both operational data and regulatory data (e.g. A3 table for QFCs) in an ad-hoc manual way. Digging into this further, we noticed that many of the banks received slaps on the wrist for poor compliance with QFC due to poor data quality or processes. At the same time, the notion of AI reading and answering complex regulatory questions on an automated basis did not sit comfortably with the various committees within these G-SIBs.

But something suddenly accelerated AI adoption here. We suddenly went from only serving Goldman Sachs in ISDA digitization to having processed more than 50% of all ISDAs in the US market in the last 18 months. Having talked to many customers, including Heads of Derivatives Documentation, Legal COOs, Capital Market COOs, Chief Risk Officers, and General Counsels across most of the G-SIBs, I have reached the following conclusions as to what is the catalyst for rapid adoption of AI-driven ISDA digitization:

1. Regulatory Push – Regulators are much more diligent than they have been in enforcing and auditing various regulatory requirements such as Dodd-Frank, BRRD, etc. Many of our customers have started to digitize their ISDAs either as a result of/or to prevent slaps on the wrist from regulators.
2. Inflationary Costs – wage inflation has added significant cost pressure for manually collecting data from these documents. Many banks still employ fully qualified attorneys to extract data, but we have seen extreme wage growth amongst junior lawyers making this an unsustainable option.
3. Extreme Market Volatility – from COVID, Archegos, and Greensill, to SVB and Credit Suisse, counterparty risk and market volatility have been at historic highs. Risk Management is starting to mandate a much more comprehensive understanding of counterparty risk and legal terms and for good reason.
4. Competitive Data Landscape – as more banks and buy-side firms digitize their documents, including ISDAs, those who do not have complete data are at a disadvantage when it comes to negotiation and positioning. As such, it is becoming increasingly important to have complete data just to keep up.
5. “iPhone moment for AI” – the ChatGPT phenomenon means that AI is now in the hands of the consumer, and anyone can use it. As a result, every Board has AI on top of their agenda, and we notice that digitizing documents is an obvious and easy way to leverage LLMs to add value and ‘check the box’.