Machine learning is set to take an increasingly important role in the move away from Libor as firms begin the switch to risk-free rates in earnest after emerging from the Brexit transition period, according to Delta Capita.

“People are taking this much more seriously,” the London-based firm’s head of consulting Rezwan Shafique (pictured) told Global Investor. “They understand and appreciate the 80%-20% rule, that from a volume perspective it’s the 20% that’s going to be the problem.”

A year ago, Shafique said, firms and their budgets were more focused on Brexit, but now transition programmes are fully in flight. Delta Capita’s clients are seeking complete capture of all impacted contracts, and are fully aware of the need for digitisation so they can apply machine learning and artificial intelligence (AI) to the process.

“From a technology point of view, before we get anywhere near AI, the real issue is converting and digitising. Shafique said. “If you don’t apply machine learning and elements of AI, you are not going to get completeness, which opens you up to litigation risk, and you’re not going to meet the deadlines on time.”

Key challenges for Delta Capita’s clients include products that add an extra level of complexity such as transitioning non-ISDA and exotic derivatives, as well as complex instruments including structured bilateral loans.

But, as head of regulatory change and solutions Karan Kapoor pointed out, investing in AI for the transition to risk-free rates has “repeat and play” potential as it will not be the last major rule-change to raise compliance challenges.

“Not just with Ibors, but every single regulation that comes through, whether it be Mifid or anything else which has any kind of financial consequences…. going back and identifying those contracts is a big problem,” he said. “You can start reinventing the wheel every single time.

“From a regulatory point of view, the key issue that our clients want to get under the skin of is the entire conduct risk and litigation risk aspect. At this point in time, nobody can say consistently across all the jurisdictions how the conduct risk aspect will work.”

As firms await further regulatory clarity, Delta Capita is focusing its clients on what it describes as a medium-risk approach, which involves reviewing, repapering and digitising Libor contracts using optical character recognition of datapoints, reassessing portfolios and focusing attention on contracts referencing priority risk-free rates.

Once this work is done, clients can fully benefit from machine learning, a service the managed services and consulting firm is providing via AI platform Luminance, which was originally designed for legal professionals. Luminance’s technology can instantly surface key datapoints, identify the parts of all contracts impacted by Libor and learn from the interaction of the reviewer.

“When you add the data to Luminance, it can start learning from what you did the last time and doing it for you,” Kapoor said. “Then AI offers another level up because, without seeing what you’ve done in the past, it can understand what you’re going to do in the future.”

Steve Vinnicombe, Delta Capita's head of consulting and solutions, has said that stock loan and repo reporting rules across the UK and Europe must align in order to maintain “harmonisation” across the broader European financial markets.

Delta Capita in November partnered with GBST, a capital markets fintech that has developed a Financial Transaction Tax (FTT) solution dubbed Syn-FTT.