The View from London - Five RegTech Takeaways

Blog | 12 Nov 2017

I crossed the pond last week to speak at “RegTech Rising” the first conference on this subject by the well regarded team at Informa.  They were kind enough to invite Ayasdi to speak on a panel about “RegTech Pilots and Success Stories”. While there is tremendous heat and light around the subject of RegTech, the public success stories are harder to find. Our work with Citi on stress testing is often used as an example of how AI can automate key processes while enhancing explainability and doing so in a fraction of the time and resources. Still there are others (AML, Credit Forecasting, Customer Intelligence, Complaints Prediction) and that was what we were there to talk about.    

We also listened.

There were some superb panels across a number of different subjects, all of which got us thinking about the space in general. The following are some key takeaways we thought were worth sharing:

  1. RegTech needs to augment existing processes.
    This is a point that we’ve touched on a lot at Ayasdi, but was reiterated by banks and VCs at the conference. No matter how good the technology, the implementation cycle becomes dangerously long if that technology requires a forklift upgrade of a key system. The reasons are manyfold – from IT to compliance and regulatory approval. Regulatory processes are mission critical and wholesale changes are not welcomed – no matter what the performance improvement. The focus should be more around augmenting existing processes, rather than rip and replace.

    We know this first hand from our work in AML. Our customers have carefully calibrated systems with regulatory buy in and established processes. By finding the lever points in those processes and applying technology precisely one can achieve the same performance improvements and with an operationalization cycle that isn’t measured in years. Take a look at what we did with Intelligent Segmentation and what we are now doing with Intelligent Alerts and Intelligent Typologies to get a better sense of what we are talking about.

  2. Domain expertise is critical.
    One panel mentioned that the most successful RegTech companies were the ones created by entrepreneurs who were passionate about risk and compliance. These founders strived to understand the challenges and opportunities inherently and often come from inside the industry.

    We learned this early on and built a team of practitioners – from JPMorgan, AIG, Moody’s, Morgan Stanley and others. While domain expertise is critical in any industry, it seems to matter even more in financial services where expertise is defined narrowly and precisely,  Further, expertise begets opportunities to get even smarter – such as Executive Chairman and Co-Founder Gurjeet Singh’s appointment to the HSBC technology board or the myriad of panels and talks the team is asked to give.

  3. Competition in RegTech is driving expectations on returns.
    During one of the London talks a panelist correctly advised, “make sure that the service you provide is better. And not in a small way. You need to be 5-10x better – in terms of cost or false positives.” While this seems steep it is, in effect, the price for change. As a company in the RegTech space we see competition from incumbents and emerging companies. Add to this the trend of banks reducing their exposure to third-party risk by strengthening vendor assessments and cutting vendor relationships and you have a formula for failure if you cannot bring material, in-fiscal year benefit.  At Ayasdi, we continuously strive to double-down on the areas where we can provide strong value to our customers. Great examples are our reduction of AML false positives by more than 20% at HSBC (10x the industry standard) or helping streamline Citi’s PPNR modeling by more than 50%.

  4. AML is a particularly fruitful area innovation.
    During the “Expert Insight on RegTech for Managing Financial Crime” presentation, the Institute of International Finance’s FinTech lead, Conan French, discussed the fascinating ways that existing AML processes are being augmented – ranging from Biometrics to Data Sharing to Artificial Intelligence. These innovative approaches are re-inventing almost every element in a bank’s AML process from KYC to SARs. Conan also touched on what may be the hottest topic in AML – data sharing. In light of recent terrorist attacks, there is increased political will to solve the jurisdictional and regulatory barriers to sharing information. The IIF’s leadership on this point is building a coalition that includes the FATF and Europol and will materially move existing efforts if successful – unleashing another wave of innovation.

  5. RegTech isn’t about big data – it is about complex data, regardless of size
    This is a theme that was brought up multiple times – especially on the topic of operational risk. Strong RegTech solutions must be equally adept at both small and large data sets. For example, in some scenarios, data may be scarce, and in the words of one panelist, will need to be “squeezed” to get the most value from it. In other examples, such as conduct surveillance, there is a huge lift in combining large amounts of data to uncover hidden patterns.

    Finding value in data, despite size is a function of the technology applied. When presented with a small data set, the technology needs to transform, automatically, that data to create more features and needs, in turn to be able to handle that expanded set of features. Further, the technology needs to scale to billions of rows and thousands of columns. That means the math has to scale, the compute has to scale, and the human interaction has to scale.

As you can tell, this was a content rich two days and speaks to the dynamic nature of RegTech – both in London, but also in places like Singapore, Hong Kong, NYC and Silicon Valley. There are additional takeaways that I would be happy to engage on just ping me at [email protected] and we can find some time.

Needless to say, we are always happy to talk through how these themes apply to our work – from operational clients to ongoing proof of concept work.