5 Minutes With an Expert: Andrew Weston

An interview with Andrew Weston, Co-Founder and Chief Technology Officer at Avid AML, and Managing Director of PropellerHead.

Tell Us About Your Current Role and Your Career

Currently, my primary role is that of CTO for Avid AML. At this point, that means actively designing and building Avid's core platform, as well as planning for its longer-term needs.

A small amount of my time is still committed to running Propellerhead, a software services business specialising in enterprise solutions. Prior to joining Avid AML I dedicated 20 years to growing and guiding Propellerhead to the point where the company could act as a self-managing entity.

Over this time we have delivered large-scale solutions for Auckland Transport, Spark, MYOB, government agencies such as the Ministry of Education, NZ Post, and NZ Customs.

My focus has always been on the challenge of delivering reliable, secure software at scale. This type of work demands we operate at our technical best and be extremely well honed in our delivery practices.

What is the Most Rewarding Part of Your Role?

One of the aspects I am finding most rewarding with Avid is the challenge to craft a platform that will scale and is sufficiently reliable to meet the needs of a global user base. It demands my full focus and builds on my many years of working to solve similar issues. One big difference these days is the ready availability of very "serious" open source technologies that we can employ to help us create an enterprise-grade platform right from the start.

For instance, Google has made Istio, their service mesh technology, available for organisations like ours to use to deliver microservices within a "trustless" architecture. Just a few years ago this approach would only have been available to large, well-resourced organisations. We are also able to achieve immediate scale in data analytics using cloud-based services from Elasticsearch (Elastic Cloud) and MongoDB (Atlas).

While the technologies themselves are cool, it's even more exciting to see them being assembled to create a platform that has a "whole of customer" approach to anti-money laundering (AML). That is, the technology has made it far easier to integrate aspects of anti-money laundering components like electronic identification, watchlist screening, nature and purpose risk profiling and transaction monitoring that have previously been met by separate, not well connected, software systems.

What Made You Interested in the RegTech Industry and AML/CFT?

First and foremost, I like the technical challenge of working on a regtech solution with global ambitions. It has a mix of challenges I personally find invigorating; it has to scale, it must be secure — our clients need to know their data is well protected, and it has to be "smart" — we should be able to help our users detect fraudulent or risky behaviour in their customer base.

Technology aside, I also like to think I can contribute to a fairer world — one that sees a more just distribution of wealth, less corruption, and less harm from criminal activity. If I step back a little it is easy to see how we might contribute to such things as reducing tax avoidance and helping in the detection of human trafficking. I love the fact that we can use technology for good in this way.

Interleaved amongst all of this is the ability to use more recent advances in technology to lower the overall cost of solutions such as ours. It means we can create solutions that are both powerful and affordable. One of our primary objectives is to put anti-money laundering software into the hands of those that might not otherwise have had access to such software.

What Technology Trends do You See Playing Out in AML?

At the top end of the market — banks and large financial institutions — there is an increasing trend away from pure rules-based systems toward employing AI to detect and flag risky behaviour. In an increasingly sophisticated digital world, deterministic rules are no longer sufficient to detect bad actors. They can employ technology to open up a greater number of "entry points" into financial systems than we have the resources to manage. Entry points can range from traditional financial transactions to prepaid debit cards to the trade in crypto assets such as nonfungible tokens (NFTs).

Machine learning has the ability to help evolve anomaly detection simply by identifying unusual patterns of behaviour in the data. This approach alone won't be sufficient to filter out bad actors — the complexity and nuances of transaction co