andreyphoto63 - stock.adobe.com
The massive money laundering case that gripped Singapore last year has cast the spotlight on the state of anti-money-laundering (AML) capabilities across the region, driving financial institutions to shore up detection of financial crimes.
SymphonyAI, a software company that builds artificial intelligence (AL) solutions tailored to specific industries and use cases, has been active in the AML space, having built a technology platform that is now used by major financial institutions to investigate and manage money laundering cases.
The company was founded by serial entrepreneur Romesh Wadhwani in 2017 and has since grown its business by 30-35% year on year (YoY). In an interview with Computer Weekly, Mike Foster, president and CEO of SymphonyAI’s AML technology specialist Sensa-NetReveal, talks up the use of AI to fight money laundering activities, the company’s expansion plans and what lies ahead for its technology.
Tell us more about SymphonyAI
Foster: Our founder Romesh Wadhwani, a serial entrepreneur, recognised very early on that AI was going to be transformational in the technology industry. But our key differentiation is that we don’t want to build generic AI. We’re very focused on a small number of deep industry verticals such as retail, industrial, media, as well as financial services that’s very focused on anti-money-laundering and compliance.
We’re based in Palo Alto, California, with about 3,000 employees across 22 countries. We’re profitable and we’re seeing growth levels of 30-35% YoY. We have a fabulous success story and lots of big ambitions this coming year and beyond.
How is SymphonyAI addressing the challenges with AML through the Sensa-NetReveal platform?
Foster: I think we have an advantage where every regulated financial institution has to adhere to deal with money laundering from offences such as terrorism, human trafficking and sexual exploitation. That said, every financial institution’s technology environment is unique with a mix of data, functions, products and services. It’s a very complex environment and there are only a number of technologies that can deal with that because it’s not just about being able to consume all of that transaction data, which has probably doubled over the last two or three years.
The systems that we deploy to look at each of those transactions not only need to be extraordinarily scalable, they also need to run at near real-time because after the fact isn’t particularly helpful in AML. We’re very fortunate that we have a combination of an AML platform that has a heritage of over 20 years of technology architecture, industry subject matter expertise and the latest predictive AI models and platforms that had been developed and scaled within the SymphonyAI business.
Through the platform, we can apply basic rules and flag any transaction greater than $10,000 or whether a particular transaction is normal for the person making that transaction. If it’s a problem, then there will be an investigation but as you know, that’s not sufficient anymore because the bad guys are quite good at laundering money somewhere in the region of $4tn a year.
Mike Foster, SymphonyAI
They have access to the same technology and technical expertise that financial institutions do, so we need to use AI and generative AI that can continuously optimise themselves based on a particular threat vector or criminal behaviour at any given time in any one financial institution around the world in almost real-time.
The considerations when it comes to AML processes vary across different jurisdictions and banks. How do you cater to all these differences?
Foster: If you look back two years ago, a financial institution would typically deploy an AML platform that’s heavily tailored to its technical and data environment. There’s a huge amount of effort required by the financial institution’s technology team and software provider to complete that integration. That could take weeks, months or even years with a lot of complexity.
What we’ve been able to do now with more dynamic AI models is to provide customers with more dynamic configurations. They can configure our platform in a way that’s specific to their needs, data and flows, reducing time-to-value or time-to-effectiveness from months to eight to 12 weeks.
And how a customer chooses to consume the platform and whether they deploy it into their cloud doesn’t matter as long as the environment is open to us. It’s obviously secure and protected, where we’re able to provide dynamic updates, just like you would with the iOS operating system. With generative AI (GenAI), we’ve moved beyond the heavy lifting and upgrades that require significant technical expertise to drive.
How can your solution augment existing the AML models and capabilities that banks have built up on their own?
Foster: Every customer we work with has some sort of in-house AML capability and models. A reasonably large financial institution could employ over 3,000 people just in its level one investigation team alone.
It’s extremely disruptive to rip out the technology you’re using to meet regulatory requirements around AML compliance because the transactions never stop. So, what do you do? We have many customers that are replacing legacy platforms with our technology, but they run the old platforms in parallel for a year before they switch them off. It’s quite a significant amount of work, it requires commitment and it’s not cheap.
We’ve developed a technology called Sensa that sits alongside or on top of your existing AML technologies. It has the latest AI and machine learning models that operate in near real time, allowing you to take alerts and put them into flows. It does its analysis while driving down false positives that plague every financial institution.
You’d also want to see if you’re missing any risk or illegal behaviour. You can consume the platform in your environment or through our instances in Microsoft Azure, Google Cloud or Amazon Web Services. That’s how we augment your existing investments where you understand your workflows and your people are comfortable with your applications. We can deploy within weeks and it’s a nimble, lightweight change that any financial institution can make.
Now, a lot of people still see SymphonyAI as a startup. Has that stood in the way of getting into deeper engagements with many of the world’s top financial institutions?
Foster: SymphonyAI is about five to six years old but we’ve been able to demonstrate very strong financial performance and revenue growth. That said, the barrier to entry into a large financial institution is very high for a startup.
We have an advantage in that SymphonyAI acquired NetReveal in October 2022, so we’ve just gone beyond the first anniversary. The acquisition came with a world-class customer base of 180 blue chip banks, insurers and other financial services customers. And we’ve been able to take that customer base that had a predominantly rules-based platform and augmented their capabilities very quickly with Sensa AI technologies.
While we’ve dealt with that problem by means of acquisition, we still had to go through the due diligence that any bank would undertake to accredit their suppliers. But I would imagine today that if I were a brand new startup trying to sell to a tier-one financial institution, that’s going to be tough because technology moves so quickly. By the time you get your supply chain accreditation, your technology will be almost out of date so it’s a very difficult environment right now.
Besides large financial institutions, are you targeting emerging financial technology startups that also have to comply with AML regulations?
Foster: Historically, we’ve focused on customers in the top quartile in terms of scale, not by design but by the amount of effort, complexity and cost to deploy the platform two years ago. Our new product suite is born in the cloud and is a true SaaS [software-as-a-service] deployment that scales from the smallest to biggest companies. That opens up a significant new market for us while enabling us to maintain our position with bigger clients as well.
There’s a different set of requirements when you’re working with smaller financial institutions. They haven’t got armies of people, maybe three or four people who don’t need big case management capabilities. They don’t need dynamic models and they don’t have bazillions of transactions going through. Still, we engage them early so we can get to a point where we can support them as they leap off a fairly complex spreadsheet to manual compliance to consuming industry-grade software.
Can you kind of give me a sense of what’s next moving forward for SymphonyAI Sensa-NetReveal in terms of making deeper inroads into the Asia-Pacific (APAC) region or building specific technology capabilities?
Foster: We’re very mindful that we don’t want to dilute our capability too quickly. We operate out of three regions, predominantly Europe, Middle East and Africa, North America and APAC. Historically, we’ve been very strong in Singapore, Malaysia and Australia so we’re expanding much more into the region and building partner networks across Southeast Asia.
From a technology point of view, we are investing very heavily in co-pilot capabilities for our investigation hub, a case management platform consumed through the cloud. It sits alongside the investigators and does a lot of the mundane work for them, by searching the internet and gathering data across the financial institution. It then builds a dossier of intelligence and then writes the narrative for the investigator who then assesses if human eyes are needed. The impact of that is phenomenal, making investigations about 70% more efficient.
Today, we do that specifically for AML but we’re building many more of these co-pilots to augment KYC [know-your-customer] processes and improve payment dispute resolutions. We’re also going to start moving into this world of screening and sanctions.
The political pressures around the world are quite significant right now and I don’t think they’re going relax anytime soon. I would like any company to be able to access a screening service for anybody it wants to do business with. That will have a profound impact on the whole money laundering problem globally.
There’s one more thing that’s early in our thinking. The issue we all face is that the bad guys are very good in the way they launder money and hide themselves. Today, financial institutions are able to share data around risks and best practices on how they’re dealing with new risks.
We’re working on a safety and compliance network that would deliver exactly that. It’s still early days, but if we can push the envelope – and we’re working on how personally identifiable information can be managed and accessed – then AML can become more predictive rather than reactive. My vision is to get to a point where we are able to identify a risk and threat before it happens on the global stage.
Read more about AI in APAC
- Organisations in APAC are deploying and experimenting with generative AI in healthcare, citizen services and other use cases amid cost-related concerns and other challenges.
- AI can be transformative in government, but its implementation has to be done in a way that drives positive outcomes while mitigating its downsides, according to the head of Australia’s Digital Transformation Agency.
- Culture Amp is building a generative AI capability that summarises employee survey responses, automating a process that typically takes HR admins up to hundreds of hours to complete.
- The multidisciplinary nature of AI offers career opportunities not only in builder roles like engineering and data science, but also in AI ethics and applied AI.