In its purest form, Open Source Data is just data that’s been made publicly available. This could be through a number of sources like social media, government websites, registries, courts and legal services as well as through regulatory bodies like OFAC, FINCEN and FINTRAC.
There are many data sources out there, including paid sources that are typically data providers that curate their content and provide it to users for a fee. This can be beneficial as it helps clear out the noise from what’s accessible on the web, but comes at a cost making it difficult for financial services organizations to navigate and implement at scale. External data can be overwhelming for investigations teams and their internal technology counterparts to source and aggregate in a meaningful way, so often teams avoid using open source data for AML use-cases or have limited access to reputable sources.
By accessing publicly available information, it gives the investigations team access to much more information than the customer provides at onboarding. It allows investigators to have a 360 degree view of the customer, which they traditionally miss out on by focusing only on internally generated data such as products used and transactional activity. This results in consistently better risk decisions that management can stand behind.
A great example of this is the recent Pandora Papers investigation initiated by the International Consortium of Investigative Journalists (ICIJ). The very same day that information was being released into the media, MinervaAI was able to access published information directly and incorporate it into the risk ratings of the individuals and entities named in real-time, and allowed us to offer immediate support to our customers.
This is ground breaking for teams that need to very quickly identify their potential risk exposure to their executives. Having lived through similar events like Panama Papers, this can take dozens of investigators several days or even weeks to compile.
The very same day that [Pandora Papers] information was being released into the media, MinervaAI was able to access published information directly and incorporate it into the risk ratings of the individuals and entities named in real-time, and allowed us to offer immediate support to our customers.
It’s interesting as there are lots of really great examples globally where publicly available data and AI have been used to detect significant issues with tax evasion, online gambling and embezzlement, all of which are precursors to money laundering. Even in traditional Risk Rating and transaction monitoring systems, an element of AI is applied.
Having implemented AI solutions within some of Canada’s largest banks, my experience has been that we need to build trust across all levels of the organization through accurate results supported by processes that make sense to the teams. Usefulness and quality need to be the top priority in order to effectively automate and create consistent outcomes.
At MinervaAI, we’ve built that philosophy into our delivery model. We work with our customers to understand their use cases and what’s important to them. Then we leverage our internal anti-money laundering and AI knowledge as well as our network of partners in industry to help guide our customers through implementation and adoption.
We constantly put ourselves in the position of our customers to ensure that any choices we are making about data acquisition, presentation and new features will not only work well, but will add value to what they do on a daily basis. We also work closely with our customers to ensure that the data we have available for them meets or exceeds their needs.
What advice would you give to Compliance Executives looking to implement an AI tool like MinervaAI into their regime?
I think the most critical aspect of implementation is to create user champions within the organization. People that embrace technology and understand that by using a tool like MinervaAI, they are taking the guesswork out of sourcing information for their investigations. MinervaAI calls all critical data sources such as sanctions lists, high risk country lists, legal and disciplinary actions lists along with adverse media every time. Teams that are used to doing that manually may need encouragement to modify their approach.
This allows the investigator to focus on assessing the output and making the best risk decisions for the organization. Having people on the team who can encourage that shift in thinking are invaluable to the process. Our customer onboarding approach focuses on this key aspect of behavioural transition.
We constantly put ourselves in the position of our customers, to ensure that any choices we are making about data acquisition, presentation and new features will not only work well, but will add value to what they do on a daily basis.
There are a lot of really good sources out there. We notice a real difference in working with agencies and governments who’ve progressed their digital transformations and adapted an API model for data access. I think that the public and private sector should continue to work together to identify and prioritize areas that are less consistent or need more centralized information, to create solutions together to enable access for investigation and detection needs.
Making it easier for firms like MinervaAI to access data and present it to investigations teams in a meaningful way helps everyone live into their obligations more effectively.