Artificial Intelligence (AI) is one of the hottest topics in technology at the moment. Despite this, Hentsū has seen that a large portion of client’s compliance and regulatory functions are still being performed using antiquated tools developed decades ago where simple keywords, phrases, and statistical models are thrown at a database of data with the hopes of detecting risks. Even primitive spam filters using AI from the 1990’s are more advanced.
This is absurd considering that the increases in computing power, decreases in the costs of data storage, and rise of distributed computing over the years. It is now possible for anyone to “spin-up” an AI project within a few minutes on AWS, Google, or Azure. While the promise of leveraging human cognition for machines to solve crucial problems, such as finally finishing the last book of Game of Thrones, seemed quite impossible only a few years ago, we’re now using technology, like Hentsū Regulatory Ecosystem, to ensure unnecessary risks are not being taken.
While our customers generally use AI and Machine Learning (ML) for research models and algorithms to generate returns, the revolution in regulatory technology (“RegTech”) sees human cognition applied to reducing risks and costs, such as performing market abuse surveillance.
Hentsū is seeing demand from firms into technologies that are able to process a lot of different types of data, such as trades, communications, and personnel data to uncover hidden relationships and anomalies. These are driven by ever increasing regulatory fines, reputational damage, and a push across the industry to reduce operational costs.
A prerequisite is having the right data platform to collect and centralize content into one place; similar to a data lake. Once the data is properly enriched to suit the use cases one or more of the following techniques are employed to uncover risk in a way that using just plain rules and keywords cannot do:
Using market abuse as an example, firms may train their detection scenarios to identify when a trader is chatting with another market participant with whom they have a personal relationship, as determined by historical conversations. When Natural Language processing identifies potential collusion in a chat, the detection scenario can then look at all channels of communications, as well as trade and market events, to detect correlations. Having all relevant and historical data, and the power to process it and make correlations, it possible to uncover when the same group of individuals routinely move their IM chats to WhatsApp and then are involved in the same trades at key times – such as during a benchmark setting.
In the case of transaction or trade monitoring, having the historical trading pattern of an individual can detect whether “spoofing” has occurred or whether the activity was a proper part of the trader’s style or strategy. Without the pattern recognition, cognition, and the ability to “learn” over time, traditional tools will just continuously create alerts each time certain fixed thresholds were triggered. Furthermore, having a library of past enforcement cases to “learn from” would allow for pattern recognition to replace hard-coded rules that never self-adjust via Machine Learning.
Compliance and simplicity are two words not often associated with one another. However, the Hentsū Regulatory Ecosystem is built on cloud platforms and using advanced AI. It enables the collection, monitoring and storing of disparate sources of data through our suite of best of breed compliance and regulatory technologies. It also offers one-stop subscription to the top artificial intelligence consultants and solutions on the market. There is no need to still be relying on outdated tools when simpler, more innovative solutions are readily available.
Have compliance and regulation questions? Come talk to us… you can contact us at: firstname.lastname@example.org