Artificial Intelligence

Artificial Intelligence

In this era of Data Science, the broader Artificial Intelligence topic frequently reappears on the agendas of Treasury and Risk Managers. It is often challenging to establish which elements of Artificial Intelligence could be beneficial for your day-to-day work. Read more for our view on AI.

Translating business needs

We support our clients with their venture into ‘Artificial Intelligence space’ primarily by translating between the risk & treasury business needs and the consequent implementation and support by domain-agnostic data scientists and IT professionals.

This involves defining business objectives and requirements, perform data availability and feature analysis, prototype algorithms and assessing their business benefit, aligning business processes with system infrastructure and data processing pipelines and validate the results in line with regulatory requirements.

Some lessons learned

Some of the lessons we learned we can share on Artificial Intelligence:

  • Advanced machine learning algorithms provide a computational advantage over ‘classic’ statistical models (only) when the complex relationships exist between features and the data reveals these. Data and information are two different entities.
  • In the financial industry, machine learning methods are used sparingly to support processes under regulatory scrutiny. However, for benchmarking and validation they are very useful: with relatively little effort can you show that the chosen features and model structure capture the available information appropriately.
  • A good machine learning algorithm, even a “self-learning” one, is more work than creating a good ‘traditional’ model. Which makes sense, because both the amount of data typically processed and the complexity of the modelled relationships are higher. Manage expectations accordingly.

Interested in
Artificial Intelligence?

Jeroen van der Heide
Get in touch with Jeroen van der Heide for more information about Artificial Intelligence.