GAME OVER FOR COLOUR CRITERIA? THE RISE OF “AS WE SPEAK MONTE CARLO SIMULATION”

Being a nerd, I’ve been following with enthusiasm the fast adoption of chatbots backed by GPT-LLMs (Generative Pre-trained Transformers and Large Language Models).

I know that AI can be very dangerous to society, and understand that regulations and policies are needed, as soon as possible. As risk people, we ought to be engaged in AI systems development.

On the other hand, AI has a great potential to solve some of the current shortcomings of Risk Management, Compliance Management and Internal Auditing, as most organizations practice those disciplines today.

Strangely, I am not finding much written material here on LinkedIn about the potential impact of chatbots and AI Agents on GRC. The only two direct mentions I’ve found were the news that MindBridge is partnering with KPMG to improve the latter’s auditing capabilities and a video by Melbourne-based 6Clicks about their unified compliance and control platform now being capable of interacting with the user via a bot called Ask Hailey.

 Maybe my love for technology makes me see purple unicorns, but the fact is that I envision a new era in our field. Let´s check some potential developments with tools that are available right now.

 1) You will be able to monitor leading indicators and other relevant information in the manner of a news feed. AI embed in data analytics software will warn when things are awkward, out of tolerance, or going astray. This will pave the way for nimbler internal auditing.

 2) Benchmarking will be made easier with web scraping apps that will know when and where to look for comparative data, while you do other more interesting things. No excuses anymore for not having information from outside: GRC functions desperately need comparative data to improve the quality of assessments, with more powerful interpretations.

 3) You can create a corporate chatbot (as of now) for risk management, compliance and audit support. If the user has a doubt about a control, a requirement, criteria, policies, tolerances, codes of conduct, etc., the bot can help her right away with decision making, and the user doesn´t need to wait for Bruce from Compliance to answer the e-mail three weeks later.

 4) The corporate chatbot can also help an analyst assess a risk situation, or risks in a project, or risks for an objective, by selecting the proper assessment methods for the case at hand. The chatbot will follow a script designed specifically for that type of situation.

 5) Last, but not the least: imagine you are at an important meeting. People are talking about a decision that must be made and they start using colours to refer to a risk (see note A).

 You think: “I wish I had time to run a Monte Carlo simulation and show them the real thing, with fat tail and impacts for certain specific cumulative probabilities, such as P90”. Well, it is over. You can do it during the meeting by just asking a chatbot (using your voice or texting). The only things you need to input: the shapes and ranges of the distributions for frequency and impact. On the other hand, if you can’t do that, you shouldn´t work in risk management after all.

 The chatbot will provide the loss exceedance curve in seconds, and you will display the risk visualization right away, during the meeting — hopefully, before they make a poor decision.

 Within 3 years, give it or take it, you will be able to display risk as a movie where you can explore the potential scenarios, instead of a dot in a static, and often wrong, heat map.

 (A) Colour codes to identify a risk situation will not go away, because they are ingrained in our communication protocols. They are useful in a pre-crisis situation. What we would like to see is a better understanding of what is behind the criteria and the assessment methods used, so even POTUS and CEOs can get a brief explanation of why the situation has a certain colour condition and how the action plan addresses the concerns mapped by the assessment. Well, for POTUS it would be a color condition. For Charles III it is a colour condition, I guess.