Posts

Note that things are moving very quickly with AI and OpenAI. CatGTP in particular so some of the issues below may have been fixed by the time you read this. (Sadly, not the bits that involve humans.) Some things to think about before you dive into GTP3 Like most of us, I love new toys. And like many others, I’ve been playing with OpenAIs tools since access opened up last year. With GTP4 scheduled for release soon and rumors that Google and Facebook both have pending AI releases, a lot will happen with AI in the near future. ...

(Read more about the origins of the risk metrics project here and here) The Risk Metrics White Paper The risk metrics project has come into sharper focus over the last couple of weeks and I’m now comfortable with the foundation of the system and how to manage the data. I’m still a little unsure as to which metrics will be most useful but I have an initial set of 12 to start with and user feedback will be the best judge of what’s relevant. ...

I chatted with OpenAI’s Chatbot the other day and asked it some questions about risk management. The answers were clear, nuanced and thoughtful, supporting all the hype surrounding the tool. It’s a text compiler so therefore not an analytical machine for the purposes of risk analysis – other tools of that kind will come on the scene eventually but I’m writing a paper as to why that’s a little ways off yet – but there are some uses I can see in the risk management space. ...

For most people, this will be a busy time of the year: end-of-year performance reviews, last-minute budget planning, holiday parties, or a final push to achieve their annual goals. All of this adds up to a very busy time when all you want to do is take a break. However, there’s one other thing I’m going to suggest you squeeze in before you start next year’s ERM program. An end-of-year review This review will neatly tie up everything you did over the past year and get you ready for the next. And best of all, it’s something you can do in an afternoon. ...

It’s no secret that I love standards (I even confessed to this in a previous blog post). However, there’s a problem: as soon as you introduce a standard, you’re in danger of turning things into a compliance exercise. Initially, that might not sound so terrible. If you’re compliant with a risk management standard, you’ll be managing your risks properly, right? Unfortunately not. Because compliance, or anything else that can be reduced to a ‘check the box exercise,’ very quickly moves the focus from outputs to inputs. So, as long as you have written the procedure, conducted the assessment, and established a governance board, you’ve complied with the requirement set out in the standard. But that doesn’t mean that the procedure is fit for purpose, that the assessment effectively evaluated the risks, or that there’s effective governance in place. It just means you’ve checked off these requirements in the standard. ...

*The Signal and The Noise Why So Many Predictions Fail–but Some Don’t by Nate Silver *is the 2012 best-seller from the then New York Times columnist who now runs the FiveThirtyEight election analysis and prediction site in the US. TL;DR – This is a great read that explains why and how you need to mix statistics and experience in your predictions and decision-making. It can feel a little technical and heavy on the math at first glance, but Silver explains everything clearly, making it an easy read, given the subject matter. ...

Something that’s very hard to do as a consultant is to say,* ‘I don’t know,’ *or confess that you don’t understand something. But you’ve got to overcome any discomfort and learn how to do this because it’s critical that you’re honest about your knowledge gaps if you genuinely want to understand an organization or situation. But the problem is that this is hard to do when you’ve positioned yourself as an expert, and someone’s brought you in as a problem solver. So you’ll feel self-conscious if one of the first things the client hears you say is, ‘I don’t understand that. Please explain’. ...

This post is an update on the risk metrics project I described a few weeks ago. You can read about the origin of the project and the intent here. If you recall, the underlying metrics for the project had to meet the following criteria: Broad, not narrow (meaning the metric has widespread effects) Publicly available Easily understandable Updated frequently Commonly used Unfortunately, for a lot of the data I looked at, I could get four out of five criteria but often one remained out of reach. Usually, this was frequency as some macro-indicators are published in annual reports, not weekly or monthly. However, for a macro trend such as literacy rates, a year-on-year measurement is sufficient, which means that the final criteria could read regular or frequent. ...

A problem I’ve observed is that we often need to differentiate more clearly between what we need to do and how we will do it. This confusion makes us think that a problem’s been solved when, in fact, all we’ve done is identify the effect we want to achieve, not how we’re going to get there. The roadmap for how to make that happen is missing. Imagine a business that’s trying to cut its costs because its cash flow is out of balance. Discussions might end with decisions to trim expenses or reduce headcount, which are reasonable ways to cut costs significantly. Everyone leaves the meeting thinking the problem’s been solved. However, they still need to determine what costs can be cut without compromising operations. How to manage layoffs and the effect these will have on the business? They will have identified the ‘what’, but not the ‘how’. ...

My risk management master’s course had a heavy reading list as you’d expect but a lot of these set texts – and the course material itself – was pretty thick and not much fun. However, I found a few books that I used alongside the course work which helped summarize a lot of the technical, academic stuff and, frankly, made it less dull. I wanted to share a few of these with you beginning with a book about when and how things go wrong. ...