Scouting for Growth
00:13:55 · Bastiaan de Goei, Sabine VanderLinden
I think that it's only now with the latest frontier models that you're able to really do that. [00:14:01] S1: So let's talk about Astra, Frontier Model. And you actually highlighted a scenario where Astra will go and talk, touch different systems. So to actually automate and hopefully augment the human, how do you see that relationship between human, human judgment, trust and governance, whilst we are now orchestrating processes which are managed by an astronaut?
00:14:28 · Bastiaan de Goei
[00:14:28] S2: Yeah. I actually love the use of the term orchestration that you just said. So I think that this is very much the way that I envision how not just underwriters, but anyone who is like in charge of like an important workflow would see their daily job. So I myself, when I'm looking at how I conduct my work, I obviously have my own workflows and I work together. So earlier this week we released Dots, OpenAI's version of Persistent Agents. Two days ago, I was for the first time that I did not touch or open a single application surface. I did not open my Gmail. I did not respond to my Slack messages. I did not open up a Google Sheet or a presentation, software. Instead, I was only engaging with my agent and the agent was responding for me on Slack, was preparing presentations, was preparing notes and did all of this on my behalf. And I was literally just orchestrating and validating that what it was doing was correct and asked for adjustments where necessary. And so it's a completely new way of working. Now if you apply that to that same insurance underwriting workflows, you become essentially an orchestrator of the underwriting process rather than the operator of one. And I think that that can dramatically expand the capacity of an underwriter and the way to stay in control of that, which is super important. At the end of the day, the underwriter needs to stay in control, is that, the agent can get this to a position where it is fully priced based on a series of assumptions and have a ready to go quotation and for the underwriter to subsequently say yes or no. And it can even identify higher risk to lower risk type of submission to say like, Hey, here I'm extremely uncertain. Here is something that I'm a lot more certain about. And so ultimately I see that you will have portfolios at the end of the day of quotations that are all ready to go. And so you can focus your attention as an underwriter on those that the agent was least sure of.
00:16:39 · Sabine VanderLinden
[00:16:39] S1: So let's visualize that. So I'm Bast, and I have my chatbot on my browser, but that is a space I'm working from, connected to Slack, Gmails, and everything is happening in front of me. So if you actually look at this claim, chief claim officer or this underwriting manager, right, and all are AI enabled today, describe for us properly, you know, because most people will hear things in a different ways, right? Picturing it is going to help us understand really what we see on the screen. What does that underwriter or that claims officer, what do they see when they arrive at their desk at 09:00 in the morning and open that beautiful browser?
00:17:22 · Bastiaan de Goei
[00:17:22] S2: So the persistent agents or DOTS in in this case, they operate in the work surfaces that you already engage with. So yes, you can engage with this via chat GPT if you so wish. If you prefer to work together in Microsoft Teams or you prefer to work together in Slack, you can do that too. Okay. And I myself, I actually work a lot on Slack. And so like when I start my day, my dot has prepared my agenda and has shown all of the things that need attention for me and has already prepared responses for all of the urgent things that require my attention. So if you actually you know, like make the parallel to an underwriter, then what an underwriter would see is that you've effectively abstracted away the entirety of like the underwriting workbench and and the email inbox. So instead you'd open your preferred way of communication, could be Microsoft Teams for a lot carriers, where the agent has already prepared. So, okay, overnight I received the last 12 submissions. I've already analyzed them. Four of them are well outside of our risk appetite. There are another four that are from your preferred broker relations and the remaining... Of the remaining four there is two that I think could be very interesting and two others that are maybe. And I've already prepared a quotation for them. Do you want to review them? And so that could be like the way that a underwriter starts their day.
00:18:56 · Sabine VanderLinden
[00:18:56] S1: So as Asian carry more of the workflow or the task we do every day end to end, I think the question stop being, what does the machine can do and become who is accountable for the work, right? And we know we need that human in the loop. We are in a regulated industry. So we need to know what may go wrong as well. Pricing coverage was a big topic yesterday. Declare true because we were talking about agentic distribution settlement. So this changed someone's life. This changed someone's rules, accountability, responsibility in a regulating framework. So where especially does a human have to stay, right, in that loop? And how do you tell a decision that can be delegated from one which should not be delegated at all?
00:19:46 · Bastiaan de Goei, Sabine VanderLinden
[00:19:46] S2: So I think that, first of all, what is important is that the... When staying with the underwriting example for a moment, the ultimate quotation is the underwriter's responsibility and that person is accountable for that decision. And so the tooling though needs to help the underwriter understand what they are saying yes to and what decisions were being made, what data was being used as input for those decisions. And I see it as something that we are investing in, in order to make that really easy. As an example, and this is a very small example, we... Earlier this year, we released ChatGPT for financial services. This is a tool that is focused on banks... Banking right now. But as a very tactical example, we noticed that analysts, banking analysts, when they were preparing investment presentations and using AI tools, were double checking a lot of the data points and that went into those presentations. [00:20:44] S1: 100%. [00:20:45] S2: And so... Because it needs to be precise. And just like in insurance, banking is a business based [00:20:51] S1: And on last, you know, I do the same. I still check everything because it's about reputation now.