Sep 27, 2026 · 58 min · 11 segments
Sam speaks to Dr Tom Hebden, digital education fellow at Quesmed about all things AI in PACES including a try-before-you-buy preview of Quesmed's PACES AI platform. Tom provided some helpful…
So, Tom, I guess just to start off, I wonder if you're able to, for anyone who, I mean, I tried to describe it at the top of the podcast, but I wonder if you can just sort of outline the new AI platform which you created at QuestMed.
What was the sort of first idea behind the process and how did you move towards the finished product, which is now available?
So there's always been, I think for the past year or so really, there's always been this simulator, this AI simulator idea in the pipeline at QuestMed.
And I don't know whether it was coincidence or whether it was me joining the team in February that really kickstarted the process, but we started off with OSCE AI as our guinea pig product, if you will, building that out.
There was planning for it and there was research and obviously the build of it and we launched that in may and it's it you know had a fantastic response we've had 350 000 minutes worth of voice you know users using our voice ai within our ski ai and we've had fantastic feedback so on the success of that well so we thought we could do a copy and paste technique onto paces It turns out it's a little harder than that.
But it also gave us an opportunity to build out some new cool features, which we'll come on to discuss and take everything that we learned from the OSCE AI product and improving it for Pacers.
So early days, and we've had some beta testers and had fantastic feedback, some things to work on, and we'll continue to do so.
And I think, I guess the thing that listeners need to bear in mind is that these products are hugely complex.
Like you think paces in itself and managing patients and medicine is complex without trying to integrate a simulation level into it.
So having taken the trial run, which the listeners will hear shortly, I was seriously impressed with the product and I hope they are as well when they hear my run through.
And obviously I had to put it on easy mode and my station I had to go for was chest pain, which as a cardiologist approaching CCT, it's a bit of a tap in, but I'll take it.
But Tom, if we could just sort of zoom out, so to speak, I wonder if you have any insights for our listeners in terms of there are so many different models And some of them are used for different purposes.
I think probably at the moment, ChatGPT is the one which listeners might find most popular, most be able to identify with.
And I guess sort of my broad question is, what features do you think are present in these models that our listeners may be able to help in their paces preparation? what things do you think would be most valuable for our listeners? And maybe after that, later on, we can start to talk about the things which listeners may think are good, but actually when it comes to these AI models, they're actually not very good.
first thing to say is, and it's a never ending challenge, if you will, at QuestMed, which is the space is so rapidly advancing every single day.
So, you know, the kind of architecture that we built for OSCE, AI and even PACES to some degree six months ago, it's now becoming a bit outdated in all honesty.
So, Tom, I guess just to start off, I wonder if you're able to, for anyone who, I mean, I tried to describe it at the top of the podcast, but I wonder if you can just sort of outline the new AI platform which you created at QuestMed.
What was the sort of first idea behind the process and how did you move towards the finished product, which is now available?
So there's always been, I think for the past year or so really, there's always been this simulator, this AI simulator idea in the pipeline at QuestMed.
And I don't know whether it was coincidence or whether it was me joining the team in February that really kickstarted the process, but we started off with OSCE AI as our guinea pig product, if you will, building that out.
There was planning for it and there was research and obviously the build of it and we launched that in may and it's it you know had a fantastic response we've had 350 000 minutes worth of voice you know users using our voice ai within our ski ai and we've had fantastic feedback so on the success of that well so we thought we could do a copy and paste technique onto paces It turns out it's a little harder than that.
But it also gave us an opportunity to build out some new cool features, which we'll come on to discuss and take everything that we learned from the OSCE AI product and improving it for Pacers.
So early days, and we've had some beta testers and had fantastic feedback, some things to work on, and we'll continue to do so.
And I think, I guess the thing that listeners need to bear in mind is that these products are hugely complex.
Like you think paces in itself and managing patients and medicine is complex without trying to integrate a simulation level into it.
So having taken the trial run, which the listeners will hear shortly, I was seriously impressed with the product and I hope they are as well when they hear my run through.
And obviously I had to put it on easy mode and my station I had to go for was chest pain, which as a cardiologist approaching CCT, it's a bit of a tap in, but I'll take it.
But Tom, if we could just sort of zoom out, so to speak, I wonder if you have any insights for our listeners in terms of there are so many different models And some of them are used for different purposes.
I think probably at the moment, ChatGPT is the one which listeners might find most popular, most be able to identify with.
And I guess sort of my broad question is, what features do you think are present in these models that our listeners may be able to help in their paces preparation? what things do you think would be most valuable for our listeners? And maybe after that, later on, we can start to talk about the things which listeners may think are good, but actually when it comes to these AI models, they're actually not very good.
first thing to say is, and it's a never ending challenge, if you will, at QuestMed, which is the space is so rapidly advancing every single day.
So, you know, the kind of architecture that we built for OSCE, AI and even PACES to some degree six months ago, it's now becoming a bit outdated in all honesty.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
