The Copilot Connection

Ep 28 - Copilot and AI in Healthcare with Gary Raife

Share episode

Send us a text

In this episode of Copilot Connection, Zoe and Kevin are joined by Gary Raife to discuss the transformative potential of AI and Microsoft Copilot in the healthcare sector. We explore the challenges of interoperability, the role of AI in enhancing patient care, and the exciting developments of Dragon Copilot. The conversation also touches on the integration of personal health data, the importance of trust and transparency in AI applications, and the future of healthcare driven by data and technology.

Takeaways

  • Healthcare presents a significant opportunity for AI integration.
  • Interoperability remains a critical challenge in healthcare systems.
  • AI can enhance patient experiences and streamline workflows.
  • Dragon Copilot introduces ambient listening capabilities for clinicians.
  • Wearable technology can provide valuable health data for monitoring.
  • Trust and transparency are essential for AI adoption in healthcare.
  • Personal health data integration can improve patient outcomes.
  • AI should support clinicians rather than replace them.
  • The future of healthcare will be increasingly data-driven.
  • Collaboration between technology and healthcare professionals is vital. 

Key links

News

Healthcare related

Kevin McDonnell (00:11)
Welcome to the Copilot Connection.

Zoe Wilson (00:14)
We're here to share with you all the news, insights, and capabilities of the Microsoft Copilot ecosystem from across the entire Microsoft stack. I'm Zoe Wilson and I lead the Copilot Business Transformation Practice covering all copilots and agents for Accenture and Avanade. I'm an MVP for Copilot and Teams and a Microsoft Regional Director.

Kevin McDonnell (00:35)
I'm Kevin McDonnell. I'm an MVP in Copilot and Viva. I'm the Copilot strategy, modern workplace AI leader Avanade and many other things and a father. And I feel I've got the flow of there wrong, but let's just prove that we don't pre-record this bit. We'll be releasing episodes as podcasts and on YouTube with insights from experts from the community and Microsoft on what the different areas of Copilot are, the impact they can make to you and your organization.

what you need to do to prepare for them or start implementing them, and even how you can extend them with the beauty of agents.

Zoe Wilson (01:09)
So today we're going to be joined by Gary Raife, who's an MVP in the UK to talk all about Copilot agents and AI in the world of healthcare. But before we get onto the really great chat with Gary, there's a few bits of news and things like that that we wanted to share with you.

Kevin McDonnell (01:27)
Yeah, we've just come back from the MVP Summit and Zoe, we both took our microphones and our cameras to get lots of interviews with people. How many interviews have we got to share from that? Yeah, we failed miserably and were too in-depth in some of the NDA content and chatting to people to remember to record things. I'm going to blame the weather because last year we did some lovely ones out in the in the sun.

Zoe Wilson (01:37)
not very many.

Kevin McDonnell (01:55)
And we didn't get so much of that this time, did we?

Zoe Wilson (01:55)
Yeah. Well,

there was a little bit of nice weather. mean, a lot of people say that, you know, Seattle is like the UK and it is a little bit in that it rains a lot and it could be quite grey, but it was interesting, wasn't it? Because we had days where it was just torrential rain and grey and about eight degrees and that Celsius for those of you who, you use sensible measuring.

Kevin McDonnell (02:00)
Shh, shh, don't tell people.

Mm.

international.

Zoe Wilson (02:20)
temperature

measuring systems. And then we had like one day or two days where it was 21, 22 degrees and bright sunshine. yeah, yeah. is

Kevin McDonnell (02:26)
Yeah. And then back down again, it was mad. And of course,

I chose the day where it was raining an awful lot to go on a nice hike. So that was that was all perfect timing.

Zoe Wilson (02:38)
Yeah, it was, it was a really good week though. mean, it's one of the most valuable weeks of the year, I think. And I know there's a lot of stuff that we can't talk about, for me, just being able to get that in depth time with peers, we know with people that we know across the Microsoft community who might be in similar roles at similar organizations facing similar challenges, having similar conversations with our customers that,

that opportunity to be able to kind of share and have really thought-provoking conversations with peers is incredibly valuable. And of course the product group, the products engineering connects as well.

Kevin McDonnell (03:11)
Yeah, and I think it's interesting what for me, I take away from the summit. There's that NDA material that's been shared that we can't talk about, but it is those conversations. It's kind of bouncing a bit of that what we're doing now, what others have done forming it into what's important right now to other people that I find really interesting. I think we were lucky being at Avanade. We had so many Avanade MVPs, many of us staying in the same place. There was lots of of chats afterwards in the...

the Ubers back to the hotel and the hotel bar occasionally when it was open, which wasn't anywhere near enough. Don't mention karaoke. damn it!

Zoe Wilson (03:42)
And yeah, mean, lots of karaoke. You

know, there are some people in our community that I'll never look at the same way again.

Kevin McDonnell (03:54)
Yes, yes, but there was there was a lot of work that went on. It's it's a pretty draining week of conversations and things.

Zoe Wilson (03:57)
Yeah.

Yeah,

we did have one particular highlight though, Kevin. Do you want to tell our listeners about the highlights?

Kevin McDonnell (04:07)
I assume you're not talking about karaoke here, but it was fantastic to be able to interview Jeff Teper in his kind of executive meeting room in Building 3, not Building 4, as I realized, a little too late, within there and have that chat with him. And if you missed that, I'm afraid you missed it. We did broadcast it live. It was on the Copilot Fireside chat, but our policy on that

Zoe Wilson (04:10)
Hahaha

Kevin McDonnell (04:32)
is that we want people to be open, to be able to chat about things. think Jeff certainly covered some areas that I wouldn't have mentioned and thought I would have gotten in trouble if I brought up, but it's fantastic. So keep an eye on that. CopilotFireside.Chat in the show notes. And we will have an exciting guest we hope to announce very soon. Usually have announced by now, but we're chasing someone down who's fantastic. We just got to booking these things over the Easter holidays and spring break in the US.

Probably not the best idea, but hopefully we'll have someone we can announce on that soon.

Zoe Wilson (05:02)
Yeah, I'm really excited about our potential guest for the end of this month. And actually, I thought the session with Jeff was great. Like you said, there were some things that I'm glad he was the one who brought them up. we did make sure that he understood that the audience wasn't an MVP only under NDA audience, but he was just so open in terms of what he sees as the challenges and opportunities.

the SharePoint fangirl in me was of course super excited to be on the SharePoint OneDrive leadership floor where you could see all of the signs for where all of the different products engineering leadership teams sit. And it was just personal exec meeting room as well, wasn't it? Which was pretty cool.

Kevin McDonnell (05:40)
Yeah.

Yeah,

yeah, no, it was very cool in there. And if Jeff is listening, I would like apologise for the face I pulled when you said the words vibe coding. Seriously, it's not a phrase I like. It's a phrase we heard quite a bit that week and are still hearing. So I haven't changed my mind in it. I stick to it, but I do apologise.

Zoe Wilson (05:52)
Hahaha

Yeah,

I even saw Satya, I don't know if it was like an interview or an article or something I read, even he was talking about how he'd been vibe coding. And it's just really interesting knowing amongst the MVP community kind of what they see. It's almost like an allergic reaction, it? you know, even though it was a PhD, who is like a PhD guy who came up with the term.

Kevin McDonnell (06:24)
Mmm.

Zoe Wilson (06:27)
and a lot of Microsoft leadership seem to be talking about it. It's not a good thing.

Kevin McDonnell (06:32)
Yeah.

And Chris Lloyd-Jones, if you're listening, no Vibe Engineering is not really any. It's slightly better, but I'm not letting you get away with that. no, exactly. Talking of announcements, exciting one I announced while we there is Experts Live UK is having their first conference. I'm involved in the organizing group for that. So if you're in the UK, September 16th, keep getting the date wrong, but Tuesday, September the 16th in London.

Zoe Wilson (06:39)
It's not any better, yeah, not any better.

Kevin McDonnell (07:02)
Having our very first conference and the call for speakers is out now. So do check on the the show notes for that one would love to hear from you. We've had a really good response already. So get in there. That will be fantastic.

Zoe Wilson (07:16)
And that is a full, that's across all of the Microsoft tech stack, isn't it? So all of the different parts of the Microsoft cloud.

Kevin McDonnell (07:24)
Yeah, that's right. The only area we're not hitting too much this time is development. We are hoping to have a kind of longer conference next year and we'd love to get the deaf people in there. But check it out. If there's something you think's worthwhile, bring it along, definitely.

Zoe Wilson (07:41)
Brilliant.

the other thing that's happened recently was Microsoft's 50th birthday. And while we were on campus, we saw all of the signs and the banners and everything. And then the week after we came back, we've seen Microsoft have the official 50th anniversary.

Kevin McDonnell (07:45)
Absolutely.

Zoe Wilson (08:00)
And to coincide with that, there have been lots of announcements, lots of new agents, new updates, things like that. So one of the ones that I think is really interesting is the researcher agent, which is utilizing some of the deep thinking capabilities. And this was something that Jeff talked about actually when we did the Copilot fireside chat with him, where...

He was saying like, if he has a scenario where we need to produce a research paper, how he might have done that previously was, know, we'd have taken research papers or white papers from partners and industry analysts and all of these things. And then, you know, it might have taken him as long as a week to be able to digest and understand and then come to conclusions and make decisions and that kind of thing. And he was saying with the research agent, he's able to just point it at all of this source material.

And in 10, 20 minutes, that's able to output kind of the, what it outputs is kind of taking care of that initial heavy load of the researcher agent. And what that then does is basically frees up his time to just be able to focus on understanding what the research agent has given him.

make decisions and come to conclusions and you know, it's more like it's more high value for him to actually be able to take that source research as a summary and then think about what he needs to do with that rather than doing all the drudgery of going through all the different research papers. And then do you want to talk about the analyst?

Kevin McDonnell (09:35)
So, yeah, I think the the analyst agent within there, it's all about reasoning. It's about looking through data analytics, which I think is a fascinating area within there. So from what I can understand, and I haven't had hands on and

detailed view of some of this but it's effectively kind of recommending Python scripts to be able to run your data query. So a little bit, and I'm going to get in trouble from this one, that it's not picking up...

It's not picking up the it's a bit like Copilot and Excel. So picking up the details there as well.

I think the the analyst agents is going to be a lot like that copilot in Excel and ways to kind of make it easier to get to your analysis, that natural language way to generate the kind of Python to run things on there. But I do want to get my hands on that and give it a try because it looks very interesting. And I think in the same one, one thing I have had a try with is the deep reasoning

an agent flows within Copilot Studio. So that's been really nice and capable with that. So do go and give that a look.

Zoe Wilson (11:05)
Yeah. then as well as the, so there was another agent that Microsoft announced actually, which was the interpreter agent as well. I know you, you posted on LinkedIn about this, didn't you, Kevin? You were having fun with it.

Kevin McDonnell (11:12)
Mmm.

Yeah, I had two posts to this. The first was when Simon Owen and I tried to do it and we recorded it using Teams. I had a very, very entertaining conversation when we tried to convert it back and forth with Italian and failed miserably on that. Unfortunately, we found out the Teams recording didn't work. So that highly entertaining one had to be binned because it was just us laughing.

on there. Simon, if you're nasty to me, I do have that for bribery later. But then, and I'll put a link to this, I did manage to demo it. I got Copilot translating to German, read that out and listen to it go through there. You can see it's a little bit staccato at times when it kicks in and kicks out. I did notice, I think here, Craig White picked up, this is the action Philly Philly Diarrhea, where I think it thought I was speaking.

speaking English when I was speaking German so it works pretty well. I haven't tried it in anger at all. think Zoe you said you'd or you know someone who's given it a go as well.

Zoe Wilson (12:25)
Yeah, so after you posted about this online, I was chatting to some of our colleagues who work in the APAC region. And I mentioned this to them, they were talking to me about a workshop with some of our colleagues in Japan that they were joining. And I said, have you looked at it? And we went and looked at the blog and realized that Japanese is actually one of the supported languages. And I talked to one of the people who was in that meeting yesterday and he said it was so bad.

that they had to go on camera because they were laughing so much. And he said, he said he wasn't just a little bit wrong. It was completely wrong. Like instead of the, you know, the business speak.

Kevin McDonnell (12:51)
it.

You

I'm not sure if it's just me Zoe, but I think your microphone might have kicked out

I'm loving this agentic world where this bright new future and we just can't get a damn podcast to work because the networks keep kicking out and microphones keep kicking us. But let's gloss over that a little bit and crack on with some of the other news. I think the other one, so we had a lot of these announcements around Microsoft's 50th, I think the other one there was the commercial copilot.

announcements as well which I think is very interesting from there that they've been looking at. Mr. Suleyman gave a great session. If you haven't seen the videos there's some really good as part of the 50th of Sachin Adela, Bill Gates and Steve Barmer chatting together. They were chatting using the the Copilot app. I'm loving the looks on Steve and Bill's faces.

as it gives responses and roasts people on there which were highly entertaining. But I think we saw some really interesting announcements that's coming up and it's always, I find it intriguing, I think it's that came up and I can't remember where so I'm going to suddenly keep quiet. But the kind of fact that the M365 copilot is slightly behind the consumer because they need to guarantee the quality of things on there.

But you can see some of the stuff that's going to come. So the ability for copilot to have memory, know who you are, know what you've talked about and use that as part of the conversations as well. More actions and actually be able to do stuff within there. And that combined with vision and pages, that ability I think they talked about here, the shopping, so you can kind of ask it to go and do things like get a ticket.

I think we've seen with Orchestrator from OpenAI similar kind of capabilities coming to the labs within that as well. And podcasts, so you can actually, I feel a little bit nervous saying this, but you can actually ask it to generate a podcast of the latest news on there as well. So please, we hope our humanity, even with all the technical issues we're having, will bring some benefit to this as well.

So please don't lose us. Moving on.

Zoe Wilson (15:17)
Yeah, I mean,

with the podcast generation, I still feel like there's value in actually having the human interpretation of it rather than just, you know, just a readout of news.

Kevin McDonnell (15:29)
Just that.

Yeah, but I think it would be really useful at getting quick summaries on there. It has been known that we talk quite a lot and don't necessarily get to the summary. So having AI do that might be useful. I think it was really nice to see not just that consumer, but some of the AI foundries, some of the developer, agentic views and capabilities announced within there. I think the agent service going to GA, so you can actually create agents, including multi-agentic systems.

So taking some of that semantic kernel capability to kind of define your different agents and how they connect to each other is really interesting. But beyond that, also things like the A-B testing, the red team tools, the red teaming agent. So making it easier. All those things we spoke with Chris Huntingford about last week.

bringing some of this capability to be a lot more integrated into AI foundry is really powerful as well. So great to see that responsible element too.

Zoe Wilson (16:31)
Yeah, think it's

really interesting to see the developments here. And then if you look across at the Copilot Studio world as well, it was great to see the updates announced to the Copilot Studio kit by the PowerCat team, which in full transparency, I haven't actually looked at this in any great level of detail yet, but from the...

from the information that I've read about this, looks like it will start to help with that reporting of agents across the organization, being able to help with the testing of low code as well as the kind AI foundry pro codes type approach. So it's just great to see kind of Microsoft making good strides across some of the big governance and visibility concerns with agents in both.

Kevin McDonnell (17:24)
Absolutely, and I think your microphones dropped out again Zoe so I will put in the show notes a link to to that this isn't just theory and something you can get but That studio kit is available to use on there The final thing that we're to cover just very briefly is around MCP the model context protocol on there

You'll see a lot of this. Microsoft really seems to be going after this. It's a way to connect all your different skills and knowledge into servers so you can service things like your databases as it puts down here, your remote services and bring these into your agents and extensibility easily within that. We've also seen Google announce A2A but...

not going to talk about that too much because we think it deserves its own show so we're going come through and pick that up as well from that. So really interesting on that. But I think for now, and I know Zoe's got to run so it needs to be now, I think Zoe will hand over to us and Gary for the interview about healthcare and copilot.

Zoe Wilson (18:33)
Yep, that sounds like a plan, Kevin. So really great. Just being transparent as well, we've already recorded the interview with Gary. It was a really great discussion and I'm sure that you'll all find it as interesting and insightful as we did.

Kevin McDonnell (18:45)
Yeah, absolutely. So, Gary, Kevin and Zoe, over to you.

Kevin McDonnell (18:53)
Hello everyone, so Gary, would you like to introduce yourself?

Gary Raife (18:58)
That's a very good question, Kevin. Yes, I will. So hello, everybody. My name is Gary Raife. I am a MVP in M365 and I have quite a long standing history of working in healthcare IT systems from all sorts of imaging platforms from ultrasound and maternity systems and lots of fun integrated healthcare.

record systems for many, years. And then I decided to change fence and go into Microsoft. Yeah, so I've been trying to square the circle of Microsoft and healthcare for the past few years, which has been an interesting journey to say the least. But yeah, so obviously why we've decided to have a discussion today about healthcare and copilot and the potential opportunities.

Kevin McDonnell (19:29)
Yeah

Yeah, absolutely. And I think it's interesting timing with the Dragon Copilot announcements, but I think this healthcare space and AI is a massive opportunity. And so I think there's certain things under the Copilot brand, but I think there's a lot more going on in the Microsoft healthcare spaces on the broader aspects as well, isn't there?

Gary Raife (20:02)
huge. I mean, you got the whole, and this is the part of the reason for forgetting interested in obviously the Microsoft space in health specifically, obviously, you know, I've got passions in M365 and Power Platform and all sorts of stuff going back many years. Interestingly, which started while I was at an NHS trust when I first came across Power Platform and saw how, you know, it could be really quite transformational to, you know, be able to develop rapidly sort of, you know, low code solutions to improve.

sort productivity in the workplace. then obviously Microsoft then started talking about the cloud for healthcare, which is obviously effectively dynamics, you know, first party solutions, but actually with a wrapper of integrated, you know, healthcare capabilities such as fire and HL seven and DICOM, which is obviously all about, you know, medical imaging. So that's where it comes in with your AI question, which is why I was getting there.

Kevin McDonnell (20:49)
OK, OK, just.

Zoe Wilson (20:52)
Hahaha

Kevin McDonnell (20:53)
Just for our listeners, just for

our listeners. So I'm thinking they'll get the dynamics. They'll probably go all one of those Microsoft Cloud things that's vaporware and things like that, which they're not, which they're not. But then you talked about Fire and EV7. Do you mind just going through what some of these are?

Gary Raife (21:02)
Yeah. No, no, it's not vape boy, it's real.

Yeah.

Well, it's an EV7, it's not an electric vehicle.

It's the HL7, which is affected. They're basically integration standards in healthcare. So HL7 is something that's been around for years and it's all about like order comms and effectively creation of reports and sending that in a structured format between systems. DICOM is another one.

Zoe Wilson (21:13)
Hahaha

Kevin McDonnell (21:32)
So is it things

like when you go to a doctor, they'll kind of take notes and then you get scans and all those goes into kind of a central database, putting that into a standardized format, basically.

Gary Raife (21:36)
Yep. Yep. Yep. Yep. Yes.

Zoe Wilson (21:44)
Yeah.

Gary Raife (21:44)
Absolutely. Sounds like you've been doing my job, Kevin.

Zoe Wilson (21:48)
I think in healthcare that question of interoperability is one that we've been trying to solve for many years because when I first started my IT career working in the National Service Desk as part of the Connecting for Health National Programme for IT, this is way back in the early 2000s. And even though that whole programme probably wasn't that successful,

this was one of the challenges that they were trying to address then, like, you know, how can you get that standardization into healthcare records? And when you're dealing with really sensitive patient data, how can that be passed from system to system or provider to provider in a way that is compliant and safe and secure?

Gary Raife (22:29)
100%. It's one of the biggest challenges. mean, you talk to any clinician, any healthcare provider, and one of the top challenges is how do we actually share our data securely, well, to the right audience at the right time to actually improve the patient experience. And this is the whole thing in healthcare and IT. think everybody who works in that space generally works there because they want to try and improve patient care. But quite often, some of the biggest things are actually

you know, affect patient care is these big systems and the fact that they don't often interrupt well. And it's, you know, it's been a bit of a bug bearer of mine for many years. you know, obviously Microsoft system processes, and this is the thing, and you get these giant electronic patient records, know, EPR systems, you know, they do, they cover a lot of things, but quite often what they...

Kevin McDonnell (23:09)
systems and processes, I think.

Gary Raife (23:23)
what they do is go broad and they sort of go relatively deep in terms of the different clinical specialties. And you think about a hospital, how many different departments and workflows and everything that goes on in hospital, it's just mind boggling. So you get these giant systems that try and cover it all. And obviously they can't cover everything to a depth, but they can go quite broad. So then you end up getting specialty clinical systems that have been developed for a key department, a key workflow, a key...

a key speciality and then you've got the challenge of right, well, the consultants and the department want to use this specific system to do this specific care package effectively, but then they need to try and get the data centrally. So obviously you've got this EPR sitting across the top and quite often they don't speak the same language in essence. And that's one of the key challenges. And then obviously you've got

Microsoft coming into the field with, you know, their cloud for healthcare, which is great, but it is obviously very American is Americanized to the American American healthcare system, which is quite different to what we've got in the UK. So you've kind of got this whole dynamics world of, of managing the patient experience and, know, in, the private healthcare space, it makes a lot of sense, but in the NHS, for example, it struggles. And I don't think it's, it's the right platform for that space.

But in the wider cloud for healthcare, there is a lot of capabilities in the interop space such as FHIR and HL7 and DICOM. And quite often trusts will potentially go and procure, you know, pack systems, which are your picture archiving systems, and they'll be stored in a server in the server room with a huge amount of storage, which they're hundreds and thousands of pounds for. But then obviously you've got a DICOM capability in Azure.

Kevin McDonnell (25:04)
You

Gary Raife (25:11)
which they could potentially leverage. And that's all part of the cloud for healthcare space. And then comes Dragon Copilot. And this is where obviously, I potentially have got quite excited about it and seen the capabilities of Copilot because obviously Copilot is essentially about giving you an intelligent conversational experience over your data.

And then obviously you've got the extensibility of copilot, which is great because you can hook it into the many disparate systems and potentially start to get a central picture of central view across multiple systems, you know, using AI and LLM to effectively, you know, make sense of it. So, you know, that to me is pretty exciting, but it's, if you take it in the sense of M365 copilot, for example, right, that that's obviously a lot of, you know, the M365 space, but to me,

extensibility bit of that in healthcare is probably the most exciting. And I'm going to stop talking because I've talked for a long time now. It's somebody else's turn.

Zoe Wilson (26:05)
Yeah. Yeah. Well, I've

got more questions for you, Gary. So when you think about all of the problems that you articulated and the things that you've seen in healthcare, you know, where do you think we're going to see the copilot plus agent spectrum and things like Dragon, you know, where do you see it having like the most immediate impacts and then longer term, what do you think those transformational opportunities are?

Gary Raife (26:12)
dear.

Hmm.

Yeah.

So I think in the immediate sense, I am quite excited about what Dragon potentially could offer because obviously, in healthcare have been using dictation in its simplest form for many years. Even when I was a younger person running around server rooms, installing servers and stuff, you quite often go into the departments and you'd see...

clinicians sat there with microphones and they'd literally just be dictating notes. instead of them type, think of a clinician, they're sitting there doing 30, 40 people through a day, they get like, whatever, 20 minutes to do their exam. And they then obviously have to type up all the patient notes. So quite often at that point in time, they were just physically dictating the notes. nuance at that point was effectively creating the patient notes as they spoke.

So obviously they didn't have to type because obviously you type in all day long, you're not really helping patients. So obviously seeing that morph into obviously now Dragon Copilot is I think a pretty exciting thing in the fact that it's obviously taking the dictated voice, but it's now also pairing it up with the sort of, well, ambient listening.

which is quite an interesting one in itself and the fact that it's actually constantly scanning what's going on and listening. And then people are having conversations with patients that's actually being captured now as opposed to a retrospective view from the clinician, which that in itself is quite interesting. And then obviously it's pairing that up with the LLM side of things to obviously help the creation of the notes themselves. instead of, if you actually read into the research about this, go on.

Kevin McDonnell (28:05)
And, and, and sorry, very

quickly, Garrett, it's not just create the notes, but it's kind of forming into the different, you know, his, his, your recommended medicine, his different topics that come through, which is really interesting.

Gary Raife (28:13)
That exactly, exactly that. 100%, 100%. And that's exactly

it. And historically, there was an interesting article I read about effectively, when it was just the voice dictation, there was something, there was a very high error rate in terms of actually how it was dictated and what it would capture. And it was getting to a 50 % rate of actually.

you know errors so, know, you would whilst it was it may be doing a very analog job of just recording Calling what you're saying and transcribing it into text. There was quite a lot of errors So there's still quite a lot of back would you know having to go fix it all? obviously with introduction of the LLM side of things and and That's actually improved that quite significantly So you're already getting an improvement in the amount of time someone's having to spend to a record it and then b You know pre-edit, you know post-edit it

But then, you know, I think with that ambient side of things, that then again is potentially capturing stuff that wouldn't have been captured historically. And probably the bit that's really, really quite exciting, which I don't think is actually talked about enough, multilingual support. Now this in an NHS trust, go on.

Kevin McDonnell (29:27)
Interesting. So is

that thinking, kind of translating conversations while the doctor or the healthcare professional? Interesting. Okay.

Gary Raife (29:33)
Yes, yes, yes, yes, yes, because,

and I shall tell you for why, and you'll laugh at this, in NHS Trust, right, in a room, effectively a clinic room, well, an exam room, there's quite often a phone with two headsets on it. One for the doctor, one for the patient, and then...

Kevin McDonnell (29:42)
Hahaha

Gary Raife (30:01)
the trust will be paying somebody to sit on the other end of that phone and translate between the patient and the doctor, right? However much that costs, I don't know. But if you extrapolate that across every single LHS trust in the country, how much money is being spent on translation services just to have a conversation with a patient? The mind boggles. All of a sudden, you've got Dragon Copilot coming in with its ambient AI.

note creation and its actually ability to translate on the fly. I mean surely that's just a no-brainer. I mean it's...

Zoe Wilson (30:40)
Yeah,

Kevin McDonnell (30:41)
Well, interesting.

Zoe Wilson (30:41)
it's really interesting because I've got friends in Bradford who I worked with early in my career. One person in particular that I'm thinking of, his mum doesn't speak English. His dad and all the brothers have basically taken care of everything, so she's never really been fully

integrated into the English speaking community. And it's a huge burden on the family because as her health has declined, as she's got older, they've had to go to a lot of the appointments. And, you know, I'm not saying there's anything nefarious in this particular instance, but you know, you could see a scenario where you've actually got a family member taking the role of that translator and are they actually putting everything on? Are they representing the needs and advocating for the person in the right way?

Gary Raife (31:22)
Pond of Sats.

Zoe Wilson (31:26)
So in terms of that access to healthcare in communities where they maybe struggle to get that, or if they don't have the support network, if they don't have the right advocates, I think that has huge potential. And actually, you were talking, Gary, sorry, Kevin, when you were talking, Gary, one of the things that I was thinking about the ambient listening is, do we think we'll see people potentially having a bit of an allergic reaction to something that's listening to them talk about things that are of personal?

Kevin McDonnell (31:48)
Yes.

Gary Raife (31:50)
Yep, can almost guarantee it.

Zoe Wilson (31:53)
So there's this real kind of, I see this in many places, but there's this tension between what the tech will enable and what people will actually be comfortable with.

Gary Raife (32:03)
They're 100 % and especially in healthcare, it's often a challenge and this is, you know, it's...

Kevin McDonnell (32:03)
I

Gary Raife (32:10)
I suppose one of the, not issues, but everybody wants to do innovation in healthcare, obviously for the right reasons, but there is this flip side rob of actually things generally move slower because of the fact that the stuff that you're doing will have a direct impact on patient care. And obviously there's a lot of, know, controls and governance that needs to go on, especially with the adoption of new technologies, which is why it's often slower, but.

That being said, if you can get it right, when you do put this technology in, it can have a massive impact, which is obviously awesome. But yeah, so I think, you're right Zoe, there will be a bristling, I think probably of the thought of, you know, something that the system recording people talking, but it probably be from the patient experience as opposed to the clinicians, because they've been dictated for many years. So that's nothing new to them.

Kevin McDonnell (32:54)
I think what

Yeah.

Yeah, I think what it was actually that that I loved about Dragon Copilot and the idea that it can take that conversation to specific elements of the notes, that it could almost isolate, oh, how's your husband doing? How's his, how's the kids and that kind of thing that where you're going on holiday, it could almost isolate some of that and then bring it back to kind of the core context of the conversation and

Gary Raife (33:10)
Mm.

Kevin McDonnell (33:26)
I like this idea that you can kind of show that to the patient, here's what we've captured, we're not going to store this recording, we're keeping these notes, do they all make sense to you? Are happy with that? And to remove that. So I think if that can develop and expand well, I think that will give people the confidence. And I think it's about trust and transparency as well. I'd be interested to know how many people knew their conversations were being recorded, what notes were being taken. I think if you can get this transparency out from there, I think we've seen that.

lack of trust from the acronyms doctors used to use that have now, I think, pretty much gone away as the internet's come along and there's been more kind of openness on that, which has really helped. So I hope that this will kind of help make that more effective for people as well and that healthcare professionals, you can tell my wife's a pharmacist and I don't just talk about doctors.

can have those conversations and be open about what's going on and be clear to people of what the restrictions are. Because I think if you hide that away, that's when people get very nervous about stuff.

Gary Raife (34:29)
Yeah, 100%. And you it's, I mean, you hit the nail on the head to be fair, you know, doctors quite often do talk in a different language. you know, it's one of the, you know, whilst they obviously have to have a level of communicating that is potentially talking about things that, you know, patients wouldn't necessarily need to hear that second, for example, but you know, it is, it is a complex world of, of healthcare and communication in that space and anything that can actually

democratise a bit more actually what's happening is a good thing in my opinion. But it's an interesting challenge and I think certainly off the back of something like this dragon, and bearing in mind I haven't seen it in the flesh as yet because it's literally so new. It's... yeah.

Kevin McDonnell (35:14)
I was going say it's in private preview at the moment coming

to the US and Canada in May, I think I saw, and then the UK, Germany, France and Netherlands sometime after that. And I will put the link to that blog post that I'm blatantly reading right now into the show notes.

Gary Raife (35:21)
Mmm.

You

Zoe Wilson (35:30)
Yeah. So that list of countries is really interesting actually, because if I look at countries where a general copilot has had challenges with things like workers councils, I'd definitely put Netherlands and Germany both firmly in that list. So it's going to be really interesting to see what the reaction is in these countries.

Kevin McDonnell (35:44)
Yeah, yeah, interesting. That's a point.

Zoe Wilson (35:52)
Because it's not so much the workers, I mean, there's a little bit of the workers council aspect with the clinicians and the other healthcare providers, but it's much more focused around like the citizen benefit or patient benefit, isn't it? So I don't even know how you deal with that.

Gary Raife (35:52)
Yeah.

Kevin McDonnell (36:04)
Yeah. And

I thought it was interesting, Gary, you were talking about that kind of translation, because one thing I was going to comment that I've seen with Dragon Copilot is it's very much about enhancing the patient's and health care professional's experience. This isn't about cutting jobs within there because, you know, the doctors still have to write up notes effectively to enabling them.

Gary Raife (36:26)
Yeah, I'm sorry.

Kevin McDonnell (36:27)
them to do more. And I really like that story until you bloody talked about translators and these people on the phone that are no longer needed and you've killed that idea on there. but I would say that I think they've had a real capacity issue that there aren't enough translators to be able to do this. So I think it is about growth and enabling more of that. You you spoke, Zoe, about bringing family members in, that there's not enough translators to support everything. You can't have people

Zoe Wilson (36:32)
Hahaha!

Gary Raife (36:36)
Yeah, but that's the thing. They are clinicians. They are clinic people.

yeah.

Kevin McDonnell (36:55)
sat there. I'm hoping that this is, maybe I'm just being rosy-tinted again, but this isn't just about kind of cutting jobs and I don't think this would be.

Gary Raife (37:04)
No,

and this is the thing and this I think is why AI in health is probably a bit more of an actual interesting case because I mean it's obviously my opinion but I struggle to see how AI will replace a qualified healthcare professional in healthcare, obviously because where else would they be? But you obviously still need a person there.

to make a judgment. And that person's judgment is based on their experience from the clinical standards, from their guidelines, from their learnings, from all of that stuff. And AI can suggest, but it will never get the holistic view of the person or the patient. It will look for commonalities, it will look for matches, it will look for using its knowledge, but it will never have.

an actual holistic view of that person sat there in the room. And you will always need a person on the opposite side of that fence to be able to do that. So AI can suggest on data and indicators, but there's still somebody needs to take a look at the person as a whole. this is why.

Zoe Wilson (38:04)
you

Gary Raife (38:20)
you know, the responsible use of AI in healthcare, I think is the interesting conversation here because actually I think it won't be an issue until people get lazy, for want of a better word.

Zoe Wilson (38:33)
Hahaha

Kevin McDonnell (38:34)
It won't be an issue until it is.

Gary Raife (38:36)
Yeah, no,

but it won't, no, but seriously, it won't be an issue until if the bit where I think it will be negative will be if it gets to the point where the clinician starts to rely on what the computer is saying in essence, or the AI is telling the clinician provided the clinician knows that it's there as an indicator, it's not there as a diagnosis, the clinicians job is still to make the diagnosis then, then

AI can support and improve and enhance the time to present the evidence in a way that actually helps them make that diagnosis more in quicker way, effectively, and in a better way. That is the best world scenario in that perspective, I think.

Zoe Wilson (39:20)
Yeah, I

mean, if we look specifically at the UK as well, I haven't validated these numbers, but I was reading something just a few days ago that was saying, I think in 2012, there were about 3000 people who had been waiting more than 52 weeks for treatment. That's currently over 300,000. So when you think about the level of pent up demand, the challenges that we have in the UK healthcare system,

Gary Raife (39:47)
Mm-hmm.

Zoe Wilson (39:48)
the

capacity issues. I do see that being able to help the clinicians get to that diagnosis quicker is something that could be really beneficial, recognising obviously that this, like you say, it's not a panacea, it's not going to be the thing that just gives them all the answers.

Gary Raife (40:09)
Yeah.

Zoe Wilson (40:10)
But our healthcare system's in crisis and anything that we can do to help make it easier for the clinicians and the healthcare providers to deal with the pent-up demand that we've got, I think, can only be a good thing.

Gary Raife (40:22)
Yeah, absolutely right. And you back to the interop question, an extensibility question, the amount of time that clinicians waste just jumping between systems to be able to actually get a view of the patient data as a whole is obscene. Considering the technology we have at our fingertips and you know, the scale of the NHS, the amount of time wasted, and I've seen some

Kevin McDonnell (40:39)
Yeah.

Gary Raife (40:51)
ridiculous things and practices over the years. It's just ridiculous. And there's so much capability here. And obviously, you know, this is where the extensibility side of copilot is, is, I think, quite exciting. And obviously the fact that it can actually start to pull data from all of these disparate systems and actually provide a view of that and any intelligent conversational way. That is the exciting part, I think, of copilot. And obviously,

the whole dragon piece is actually bringing in the conversational aspect of talking with patients and then being able to use the extensibility, the conversation base, and then obviously enhancing the workflow in a sense of creating reports, sending documents to other systems. All of these things are all capable and able to be done. That then starts to obviously enhance the clinicians workflows and the amount of time they're spending.

potentially trying to source the data, make a diagnosis and then report on that diagnosis. Those are the obviously the three key things they're trying to do on behalf of the patient. So, know, anything that we can do in that space to try and speed those things up is obviously going to have an improved experience for the patient.

Kevin McDonnell (42:02)
So awkward question for you, Gary. How, realistically, how far away is this? You know, we're not talking about going to your GP, turning on Dragon, and this is going to change everything for everyone tomorrow. It feels to me like we're probably a few years away about it will work with some data sets, so it might work in certain areas and then grow in others. What's your thoughts on that?

Gary Raife (42:06)
It's not like you, Kevin.

Zoe Wilson (42:07)
Hahaha

Gary Raife (42:17)
No.

Yeah.

So this is

the thing, and this is the challenge of the NHS all up to be fair. So currently as it stands from what I've seen, Dragon is integrated with one of the big patient records, Epic was called it. It's Epic, it's big, it's massive. It's literally called Epic. So it is.

Kevin McDonnell (42:45)
Not related

to Epic Games and Fortnite in any way, just to be clear to people. Good, good.

Gary Raife (42:48)
No, no, no, no. That would be quite cool. My son would be quite like, he obviously plays

Zoe Wilson (42:49)
Hahaha

Gary Raife (42:52)
on Fortnite all the time. But yeah, so it is integrated with clinical systems. So it is already, and to be fair, over the years, Nuance has been integrated with systems in the past and stuff. that side of things is there. The interoperability elsewhere, the technologies are there. There's often...

challenges, be it commercial, be it information governance, be it operational barriers, whatever. There's more challenges in the ecosystem than actually probably the technology as it currently stands. If you think about it, historically, trusts have obviously built their worlds. They've got all of their data quite often on premise.

You know that there's been a move to the cloud which we've all seen obviously over the past few years where they've started to leverage cloud systems and you know, they've got this hybrid world going on so then you've got The structure of the NHS itself is probably part of challenge But you know, you've got effectively your ICBs which effectively look over the the trusts as a whole and you've got data going across or that level integrated care boards integrated care boards

Kevin McDonnell (43:57)
Sorry, I for those that don't know ICBs.

Thank you. Thank you.

Gary Raife (44:06)
So yeah, so obviously they look after a patch and all within that patch is a load of hospitals and GPs. And if you actually then look at data at that scale versus just the trust scale, if you take data at that scale, you then start to be able to analyze patterns in outcomes with say socioeconomic factors of, know, different areas of deprivation might have a higher prevalence of, you know, diabetes. I don't know, I'm just making things up here.

There's data all over the place and there's no real central world effectively of being able to bring these things together. And obviously there's this talk at the minute of this federated data platform, which is, know, trying to solve that challenge. But at the point of care.

Zoe Wilson (44:46)
Yeah, but

they've been trying to do that for years, so that was the whole point of the national programme for IT, wasn't it?

Gary Raife (44:49)
I know they have, I know, I know, I know they have, I

know they have, and this is the problem. And you've got, so you've got all of these EPRs in a trust that doing all of these different things, all of these bespoke clinical systems that are doing a very bespoke clinical workflow. And you wanna try and integrate all of these things. And obviously you've then got at the point of care, something like Dragon Copilot, with extensibility and all that sort of stuff, which sits in the cloud.

So, effectively you're then trying to look right well, if you think about it, it's just a big spaghetti mess of connections and things that data needs to be able to be stored. So you can either go, well, we've got to get all of this data into the EPR, which is often where trust predominantly go, because it's their single point of truth. So you get it up into the EPR and then where does it go at sort of a wider level? Well, federated data platform. But what I'm trying to say is very badly,

There is so many connections to potentially try and create in this space to get the data where it needs to be to be able to get it to something like Copilot. Each one of those has a cost. It has a change. It has all sorts of, know, things around it to actually make it work. And then you're into the skills, capacity, and let's not forget, most IT departments of a trust are literally trying to keep the lights on.

So, you know, this is what I'm saying. The technology to do it is there. The capacity, skills and finance to do it is another issue.

Zoe Wilson (46:20)
Yeah,

I mean, if you look at if you look at somewhere like the US where it's all privatized, you can see how things like this will get more traction there. Because being able to provide these enhanced patient experiences will be part of the value prop for hospitals who are potentially competing with each other. And I've got a slightly different question there, Gary. I know that at conferences, one of the sessions that you've talked about previously is

Kevin McDonnell (46:22)
interesting.

Gary Raife (46:28)
Alright.

Kevin McDonnell (46:28)
Mm.

Gary Raife (46:35)
Yep, absolutely, absolutely.

Mm-hmm.

Zoe Wilson (46:46)
patient monitoring with IoT essentially with things like an Apple watch or other wearables. When we think, you know, this is a whole different set of data that would potentially be available to improve patient outcomes as well. how do you think we'll see that evolve? I mean, I've got an Aura ring and an Apple watch as an example. So that's two different sets of data that could potentially feed into a clinician who's making

Gary Raife (46:49)
Yes. Yes. Yes.

Mm-hmm.

Yeah.

Zoe Wilson (47:13)
observations

or decisions about my health and that's true for many people. So how do we think this kind of personal healthcare data is going to evolve into this world?

Gary Raife (47:23)
I mean, again, it comes down to cost barriers and everything else, but I mean, in reality, I mean, there are so many data points available now for a person, you know, like you say, through Apple Watch and, you know, obviously Fitbits and everything else, you know, it's just the Apple show. Actually, there is a lot of potential.

should we say preventative potentially in the sense of actually the technology is there where you could effectively monitor people in the community. And you can often hear central care organizations talking about how we need to shift care back into the community, because that's where it's at its lowest point of cost effectively. So if you can monitor people in the community and potentially spot trends,

Kevin McDonnell (48:05)
Absolutely.

Gary Raife (48:10)
in their health that would potentially indicate that there's something potentially coming up or looking like they're going to have a problem and intervene at that point. The cost of that compared to them rocking up at A &E in the acute stage is massively different. So to me, that's where that sort of technology makes sense. But again, in the UK specifically, that would need to be driven by

Zoe Wilson (48:26)
Mm.

Gary Raife (48:37)
the GPs, for example, because the GPs are the primary point of care in the UK. So the GP systems, the levels of funding for them to do what they do is minimal. They get paid pence per patient. So how would a GP fund the use of something like that, right? So you then have to also, unless it was taken at a national level,

Zoe Wilson (48:38)
Yeah.

Gary Raife (49:03)
You'd also have to effectively go and implement that into every GP or ICB. If they took it at an ICB level, that'd be interesting. But you would effectively have to get every single GP clinic in the country, you know, hooked up to something that could say, right, here's a preventative care measure that we are going to offer to our patients. If you've got an Apple watch and an iPhone or a Fitbit and whatever Android, here's an app. Go and install this app.

This app will basically monitor your data and then send it into our EPR system and we'll keep an eye on it. But here's the rub. The EPR systems that are in use by the GPs are predominantly one of two flavors, neither of which are particularly...

Zoe Wilson (49:40)
Yeah.

Gary Raife (49:50)
forward thinking in terms of capability and actually given the scale of those systems are deployed at are quite slow to change. So unless you were to supplement those systems with another one, which is again, another system that the GPs have got to use, this is part of the issue. it's when you actually roll it all up, the sort of key systems and key places that historically aren't

the most interoperable. That's a good word for you.

Zoe Wilson (50:19)
Yeah.

You've got me thinking, because I also pay for, like I get an annual full blood test, like way beyond what I'd get at a GP. I haven't been to a GP for years, but you know, there are things that I do just to keep an eye on my health. And you've got me thinking about, you know, do I create an agent for myself that's looking at the data coming from my Apple watch and my aura ring and the blood tests over the last like five years, however long I've been doing that? Could I put in like how many hours I'm working, how many hours I'm traveling?

Kevin McDonnell (50:20)
You

Gary Raife (50:28)
Mm-hmm.

Mm-hmm.

Kevin McDonnell (50:49)
It's not good idea. Don't do it. Don't do it.

Gary Raife (50:49)
And then it'll go, not a good idea Zoe, you need to do less.

Zoe Wilson (50:54)
But yeah.

Gary Raife (50:55)
Stop doing podcasts. That's why. Especially when Kevin turns up late.

Kevin McDonnell (50:57)
Yeah. Well, well, well, well, well, well, well, well, well, well,

Zoe Wilson (50:57)
Yeah.

Kevin McDonnell (51:05)
Tishek. I think we talked about this in one of the early Copilot Connection episodes, didn't we, of somebody who kind of put all their details into, I think it was ChatGPT at the time, and it came up with a diagnosis which the doctor kind of went, oh, yeah, that does sound pretty realistic. I know it's interesting, we haven't heard more stories come out of that, but I think it's a really interesting idea.

Zoe Wilson (51:20)
Hmm.

Yeah.

There's quite a lot. mean, I know someone at work whose husband had some health issues towards the end of last year. And it was quite a complex case scenario. he was, this was in the US healthcare system. So he had like three or four different specialists that he was having to go see at the same time.

the health issue that he had meant that there were considerations for some of the other things that he was exploring. So there were all these interdependencies and she was using chat GPT. So she was basically taking the output from each of the sessions, all of the test results, and she was using it, asking questions, asking what types of questions she should be asking the clinician, what types of points they should be raising, and just making sure that when they went into the appointments,

she was as well educated as she could be and had a really strong point of view. it was kind of giving her indications of what might be some of the underlying factors that they could then raise with the doctors and make sure that all of these different disciplines were actually kind of considering not just the bit they were looking at, but all of it. And that's just an anecdotal.

Gary Raife (52:30)
talking to each other.

Kevin McDonnell (52:32)
Interesting.

Gary Raife (52:34)
That's the thing here.

Zoe Wilson (52:38)
experience but I think there's loads of examples of where people are using it in this way which is again interesting.

Kevin McDonnell (52:44)
But I think it's also a big risk as well. We've heard about people Googling their symptoms and going down paths that aren't needed. Yes, but I think this kind of almost makes it easier to go down a route as well. And I think it's something that the sort of clinicians need to bring together to it as well. So not kind of opening up too much.

Gary Raife (52:49)
Mm.

Well, people have been doing that for years.

Oh, does. Yeah. And this is, and this is why I'm saying back to the original bit, but this is the thing that

still needs to be the trained professional in the loop. In essence, you know, yes, no, everybody generally likes to have a better understanding of things that are happening to them. Hence why people Google all sorts of interesting things. I mean, working at Google must be hilarious, but you know, if you then apply that logic to obviously an LLM where it's got

Kevin McDonnell (53:09)
Yeah. Yeah.

Hehehe.

Gary Raife (53:29)
huge amounts of data that has been trained on, obviously it's gonna potentially throw back things that are of consideration. Maybe because it's read an article somewhere that says that there's a potential issue that might conflate with this. It's obviously gonna throw something back. But then, well, exactly, exactly. So this is why I'm saying, you know, yes.

Kevin McDonnell (53:44)
which has no validation on it either. Yeah.

Gary Raife (53:52)
know, patients could use it as a tool to help them, you know, improve their understanding, but there still needs to be the qualified medical professional there that says, actually, do you know what? That chance is, you know, one in a million, whatever it may be, then actually, you know, just brings it down a level to actually, you know, look at the person as the whole, you know, not just the data and, you know,

I've worked in and around healthcare for years and I thought I was relatively aligned to this until it's somebody you care about, all of a sudden all levels of sort of, we say, know, keeping it on the level, go out the window, perfect example. My son went into hospital for an asthma attack and I was sat there basically chatting away to the doctor, blah, blah, blah, you know, saying, we know, we're doing this.

respiratory rates and his SPO2 levels and all that sort of stuff and blah, blah, blah, blah. And he was like, he basically says, look, I get it. He's like, don't watch the numbers. And I was like, what do mean? He's like, don't watch the numbers. He's like, you need to look at him as a whole. He's like, people have different ranges all across the board. And you you sitting there saying, oh, you know, he's dropping below 90, blah, blah, blah, blah. You know, he might do.

because that's just the way he is. And you have to look at him as a whole to get to understand how he, you you look at his breathing. Is he really struggling to breathe or is he just, you know, taking quick little breaths? These are the things that, you know, you need to look at. So this is a very crude way of explaining the fact that if you pass that data to a model, that model might go, oh, he's dropped below 90. Oh, there's an issue. Oh, big red alarms going off. beep, beep, beep, beep, beep. But actually,

it takes the human in the room to potentially take that as an indication and go, well, yes, but, and that's where the clinician, think, is supplemented with stuff like this, not necessarily replaced, if you see what I mean.

Kevin McDonnell (55:52)
Yeah,

absolutely agree. So really interesting. I think we could probably do a whole episode on personalized health care and what could happen in that world. let's look at the time. where, Gary, where should people go to find out more? Obviously, we'll put the link to the Dragon copilot announcements on there. But where should people find out more about this kind of world?

Gary Raife (56:00)
yeah. As you could say, I'll talk all day long on this, so...

Yeah.

I mean, to be fair, there is some pretty decent Microsoft literature out there. been around that. mean, obviously the cloud for healthcare is a pretty massive thing. So, but specifically there are, you know, specific health and life sciences blogs that, Microsoft publish, which are, know, which are specifically about copilot and obviously the new dragon copilot stuff. And hey, if you want to go read my blog, I quite often chat about these things.

Raife.com, you can check that one out. I'm sure you'll put a link in there, Kevin. Good lad.

Kevin McDonnell (56:40)
I have it in

the notes. I thought I'd let you do the self-promotion, but yeah, we'll absolutely put that in there.

Gary Raife (56:43)
look at that, look at that.

Better self promotion, it's never hurt anyone.

Zoe Wilson (56:48)
Well, thanks so much for joining us, Gary. I mean, it's just great. mean, me and Kevin, we talk about so much on this podcast that covers a wide range of different areas. No, no, no, I wasn't. No, I was just... But it's just really great to have someone who's actually got that...

Gary Raife (56:51)
You're welcome. Thanks for having me.

Kevin McDonnell (56:58)
You're going to say so much shit, weren't you there, Zoe? I heard that pause. Yeah.

Gary Raife (57:01)
And I wasn't gonna say it.

It's fine.

Zoe Wilson (57:08)
the years of experience and that deep understanding of the industry and the technology side as well, just to kind of bring it to life for people. So thank you.

Gary Raife (57:15)
Hello. Thank

you for the opportunity. Thank you for having me on.

Kevin McDonnell (57:20)
Thanks a lot, speech you soon Gary.

Gary Raife (57:23)
Cheers!

Kevin McDonnell (57:26)
Thank you very much to other Kevin, Zoe and Gary. It's just me, just to finish off, just want to say thank you to Gary for that session, really interesting. For those who aren't in the UK, we did cover quite a few NHS and UK centric views and things, we tried to of link it back to the US healthcare, but would love to hear from you and how you think this is appropriate for your own country. So those who aren't in the UK, who aren't in the US, how do you think?

how realistic do you think some of these things could be. So please drop us a note on our social media or contact Zoe or I on that one. We'd love to see what you think from that. Just wrapping up, want to kind of draw attention again to Experts Live UK that's coming up. Call for speakers is out there. Check the show notes or go to expertslive.co.uk and find out more about that. We've got quite a few with kind of coming into peak conference season now as well. So we've got the M365 conference coming up, sadly.

won't be there for that one but Zoe will be. She'll be over in Vegas so go and say hello to her. Hopefully she'll have some stickers to grab hold of and chat there. Neither of us are going to be at Builds this year. Unfortunately for me it is far too complicated with family and sons doing exams. So much as I love travelling to these conferences there are more important things in life so putting those first for the first six months this year. So I won't be travelling over for

builds this year or the European Collaboration Summit but Zoe will be there along with a lot of others and we're thinking maybe that we'll get a session, one of our podcasts, we're going to look at some of those events that are going on and highlight what we think will be particularly useful for those in the Copilots and AI agent space so keep an eye out for that one.

I will however, if you're in the UK, at Comsverse which is coming up in June as will Zoe so both of us will be there speaking separately but come and speak to us and let us know what you're having. I'm actually doing a very demo-centric session on Copilot so getting hands-on, not just waffling as I'm doing this but showing the reality of some of it as well so come and join for that one.

I want to say thank you. If this is your first time listening, please go and hit that subscribe button. Whether it's in YouTube, whether it's in your favourite podcast app, Spotify, we're in all those places. So just search for Copilot Connection and whack on that subscribe. Tell other people about it. I love it when Zoe and I have had a few times where clients have come up and said, yeah, we've listening to Copilot Connection. We've loved to do this. Karen Deap, if you're listening.

told you should listen to last one but hopefully you picked up on this as well and yes I'd love to have you on the show at some point as well. So let other people know, share it out there, give us comments, reviews and things like that, all the good stuff that you're thinking now's the time to skip to the next episode but no please just before you skip ahead onto your next podcast just go and whack on that. Otherwise thank you very much, thank you for the messages we've got from people about the episodes.

Enjoy your time, make sure you're using Copilot and make sure you let us know your challenges you have with Copilot as well, because not everything's perfect, but it's getting better and better. Thanks very much. Bye.