00:03:53 · Sebastian Johnson
[00:03:53] S0: Can you touch on the adoption? Like, how receptive are clinicians, you know, the the real end users to to to using tools like yours?
00:04:03 · Lukas Saari
[00:04:03] S1: Yeah. I mean, so clinicians are historically not the most change prone profession, you could say. I mean, this also comes from them really having felt that they've been abused by IT systems historically. I mean, so if you're familiar with it, you know how clinicians are spending 40% of their time on admin. They're sitting in the systems that were built in the eighties and nineties and still look the parts. And commonly it's been coming like top down from the decision making and saying like, now you need to start using this tool also because then we can have less medical secretaries and we can save some money and so on. But really, it's just like adding to the burden another system, another thing to kick through and so on. But what's so revolutionary now is, okay, if you ask a clinician, what is it that you hate most about your work? It is, I'm right now sitting two hours every evening after I come home and doing documentation instead of spending time with my family. So if you go to a clinician and you say, now we have a solution that solves this for you, the clinicians are very happy to accept it and then start using it. So for us, we're actually commonly seeing that it's rather almost more of like a bottom up motion with the end users, the clinicians being the one that really are pushing for the solution. And then it's rather a question of how fast can you get it through the organization with the procurement, the legal, the compliance aspects, and so on.
00:05:18 · Sebastian Johnson
[00:05:18] S0: But it must be amazing to have almost, like, internal advocates fighting for your product or or at least this type of change. You mentioned kind of product road map and you're moving from, I guess, you're on this journey from being like a scribe tool, which is your initial product or initial kind of wedge to now this sort of like AI native operating system for clinicians. How are you deciding what to build next for the clinicians that you're serving?
00:05:44 · Lukas Saari
[00:05:44] S1: Yeah. I mean, so we're working very closely with our customers and, like, having these regular check ins and workshops with their CIOs and so on to understand, like, what are the additional challenges and issues that you are seeing. And then from our side, also seeing what is it that we can develop that we can apply across Europe, like across care settings, providers, and so on. So if you think at the core of it from, like, a patient journey perspective of how does it look today and how is it starting to look now when you use Tandem instead. So if you, as a patient, are feeling sick and you're calling into your clinic, you would today be calling in between eight and nine in the morning when there's a phone slot available. You sit in a phone queue for half an hour, and then you get told, sorry. We're fully booked today. Please try calling again tomorrow and call a bit earlier because then it might not be fully booked yet. So, like, not a great experience for you as a patient, but also very time consuming for the clinic. How can you instead have what is rather like an AI receptionist that responds to the phone call, can do triaging, can do scheduling, handle simpler tasks as prescription renewals and so on, and be available twenty four seven. And then when you come into the clinic, how can you... Instead of having the clinician sitting and reading through Lukas's past 50 medical notes to figure out who is Lukas, why is he here today, what should I do with him, Instead, just have an LLM summarize that for you. And then during the visit, if you as a clinician are unsure about something, instead of using Google or sitting with control f in various PDFs and so on, can you just have, like, an AI with you in the room that's listening and has access to the patient context, the right guidelines? And then after the visit, instead of manually doing the note taking by hand, the coding by hand, the referral by hand, the prescription by hand, the follow-up appointment booking by hand, and so on, having this, like, complete AI medical assistant that takes care of all of this for you. So...
00:07:29 · Sebastian Johnson
[00:07:29] S0: Yep. I was gonna say, are you betting on a... You know, along that stack from, you know, the phone call to the post night... Post notes, write up, whatever, are you betting on specific areas right now? Are there areas of focus that you really think is gonna unlock the next kind of area for you?
00:07:43 · Lukas Saari
[00:07:43] S1: So all of the things that I described now, we have live at various stages. So, I mean, a year ago, we were more in the scribe space, expanded to this, like, AI medical assistant, and now doing more of also the patient facing and patient flow aspects that we're bringing across. And then how I'm thinking about it with this, like, grander AI native operating system is really breaking down and thinking more first principle. Like, what are the 20 categories of tasks that you need to do at a clinic? And I listed a bunch of those in the use cases that I just described, but there's also many others. And then saying, how can we go to a clinic and say, what we now want to do is to help you sort of automatically handle all of these 20 different types of tasks. And we can either support this at a level zero, which is we don't do it at all. A level one, that is that we provide a draft to the clinician and they act on it manually. Level two, we provide a draft to the clinician and they come press yes, and then it's automatically handled. Or level three, you can just do it without the clinician being in the loop. And now wanting to see, okay, how can we go to clinics and then say, what we want to do with you over the coming few months is that we want to bring all of these types of tasks that you're doing up to, like, level two, level three, and then using that as a type of framework for how do you bring a clinic into the AI era and sort of make them more and more agentic and automatic.