Latest / Breaking the ICE / Episode 16 - What Our Customers Taught Us About Getting Started With AI

Transcript

Episode 16 - What Our Customers Taught Us About Getting Started With AI · Breaking the ICE

• 0:01
  SPEAKER_00: Welcome folks to Breaking the United episode 16, which means it's Friday, because we always film these podcasts on Fridays, which is uh which is fun.
• 0:09
  SPEAKER_00: And I've got a week off next week, so looking forward to uh well a few days.
• 0:13
  SPEAKER_00: A few days, yeah, you're off uh you're off to the well it's gonna be a bit blustery with you, aren't that?
• 0:17
  SPEAKER_00: By the looks against Devon Devon this weekend is not the right choice.
• 0:22
  SPEAKER_00: I know, I know.
• 0:22
  SPEAKER_01: The person that we're hurrying the police from going, You sure you want to come?
• 0:25
  SPEAKER_01: You know, it's quite you know.
• 0:27
  SPEAKER_01: I said, No, we'll tough it out.
• 0:29
  SPEAKER_01: It's too late now.
• 0:30
  SPEAKER_01: We're committed.
• 0:31
  SPEAKER_00: As you know, I'm in Dorset and uh and there's debris flying everywhere.
• 0:35
  SPEAKER_00: And um I've been sat in my in my room at the front of the house working today, and yeah, it's got a flat roof, and I'm like, oh god, please don't blow off this weekend, please.
• 0:43
  SPEAKER_00: I've got a week off next week, and I don't want to be uh to fix it.
• 0:46
  SPEAKER_00: Well, yeah, well, put a piece of tarpaulin up.
• 0:48
  SPEAKER_00: That's probably about as good as my uh my building skills um go, Mark.
• 0:52
  SPEAKER_00: Um how are you other than that?
• 0:54
  SPEAKER_00: All good?
• 0:54
  SPEAKER_00: Good to have you back on?
• 0:56
  SPEAKER_01: Um we had a client day last week, it's probably worth having a little bit of chat about that.
• 1:00
  SPEAKER_01: And once I get back next week, I'm going off.
• 1:03
  SPEAKER_01: I'm gonna be with Elliot, who he's not with us today, but um, we're gonna be in Istanbul, big travel conference where you know people are coming from all over to um you know sell their wares and us included.
• 1:16
  SPEAKER_01: So yeah, looking forward to that.
• 1:17
  SPEAKER_01: I've not been to Istanbul, so um yeah, yeah.
• 1:22
  SPEAKER_01: So so we had a client day recently, um had about 40 customers there um going through kind of what we're delivering on the policy and claims side.
• 1:37
  SPEAKER_01: Um but a big focus and a lot of interest, perhaps not surprisingly, um, was all around the kind of AI piece and the work that we're doing there.
• 1:47
  SPEAKER_01: We've invested big time probably last 12 to 18 months.
• 1:51
  SPEAKER_01: Um, and I think you know, in terms of where we are within the insurance domain, I think we're fairly well placed.
• 1:56
  SPEAKER_01: But I think you know, some of our customers still really yet to get going.
• 2:00
  SPEAKER_01: I mean, everyone's interested and kind of talking about it, but you know, there's a certain amount of nervousness there.
• 2:07
  SPEAKER_00: I think I think it's like a it's kind of an insurance thing, right?
• 2:11
  SPEAKER_00: We sort of procrastinate and talk about things for a while.
• 2:14
  SPEAKER_00: It's a risk-averse industry.
• 2:15
  SPEAKER_00: Yeah, yeah, yeah.
• 2:16
  SPEAKER_00: It's true.
• 2:16
  SPEAKER_00: And you know, I was on the underwriting side, so I, you know, I get it.
• 2:19
  SPEAKER_00: But I mean, you know, you I remember sort of the the telematics stage, you know, just it's a great idea.
• 2:24
  SPEAKER_00: Why are we still talking about it?
• 2:25
  SPEAKER_00: Let's just do more of it.
• 2:26
  SPEAKER_00: Um and and you're right, you know, AI is definitely in that in that place.
• 2:30
  SPEAKER_00: And I think actually what I can see these days is is is it's more of us picking it up and running with it than probably some of the businesses because I I'm using it all the time at the moment.
• 2:41
  SPEAKER_01: Yeah, I mean, yeah, I mean I completely agree.
• 2:44
  SPEAKER_01: I mean I mean a day goes by.
• 2:47
  SPEAKER_01: I don't think a day goes by where I'm not completely flabbergasted by you know the potentiality of what we've got.
• 2:55
  SPEAKER_00: I think from an ICE perspective, you know, one of the things that's been good from for me sp in particular in that kind of sales space, and and will be the same with you.
• 3:03
  SPEAKER_00: I mean, you you've obviously been with ICE a a little bit less than what I have, but actually being able to, it's been a really good tool, actually, because we've kind of constructed our tools and built them in a way that they don't necessarily need to kind of be in symbiosis with the platform policy and claims platforms and other ICE tech, and we can sell them as individual components.
• 3:21
  SPEAKER_00: It's been a really good tool, actually, to go back out and just drum up a bit of conversation and say, look, guys, have you heard about these other things we're doing?
• 3:26
  SPEAKER_00: So really quite cool.
• 3:27
  SPEAKER_01: Yeah, no, I can completely agree.
• 3:29
  SPEAKER_01: And you know, we were making the point at the um at the uh at the customer day that um you know we're drinking our own champagne um or or eating our own dog food, whichever way you want to look at it.
• 3:43
  SPEAKER_01: But you know, the tools that we're building to enable our customers, we're already using them ourselves to orchestrate, and we'll we'll talk maybe about that a bit in a second, to orchestrate the complete development life cycle.
• 3:57
  SPEAKER_01: Um, so importantly, we've still got humans in the loop at the appropriate place.
• 4:02
  SPEAKER_01: But um Al, who I think is going to come on the podcast at some point soon, he talks about the shift to the left.
• 4:11
  SPEAKER_01: So you think about the the elements of um no political stuff here, Mark.
• 4:17
  SPEAKER_01: No, no, you're saying a safe space.
• 4:20
  SPEAKER_01: When you think about the development lifecycle, on the left hand side, I don't know if this works with a camera, but anyway, on the left hand side, you've got all of that specification, requirements, capture activity, architecture, then you get into coding, then you get into testing, and then you get into release.
• 4:35
  SPEAKER_01: Now, traditionally, for kind of software companies, the focus on the technical side was very much in the middle and on the right.
• 4:43
  SPEAKER_01: And then you would have kind of business owners and BAs and product specialists who kind of lived in that left-hand part of the world.
• 4:52
  SPEAKER_01: Um, you know, and that's definitely kind of where I've sort of inhabited um my career.
• 4:58
  SPEAKER_01: But because of the, you know, the the capability of AI, we're able now, well, you know, Al will talk to it.
• 5:06
  SPEAKER_01: You know, we're less interested in the deep technical skills.
• 5:11
  SPEAKER_01: Clearly, they're important, and the stuff that Piotr and the AI team, they're really into the nuts and bolts of this.
• 5:17
  SPEAKER_01: But in terms of the end solutions that we end up delivering, it's much more about understanding the business, understanding the requirements.
• 5:24
  SPEAKER_01: So a lot more effort going in up front.
• 5:27
  SPEAKER_01: Um, and then we can just really turn the handle and get the software out as a result of that.
• 5:33
  SPEAKER_01: So that's kind of that that shift to the left.
• 5:36
  SPEAKER_01: It's really interesting.
• 5:37
  SPEAKER_01: And you know, as I say, we're we're using it internally, and it's making you know a big difference.
• 5:43
  SPEAKER_00: And I think, you know, in in fairness, you know, kind of called out the fact that people are still talking about it and not moving with it.
• 5:49
  SPEAKER_00: That's not necessarily the case now from an ICE perspective.
• 5:51
  SPEAKER_00: We've got a couple of customers now that are rolling this stuff out.
• 5:54
  SPEAKER_00: Yeah, that they're out there, they're doing it.
• 5:56
  SPEAKER_00: We've got some proof points, and and I I think, you know, certainly going back to that whole sales piece again, and I had this conversation with someone in the office yesterday, you know, when you've when you've won the first tier one, it's easier to win the second tier one.
• 6:07
  SPEAKER_00: When you've won, when you've got the next territory, it's it's easier to get the territory after that.
• 6:10
  SPEAKER_00: When you've sold the first piece of AI and prove the point, you know, again, those conversations should get easier.
• 6:15
  SPEAKER_00: So it's it's exciting to see some of that stuff in action.
• 6:18
  SPEAKER_00: And I think, you know, from my perspective, if I was to pick on one of our kind of AI solutions at the moment, and there's a fair few of them and they're changing all the time, um, the orchestration tool is is pretty mesmerizing, even for someone without a technical background, seeing it kind of in action, just doing and creating and iterating is is really cool.
• 6:39
  SPEAKER_01: Yeah, I mean it, I mean it is literally transformational.
• 6:42
  SPEAKER_01: And and and I think, you know, some to some extent that's where a little bit of the kind of nervousness comes from, because customers in a position where they can actually completely reimagine their business.
• 6:54
  SPEAKER_01: And you know, not just like you know, what's going on in my department, it's about well, what's going on in my company and my supply chain and you know, just the whole end-to-end process.
• 7:04
  SPEAKER_01: We can completely think about it differently.
• 7:07
  SPEAKER_01: And technologies like Alice Orchestrator, which, you know, by the way, is completely application agnostic, will run over any application, not just our stuff, um, you know, really kind of opens that up.
• 7:20
  SPEAKER_01: I think um and and and I think you know, terms in terms of kind of our customers, we can kind of offer them the best of both worlds, really.
• 7:30
  SPEAKER_01: You know, you can take an evolutionary approach and you can take a transformational approach.
• 7:34
  SPEAKER_01: And I guess the advantage really of kind of doing that within ICE, although it I don't think it's necessarily directly tied to ICE, is you know, you can do it both within the same platform.
• 7:43
  SPEAKER_01: So, you know, our customers, you know, what we're saying to our customers, look, just just just just find a point of friction.
• 7:50
  SPEAKER_01: It might be an email box that somebody has to check every so often to see if something's come in, you know, something fairly trivial.
• 7:59
  SPEAKER_01: I mean, that stuff like that is absolutely absolutely right to say, oh yeah, well, an AI can pick that up and give it to someone, or an AI can pick that up, read it, decide where to send it, or an AI can pick that up, read the contents, extract it, and put it in your application, you know, and then you can just gradually kind of evolve and sort of go from kind of evolution much towards you know that kind of change in the business that's truly transformational.
• 8:26
  SPEAKER_00: That's a that's a good point.
• 8:28
  SPEAKER_00: I mean, you know, insurance as a whole, and and again, I've been on the insurance side of the of the table, it it's it's literally kind of laced with you know manual workarounds, manual process.
• 8:39
  SPEAKER_00: And and I can remember working at an organization where that list about there was a project just to get that list from like 900 manual workarounds down to well, we want two, we won 300 because you know that's manageable, that's workable.
• 8:51
  SPEAKER_00: You know, actually what we're saying now is is you know, to your point, find us, find us something, give us a problem to solve, um, and we'll and we'll we'll do it using the orchestration tooling.
• 9:01
  SPEAKER_00: But I mean, uh how do you pick it?
• 9:03
  SPEAKER_00: What's what's the first one?
• 9:04
  SPEAKER_00: What how do you pick your first kind of workflow?
• 9:05
  SPEAKER_00: What what do we do with it?
• 9:06
  SPEAKER_01: I I I just you just I think you you know, I don't think you necessarily have to overthink it too much, you know, just just find that point of friction.
• 9:14
  SPEAKER_01: Um, I think you know, if you get a bunch of people in a room, you know, with around a business process, they'll very quickly say, Oh, yeah, well, I do that, and it's a bit of a pain, or I have to write that thing down four times and I don't know why I'm doing that.
• 9:25
  SPEAKER_01: I mean, literally, we were visiting clients a couple of weeks ago and we're discovering this stuff.
• 9:29
  SPEAKER_01: Um, oh, great example of we visited a customer, um, pet insurance, because obviously that's kind of um home territory for me.
• 9:39
  SPEAKER_01: Um and when the pet dies, quite often the customer's upset, they forget to cancel their insurance.
• 9:45
  SPEAKER_01: Um, and really it's only kind of after the fact that that they then notify the policy team that oh I forgot it surely my dog died three months ago.
• 9:56
  SPEAKER_01: So, what we can do is as soon as we spot like um a death benefit claim, um using Orchestrator, feed that information back to the policy team um and automatically cancel that policy.
• 10:11
  SPEAKER_01: Today, somebody fills out a form, it lands in an inbox, it then requires that somebody reads that inbox, they then have to find that record um and then go and cancel that policy.
• 10:22
  SPEAKER_01: Um and and it's super easy.
• 10:25
  SPEAKER_01: We just say, Yeah, okay, we can give that to Alice Orchestrator, and by the way, you don't need to code that, you just literally chat to Alice and say, This is what I want to do, and it goes and builds the process in front of your very eyes.
• 10:37
  SPEAKER_01: Um and so, you know, and I think you know, once you once you get going, I think very quickly, you know, people will just just completely dive in and you know, there's there's huge opportunity there.
• 10:47
  SPEAKER_01: But I mean, and with huge opportunity comes huge responsibility as well, or or with great power comes great responsibility.
• 10:54
  SPEAKER_01: So I think you know the the orchestration layer gives that will give our customers the kind of control that they need.
• 11:00
  SPEAKER_01: So yeah, super interesting.
• 11:01
  SPEAKER_00: Yeah, we've had that conversation around config before and you know making sure it's in the right hands and and kind of gated off in the right way and protected in the right way, which is is super important.
• 11:11
  SPEAKER_00: I guess you know, with AI, you know, I I think the temptation with AI at the moment is you know that everyone's looking for the biggest use case.
• 11:19
  SPEAKER_00: Everyone thinks it's this massive thing.
• 11:21
  SPEAKER_00: And actually, you know, you should probably be doing the opposite to your point.
• 11:24
  SPEAKER_00: Find a process that you know, maybe it's maybe maybe it is that.
• 11:28
  SPEAKER_00: Find a process that really annoys everybody, takes a long time and it's got loads of manual steps, and that's the first place to start.
• 11:34
  SPEAKER_01: Might not necessarily be the biggest saving, but I think I mean that's I I really think that you know our customers or just organizations in general, they can really kind of embrace this and and do it in parallel.
• 11:46
  SPEAKER_01: You can go and have a team off over there doing the blue sky stuff, thinking about transformation and using something like Alice Orchestrator to start thinking about what that's going to do.
• 11:56
  SPEAKER_01: But there's no reason that the operational team that's facing the here and now at the same time can still make these kind of operational improvements very, very quickly using AI that has immediate material benefits.
• 12:09
  SPEAKER_01: So yeah, I mean, we definitely, you know, as I say, we're engaging with a number of our customers to try and get some of these things overline.
• 12:16
  SPEAKER_01: Like the thing we went to a couple of weeks ago, without even going into kind of deep deep conversations about transformational automation, very quickly we identified seven things that would, you know, save them half a million quid a year.
• 12:30
  SPEAKER_01: So, you know, yeah, amazing.
• 12:33
  SPEAKER_00: Who doesn't want that, right?
• 12:35
  SPEAKER_00: Who doesn't want that?
• 12:35
  SPEAKER_00: Come chat to us, let's see if we can help.
• 12:37
  SPEAKER_00: I mean, you you talked about the customer day.
• 12:39
  SPEAKER_00: Did did did did the customers sort of give any feedback or thoughts or kind of um you know where their pain points were or you know, demonstrating some of this stuff and you know you had that light bulb moment.
• 12:51
  SPEAKER_00: Was there any of that kind of uh that kind of thing?
• 12:53
  SPEAKER_01: No, not so much.
• 12:53
  SPEAKER_01: I mean, I think people may be a little bit kind of reserved about talking about their individual processes when they were in a in a group.
• 12:59
  SPEAKER_01: There were a few breakouts and didn't get to all of them, so maybe kind of it came up more there, but um uh you know, I I I still think I think very quickly once you engage with with customers in these kind of conversations, you very quickly can pick up things that that we can just automate.
• 13:17
  SPEAKER_01: Um yeah.
• 13:19
  SPEAKER_01: And I think I mean one of the things one of the things that we spoke about, which is really interesting, is the kind of the transition using AI for for small incremental business gains, but then again, using Orchestrator building out an agentic workforce where you've got a number of AI agents that you literally work alongside with.
• 13:52
  SPEAKER_01: You could imagine, like, you know, I don't know, you could have a claims team and you might have some junior claims handlers.
• 13:58
  SPEAKER_01: In the future, you'll have junior agentic AI claims handlers, and then when they get stuck, they can refer to you, you know, and ask you, well, what should I what should I be doing with this?
• 14:10
  SPEAKER_01: And you know, through teams, you can chat to them and discuss the conversation.
• 14:15
  SPEAKER_01: And, you know, like like a new person joining the company, oh, they'll go, oh, right, okay, that's interesting.
• 14:20
  SPEAKER_01: So, right, well, the next time I see that, um, you know, that's the way I'll handle it.
• 14:24
  SPEAKER_01: But crucially, you haven't just told one person, you've told that AI agent that's potentially working across, you know, a whole bunch of things.
• 14:30
  SPEAKER_01: And I think the key thing, you know, in you know, in terms of what's going on there, it's not so much about replacing the people, it's taking like the best people and getting them to focus on the hard problems as opposed to having them dealing with some of the mundane things that you you can automate.
• 14:45
  SPEAKER_01: And then some of the mundane things that you are automating are getting done, you know, consistently in the same way that they would do them.
• 14:52
  SPEAKER_01: So, you know, you would hope to see, you know, major increases in kind of productivity, but also the quality, the consistency and quality of service that that we're able to offer or our customers are able to offer their customers.
• 15:05
  SPEAKER_00: Yeah, I think um I think interestingly, you know, when you one of the things I I noticed with um with kind of our orchestration tool and seeing some of this stuff being pulled together, it's like you know, everybody thinks you need that AI piece.
• 15:16
  SPEAKER_00: That's the thing that's good, that's the silver bullet.
• 15:19
  SPEAKER_00: But actually, you know, ICE as an organization, I know this, you know, having been there for for as long as I have, we've been doing workflow and automat automation for years.
• 15:26
  SPEAKER_00: Um and actually seeing orchestrator not necessarily as part of building some of these processes, needing to even go to an AI source and incur a token cost, and actually it's it's a deterministic process.
• 15:38
  SPEAKER_00: And you look at it and go, actually, it's no AI in here at all, really.
• 15:41
  SPEAKER_00: It's just if if that, then this, um, you know, once that and that, and do do that over there.
• 15:46
  SPEAKER_00: So deterministic versus AI is uh is an interesting Yeah no, absolutely interesting conversation.
• 15:53
  SPEAKER_01: And and in terms of kind of how we deploy AI or how we're advocating deploying AI, what we want to do first is make sure that the processes are firstly deterministic, and then when things get a bit more nuanced and require human-level type assessment, then you can bring in the AIs, and then when the AIs don't know what to do, um, then you can get the human, you know, to intervene.
• 16:20
  SPEAKER_01: So it's yeah, it's really it's really interesting.
• 16:23
  SPEAKER_01: And and uh and and the fact that these processes can be built literally through a chat interface, and you can see, you know, you can see the process diagram being built in front of you, it's just incredible.
• 16:36
  SPEAKER_01: You know, you know, you don't need to go on a five-day training course to build business processes within our application now.
• 16:43
  SPEAKER_01: You need to be able to articulate the problem that you're looking to solve, and you know, you're off and running.
• 16:47
  SPEAKER_01: So it's gonna be really, really interesting to see kind of how our customers build out on that.
• 16:53
  SPEAKER_01: And yeah, and and all, you know, and also whilst you're building this thing, it's testing literally you're building it on the right-hand side and you're seeing the test running and the interactions taking place on the left-hand side.
• 17:04
  SPEAKER_01: And so if it's not right, you can stop it as it's going and just correct it as you go.
• 17:07
  SPEAKER_01: It's it's very intuitive.
• 17:09
  SPEAKER_00: So yeah.
• 17:10
  SPEAKER_00: I think it's just you know, it's just working out where these lanes are, aren't they?
• 17:14
  SPEAKER_00: You know, it's right.
• 17:15
  SPEAKER_00: This is a deterministic lane because the answer is very, you know, it's very black and white.
• 17:19
  SPEAKER_00: Let's let's put it down that path.
• 17:20
  SPEAKER_00: Oh, hey, we need to just overlay a little bit of AI over here, it needs to be something slightly different.
• 17:24
  SPEAKER_00: Oh, right, we've got a process within, I mean, I think back to kind of household insurance days where we used to, I mean, I worked within an over 50s insurer, and sadly, people over the age of 50 are a lot closer to to the end than people under the age of 50.
• 17:37
  SPEAKER_00: And and we had a we put a deceased uh team in place for things like unoccupied properties when someone passes away.
• 17:43
  SPEAKER_00: Well, they're difficult conversations to have, right?
• 17:45
  SPEAKER_00: You know, you've got somebody that's grieving, they're going for a really difficult time.
• 17:48
  SPEAKER_00: I don't want AI talking to that person.
• 17:50
  SPEAKER_00: I want somebody that's got some empathy that can that can really make them feel comfortable and and and sort of you know look out for them in in a difficult time.
• 17:57
  SPEAKER_00: So yeah, it's just to your point, kind of working out where these things sit.
• 18:01
  SPEAKER_00: Um, but yeah, interesting, interesting conversation.
• 18:04
  SPEAKER_00: I think you know the answer isn't AI everywhere, is it?
• 18:06
  SPEAKER_00: It's just working out where it needs to go.
• 18:08
  SPEAKER_01: Yeah, absolutely correct.
• 18:10
  SPEAKER_01: Yeah, it it's it's where AirI is adding value, yeah, yeah.
• 18:14
  SPEAKER_01: And you know, and and keeping the human in the loop, you know.
• 18:16
  SPEAKER_01: For example, you know, and my experience at TruePanion, um they were doing high levels of automation, uh using AI based on machine learning to automatically process thousands of claims every single day, but all of the declines they would go through a person.
• 18:37
  SPEAKER_01: You know.
• 18:38
  SPEAKER_01: So, you know, you don't maybe in the future, but who knows?
• 18:43
  SPEAKER_01: But you know, I think quite rightly they took the view, okay, well we're not going to reject a claim unless a person's looked at this first.
• 18:50
  SPEAKER_01: Even, you know, even if it's just rubber stamping, so you know, human in the loop are the right, you know, where it's appropriate to do so.
• 18:56
  SPEAKER_00: And I think kind of understanding, you know, how do you prove that it's worked?
• 19:01
  SPEAKER_00: You know, how do you prove that one of these processes has been a success?
• 19:04
  SPEAKER_00: Is it, you know, I want uh I want to achieve X amount of cost saving, I want to save this amount of time, I want to be able to, to your point, deploy this really good resource over here to go and do some of the moonshot stuff that we were kind of talking about.
• 19:17
  SPEAKER_00: I think you know, you need to you need to kind of have a bit of an idea in mind before you go down the path on some of this stuff to to understand what it looks like so that you can see if it's been successful or not.
• 19:25
  SPEAKER_01: And what you want to do.
• 19:26
  SPEAKER_01: Yeah, I mean I think I think a lot of it come will come down to operational efficiency, and I think quite a lot of that is quite measurable.
• 19:32
  SPEAKER_01: Um But I think, you know, I don't I'm not sure if I'm particularly smart enough, or I can't think of any good ideas right now, or uh, or and you know I can think of some stupid ideas, maybe they are good ideas, but you know, the that kind of transformational element of of how we interact with buy insurance.
• 19:52
  SPEAKER_01: I did on client day sort of give an example.
• 19:55
  SPEAKER_01: This might resonate with you, I don't know, but you know, when you fill out home insurance, frankly, it's a real pain in the backside.
• 20:02
  SPEAKER_01: You have to fill in loads of questions.
• 20:04
  SPEAKER_01: Have you got a flat roof?
• 20:05
  SPEAKER_01: I don't know what what your location is, are there any particular things you want to insure?
• 20:10
  SPEAKER_01: You just have to fill out a long boring form.
• 20:12
  SPEAKER_01: You could imagine, and I don't think it's very far off, you just walk around uh in video mode with your phone, you just walk around the house and just talk to the talk to the talk to the camera, go, oh yeah, I've got a watch there, there are a couple of bikes, I'll just show you them.
• 20:30
  SPEAKER_01: Uh, yeah, that's it, really.
• 20:32
  SPEAKER_01: And then you know, you hang up, hits pricing and underwriting, and within, you know, 10 seconds you've got a quote for your home insurance.
• 20:38
  SPEAKER_01: Like, you know, how easy could that, you know, that feels that feels like that's not far off.
• 20:44
  SPEAKER_01: Yeah.
• 20:45
  SPEAKER_01: Yeah.
• 20:45
  SPEAKER_01: And I just think, you know, everything, you know, will just become very, very it's funny, isn't it?
• 20:51
  SPEAKER_01: You sort of talk about AI, but I think things like that, you know, that's a very kind of human way of of interacting, as opposed to, oh, I've got to fill out this form, oh my god, I hate it.
• 21:01
  SPEAKER_00: I think that's interesting.
• 21:02
  SPEAKER_00: And and I was thinking about this the other day, actually, just in in terms of acquisition, right?
• 21:06
  SPEAKER_00: Acquisition of insurance.
• 21:07
  SPEAKER_00: We all know it's you know it's not really a tangible kind of thing, it's peace of mind, right?
• 21:11
  SPEAKER_00: That's where you're buying.
• 21:12
  SPEAKER_00: But you know, when you when you think back to how insurance was kind of sold, you know, you used to have kind of Jeff knock on the door on a Thursday and he'd, you know, he'd insure you and you'd pay him a couple of quid a week, and and that's how it worked.
• 21:22
  SPEAKER_00: And then all of a sudden, you know, you had your direct lines and your people that kind of cut brokers out, and it was just going direct to the insurer, and then all of a sudden you got this wonderful thing called the aggregators.
• 21:31
  SPEAKER_00: You know, you pop all those those details that you just talked about in one box and it goes off to all the insurers.
• 21:36
  SPEAKER_00: But actually, if I look at that, I think that's that's been quite damaging to to insurance from a consumer perspective because people, what they do is race to the bottom issues, yeah.
• 21:47
  SPEAKER_01: Yeah, exactly.
• 21:47
  SPEAKER_01: It's interesting.
• 21:48
  SPEAKER_01: I mean, I just as I'm talking about it, I don't even know if it's a thing, but like the aggregators definitely brought about what was referred to as uh disintermediation.
• 21:59
  SPEAKER_01: So you You no longer dealt with your insurer.
• 22:02
  SPEAKER_01: You y you went through this third party.
• 22:04
  SPEAKER_01: Could AI actually herald re remediation or you know, that that that closer connection?
• 22:11
  SPEAKER_00: I wonder whether it goes the other way though, because actually you think about it at the moment, you know, you've got lots of people that get to the point of making a claim and they go, What windscreen excess?
• 22:19
  SPEAKER_00: £350?
• 22:20
  SPEAKER_00: What do you mean?
• 22:22
  SPEAKER_00: Ten years ago it was 60 quid.
• 22:23
  SPEAKER_00: Well, no, no, no.
• 22:24
  SPEAKER_00: You bought it on the aggregators and you, you know, you went for that policy that you got for£270 for the year that you just went for because it was top of screen.
• 22:32
  SPEAKER_00: Actually, it might have been second or third because you did go for a little bit of brand recognition.
• 22:36
  SPEAKER_00: Well, the only way those insurers manage to get to the top of screen is by trimming out some of the kind of bells and whistles, and you might have missed a few bits and you're probably playing a bit more of an excess.
• 22:45
  SPEAKER_00: And then actually that creates resentment with the with the insurers, but it's not the insurers' fault because that's the way you you've got to acquire business now.
• 22:51
  SPEAKER_00: It's really difficult.
• 22:52
  SPEAKER_00: And actually, if you want to keep them and get them to renew, in order to do that, because you you solved the policy at a loss the year before, you've got to hike up the renewal price.
• 23:00
  SPEAKER_00: So again, that really aggravates the customer.
• 23:03
  SPEAKER_00: Um, so actually, is is AI going to take away the little bit that might still exist from a brand perspective?
• 23:09
  SPEAKER_00: Because actually, you're going into your AI, you know, your chat GPT as an example and saying, find me an insurance policy.
• 23:15
  SPEAKER_00: Here's here's this video that I've just taken of the house, here's a bit more information about me.
• 23:19
  SPEAKER_00: You just go crack on and do it.
• 23:20
  SPEAKER_00: And actually, there won't be any engagement with uh with even an aggregator at that point.
• 23:24
  SPEAKER_00: You might never see an insurance company again.
• 23:27
  SPEAKER_01: Well well, I mean, you you definitely make a good point about you know how these AIs, you know, and you read it in the paper, you know, uh MetLife in America, who else has done it?
• 23:37
  SPEAKER_01: Many pets in the UK, I know have done it.
• 23:40
  SPEAKER_01: The ability to get a quote direct through the through the through the AI, through Chat GPT is the one that I've seen.
• 23:47
  SPEAKER_01: I was gonna say, um, but this intermediate this intermediation point and you know that kind of race at the bottom to be at the top challenge you know, we are looking at and I think insurers are responding to that a bit by offering upsell once you get onto their um once you get onto their direct site.
• 24:14
  SPEAKER_01: And I think AI at that point, you know, you know we've got Alice Sales, that conversational capability to understand what the customer is shopping for, you know, to take your example.
• 24:29
  SPEAKER_01: Look, this is quite a cheap policy, but just be aware, yeah, it's got a 300 pound excess on on the uh on the windscreen.
• 24:36
  SPEAKER_01: I can offer you this other one, and the excess is lower, you know.
• 24:40
  SPEAKER_01: And we've seen loads of examples or you know, if you buy a particular motorbike, it knows that this person, based on that profile, is like to have the expensive leathers, and then you can offer cover for that.
• 24:49
  SPEAKER_01: So I think I think I'm curious to see whether the industry shifts to a more modularised approach um to kind of counter some of those negative effects of those kind of ag journeys.
• 25:05
  SPEAKER_01: It's a bit like when you buy a ticket for an aeroplane on Ryanair, which is a complete soul-destroying exercise, because you then have to like avoid the minefield of being upsold and you know, an extra three thousand quid as you buy the champagne breakfast and the the helicopter to the airport, or maybe just pick a seat.
• 25:31
  SPEAKER_01: I think that could be done a lot better.
• 25:33
  SPEAKER_01: I think that could be done a lot better uh through this kind of AI chat capabilities.
• 25:38
  SPEAKER_00: That's the bit actually that I think could be really quite powerful in that kind of acquisition stage of actually, as we know, the more we interact with these kind of agents and and and kind of AI interfaces, the more they start to learn and understand and know about you in quite quite specific detail sometimes.
• 25:55
  SPEAKER_00: So actually being able to kind of interact with, say, like that aggregation type step and look at it and go, Oh Lloyd, do you realise that you know this is definitely not right for you because you know you've got X amount of some insured and you've got a couple of these bikes here, that these guys just aren't going to cover it in the event of a claim.
• 26:11
  SPEAKER_00: Um, and and be able to give you a bit of a of good quality advice as part of that.
• 26:15
  SPEAKER_00: Again, we get into the kind of grey area of providing advice at the point of that process.
• 26:20
  SPEAKER_00: Is that right?
• 26:21
  SPEAKER_00: Is that wrong?
• 26:22
  SPEAKER_00: Not sure.
• 26:22
  SPEAKER_00: But you know, I I it is interesting because I think people have tried it with like, you know, putting a witch rating on a policy or you know, I'm de facto five star versus de facto three star.
• 26:31
  SPEAKER_00: But I mean, you know, I've I've designed policy wordings where we've got de facto five star, and I've got the extra few points I needed by offering holding one cover on a home insurance policy, and it's you know, it you can game it a bit, you can game it a bit.
• 26:43
  SPEAKER_00: So I think you know, trying to having that is cool.
• 26:48
  SPEAKER_01: I think those I think those chat kind of interactions are really useful.
• 26:51
  SPEAKER_01: Like you say, you have to be careful about how you're positioning some of the things that you're proposing, but I think I think it can be done in a in a way where perhaps you know people are making on balance decisions.
• 27:05
  SPEAKER_01: And I think you know, the way it's positioned is quite fair, really.
• 27:10
  SPEAKER_01: You know, you know, for an extra 22p a day, you know, we'll we'll cover you for this.
• 27:15
  SPEAKER_01: It's quite a quite a nice way of looking at it.
• 27:18
  SPEAKER_01: What's also interesting is is like those chat things, you know, people are asking questions about, you know, do I have an access for um you know for my windscreen?
• 27:27
  SPEAKER_01: Would I be covered in the event of X, Y, and Z?
• 27:31
  SPEAKER_01: All of that information gets captured on the back end, and that's incredibly useful, like after the fact to go, oh look, we've obviously got you know a common kind of question point here.
• 27:44
  SPEAKER_01: We need to redesign our flow.
• 27:46
  SPEAKER_01: So not only I think are we able to better interact with our customers, but the journeys and the kind of quality of data that we capture as as um as customers are going through these journeys just can be you know much better as well and ultimately lead to to better outcomes across the board.
• 28:02
  SPEAKER_01: So yeah.
• 28:03
  SPEAKER_00: Yeah, I mean I guess excuse the the kind of analogy, but you know, I I get the impression that some of this stuff is going to be like a drug, right?
• 28:11
  SPEAKER_00: You you get it, you get it right, and then you want to do the next one because you know you've seen the benefit.
• 28:16
  SPEAKER_00: But you know, what if that doesn't happen?
• 28:18
  SPEAKER_00: How do you stop the first kind of workflow improvement process becoming a one-off experiment?
• 28:22
  SPEAKER_00: Because again, you know, we've we've seen these kind of innovation hubs within insurance businesses that set something up and do something, and then before you know it, it just gets kind of amalgamated onto the mothership and and it's no longer an innovation anymore.
• 28:34
  SPEAKER_01: Yeah, so I I mean I think there are definitely two schools of thought here.
• 28:37
  SPEAKER_01: You know, you've got you've got a school of thought that says, right, if we're gonna inter innovate, we're gonna go over there and we're gonna, you know, forget about everything that we've done before.
• 28:48
  SPEAKER_01: You might put that in the boil the ocean bucket.
• 28:51
  SPEAKER_01: Um, you know, that's one way.
• 28:56
  SPEAKER_01: I mean, I go back to what I said before, you know, find the friction points, just go for one thing.
• 29:02
  SPEAKER_01: I think very quickly people can get this stuff over the line, prove it, measure it, you know, measure it, deploy it, measure it, and prove it, and then just move on to the adjacent thing.
• 29:13
  SPEAKER_01: So I think, you know, as I say, there's a lot of value in just doing stuff incrementally.
• 29:17
  SPEAKER_01: Um and then maybe, you know, once you've you know kind of mined that scene, you'll have a much better sense of oh, actually, if looking from a transformation perspective, knowing what we know now, with the capabilities and the tools that we've got on our hand, we can do this completely differently.
• 29:33
  SPEAKER_01: Um, so I think it's really just a question of you know, helping our customers move up that experience curve and getting comfortable.
• 29:39
  SPEAKER_00: I think the powerful thing with with Orchestrator as well, right?
• 29:42
  SPEAKER_00: You know, I think it I I think we're looking to operate this in a similar way to how we we roll out the core application, which is you know, let's let's do a couple of these with you, let's hold your hand, let's let's go on a journey together.
• 29:54
  SPEAKER_00: But actually, you know, we've we've created yet again another really configurable application where we want to get you guys doing that and we want to and we want to show it, you know, working in that way.
• 30:04
  SPEAKER_00: And actually, once you've done that first workflow or two or three or whatever we do together, you guys are then, you know, or our customers, sorry, are are then going out and and doing it themselves with very little interaction from us.
• 30:13
  SPEAKER_01: Yeah, absolutely.
• 30:14
  SPEAKER_00: Yeah, which is definitely what we want to be pushing.
• 30:16
  SPEAKER_00: Yeah, I agree.
• 30:18
  SPEAKER_00: Cool, cool.
• 30:19
  SPEAKER_00: Well, it was uh I guess a bit of an uh a kind of you know conversational one this week.
• 30:25
  SPEAKER_00: I know we've got Al coming in, as you said, who's gonna talk a bit more about some of the actual use cases of of some of our clients using it in the real ways, which I think is gonna be cool.
• 30:35
  SPEAKER_00: Um, yeah, I guess I didn't go to the the customer day this time, but you know you did, and and yeah, big thanks to all our customers that turned up.
• 30:42
  SPEAKER_00: You know, we are a we're very much a customer focused business.
• 30:45
  SPEAKER_00: We want to we want to understand what's going on and and and help our customers where possible as well as when as well as win new m new ones.
• 30:52
  SPEAKER_00: So that kind of brings me on to just something I wanted to sort of close with, Mark.
• 30:56
  SPEAKER_00: I mean, you're at uh an event in Turkey, um, certainly different to the one I met in in Angel in London in in the same week.
• 31:04
  SPEAKER_00: Um but Ice ICE is exhibiting at the Insurance Innovators Conference Um as well on uh I think the third and fourth of November.
• 31:14
  SPEAKER_00: We've got a stand, we're a sponsor, um, we've got some guests coming along that are clients, we've got a few prospects along as well.
• 31:21
  SPEAKER_00: Um, got a number of the team that are gonna be there shows showing some of the kind of AI capabilities that we've got on offer.
• 31:26
  SPEAKER_00: So, yeah, come come and find us.
• 31:28
  SPEAKER_00: I'll be there, reach out on LinkedIn, reach out to Mark on LinkedIn, he's gonna be in Turkey talking travel.
• 31:34
  SPEAKER_00: But yeah, we've got lots going on on in the event space over the next couple of weeks, so come find us and come and learn more about this stuff.
• 31:42
  SPEAKER_00: And on that bombshell, on that bombshell, a couple of ice mascots in the background.
• 31:48
  SPEAKER_00: Okay, alright, thanks.
• 31:49
  SPEAKER_00: Yeah, cool.
• 31:51
  SPEAKER_00: Um, on that note, thanks ever so much, everybody listening.
• 31:54
  SPEAKER_00: Come find us at our events.
• 31:56
  SPEAKER_00: Goodbye to Mark and his uh and his dogs, and um have a great weekend, and we'll see you on the next one.
• 32:01
  SPEAKER_00: Cheers.
• 32:02
  SPEAKER_00: Cheers, guys.