WEBVTT
00:00:08.400 --> 00:00:10.640
Hello and welcome to CX Today.
00:00:10.800 --> 00:00:15.679
I'm Reese Fisher, Associate Editor, and today I'm very fortunate to be joined by two people.
00:00:15.759 --> 00:00:22.320
I'm here with Russell White, the transformation lead at Bluecrest, and Rhys Harper, the account executive at Nice.
00:00:22.719 --> 00:00:23.679
Thanks for joining me, guys.
00:00:23.760 --> 00:00:25.199
How are you both doing today?
00:00:26.399 --> 00:00:27.199
Really good, thank you.
00:00:27.359 --> 00:00:28.559
It's a good question to meet you.
00:00:29.679 --> 00:00:30.239
Yeah, great.
00:00:30.320 --> 00:00:33.520
It's great to be on here and uh have a good conversation today.
00:00:34.079 --> 00:00:35.439
Yeah, yeah, absolutely.
00:00:35.520 --> 00:00:51.280
You know, we're gonna be discussing uh a recent collaboration between Bluecrest and NICE, you know, and really I guess using that as a bit of a lens to dig into the relationship between contact centres and customer data, which has been a little fraught at times, I think.
00:00:51.359 --> 00:00:58.159
So we're gonna look at you know where contact centres get it right, where they get it wrong, the role AI plays, the role humans play.
00:00:58.320 --> 00:01:01.280
So yeah, I think it's gonna be a really, really interesting chat.
00:01:01.359 --> 00:01:05.359
And I guess off the back of that, Russell, I wanted to start with you.
00:01:06.159 --> 00:01:19.519
You know, some of the guess specifics of the collab with NICE, one of the big changes was Bluecrest moving from reviewing a sample of calls to analyzing every interaction.
00:01:19.760 --> 00:01:27.439
What did you maybe discover once that you could see that entire picture that you couldn't um that you were maybe missing before?
00:01:29.120 --> 00:01:48.079
Um, so yeah, uh obviously Bluecrest has uh implemented the the nice CX1 system, and as part of that package, we uh purchased interaction analytics, which enables us to uh review every single call, uh categorize those calls, uh, and also use the topic topic AI functionality within the system.
00:01:48.239 --> 00:01:57.519
We're really able to uh um capture information that we weren't able to capture at such a wide scale before, particularly with customer intent.
00:01:57.680 --> 00:02:02.719
So uh our contact center is massively important to us at BlueCrest Wellness.
00:02:02.799 --> 00:02:06.079
The contact centre really is the front door to preventative care.
00:02:06.159 --> 00:02:10.800
Every call we we answer is potentially a health problem found early.
00:02:11.039 --> 00:02:40.240
And knowing why our customers are contacting us and being able to segment that data and analyse that data um opens up a wealth of opportunity for um automation flows and and how we can um uh provide a better service by really splitting those calls out to the to the right people um and having the right automation flows or right processes in place to handle those interactions.
00:02:42.080 --> 00:02:42.879
Yeah, yeah.
00:02:43.039 --> 00:02:45.280
I guess I wanted to come to you next, Rhys.
00:02:45.599 --> 00:02:54.080
Obviously, Russell, you touched on the obvious being able to see all this information is one thing, but then actually acting on that, what you're being able to do off the back of that is obviously the really important next step.
00:02:54.159 --> 00:03:06.400
But you know, I was reading the report around this and I saw that some of the findings were used to improve access, uh reduce after call work, and kind of shape how your advisors handle those customer needs, which you've touched on there, Russell.
00:03:06.639 --> 00:03:16.319
Reese, what do you think separates a contact center that collects more insights from one that actually changes how it operates because of those insights?
00:03:16.719 --> 00:03:17.039
Definitely.
00:03:17.120 --> 00:03:21.439
So I think you obviously here at NICE, we run customer service for the enterprise.
00:03:21.520 --> 00:03:29.039
And what the what's been really great about working with Russell and the Bluecrest team is that the technology is able to deliver that for our customers.
00:03:29.199 --> 00:03:34.560
But it's ultimately about what you do with that information and how you build that into the people and the process.
00:03:34.719 --> 00:03:38.479
And I think you know, since we engaged and worked with Russell, we were always very clear.
00:03:38.639 --> 00:04:08.400
Well, Russell was very clear with us about what we were looking to achieve as part of this project and what you know the end goal was, um, and being able to continually optimize and use that to the technology and the analytics to be able to make informed decisions and further guide that roadmap moving forwards has been something, you know, as you were seeing with some of the um outcomes that Bluecrest has been able to achieve has really been based on that fundamental approach from the beginning to make the technology um work for the organization.
00:04:08.879 --> 00:04:09.199
Yeah.
00:04:09.599 --> 00:04:16.800
I think for me, yeah, our disposition has always told us like the outcome of the call, but not really why people were calling.
00:04:16.879 --> 00:04:25.680
And we guided to the outcome on the call, but having that that campaign attribution built around the the actual intent is very different.
00:04:25.839 --> 00:04:29.360
Um, and it gave us a real clear reporting picture of the customer.
00:04:29.600 --> 00:04:34.079
I mean, the the first thing you find is that the anecdote and reality are two different things.
00:04:34.240 --> 00:04:45.199
What advisors and teammates believe he's had now calls in good faith is not often what the transit trips show, um, which we'll we can talk about a little bit more further further later on.
00:04:45.360 --> 00:04:48.800
But um, and the other thing really is where the time goes.
00:04:48.959 --> 00:05:09.120
If you if you think about at the beginning of of every call, the actual the um the process of of customer identification and idea verification um is really uh you can really drill down onto how long that process is actually taking and and the different ways people handle it and unifying the approach on every call to maximize efficiency.
00:05:09.439 --> 00:05:15.120
Um so yeah, it it really was a big improvement for us having that information available.
00:05:16.959 --> 00:05:18.639
Yeah, yeah, I can imagine.
00:05:18.800 --> 00:05:25.439
You touch on the perhaps the the reality versus maybe some internal assumptions you may have had beforehand.
00:05:25.759 --> 00:05:33.920
Could you give us maybe an example where customer conversations pointed Bluecrest in a different direction from one perhaps that you guys expected?
00:05:35.439 --> 00:06:02.720
Um we weren't aware of really the the volume of calls that that where customers were reaching out with us because of technical difficulties, so it really enabled our uh IT and development teams to uh drill down and and concentrate on those um those um journeys through our website, through our well um our wellness app and identify um common issues that that customers were contacting us for.
00:06:02.959 --> 00:06:16.000
The probably the misconception side of things, um, a really good example of that is that um we uh launched self-serve um for our customers to be able to reschedule their appointments online.
00:06:16.480 --> 00:06:32.240
Uh, and obviously the business uh methodology behind that was it would there would be uh an um process improvement, so it's better for the customers just being able to reschedule online, but also that that would detract uh call volume into the contact center and be an efficiency gain for us.
00:06:32.319 --> 00:06:36.000
So it's win for the customer and win for our efficiency as well, really.
00:06:36.319 --> 00:06:44.879
Um, and so we launched self-serve on on reschedules and anticipated a drop in call volume to the contact center that didn't happen.
00:06:44.959 --> 00:07:00.240
Uh, and interaction analytics and and using AI and those tools really enabled us to um visualize the fact that the the intent, the reason people were uh the rescheduled intent was still just as high as it was prior to that self-surf.
00:07:00.319 --> 00:07:11.759
And in fact, by launching the self-surf, what we'd really enabled our customers was to do was to give them more flexibility in the bit in rescheduling, which just meant we got a higher volume of reschedules.
00:07:11.920 --> 00:07:20.240
So the business misconception was that this would be a huge efficiency gain for us, and actually it wasn't it, but obviously the benefit was still there for the customer.
00:07:21.439 --> 00:07:24.480
Yeah, yeah, that's uh yeah, that's a great example.
00:07:24.720 --> 00:07:31.040
Um, Reese, I wanted to come back to you again, picking out some more of Bluecrest's kind of results from the collab.
00:07:31.279 --> 00:07:38.160
You know, we've seen uh sort of better access, lower after call work, and more relevant customer conversations.
00:07:38.560 --> 00:07:50.560
I guess from your perspective at NICE, when organizations do deploy the interaction analytics that Russell's been talking about, where do they maybe often get the value calculation wrong?
00:07:50.959 --> 00:08:01.920
I think, you know, similar to what Russell said earlier, is not going in there, you know, with a hypothesis or, you know, for what you think is in the data and the analytics give you the answer.
00:08:02.079 --> 00:08:39.919
And I think, you know, what we we see with a lot of our customers is that changes on a week-to-week basis, and being able to have that enterprise level insight and uh changing on a regular basis to be able to guide those business decisions, like Russell mentioned, you know, whether that's a change in something we're offering to a customer or it's a change in the competitive landscape, being able to use that, and I know that Russell does this quite a lot, is feeding that back to the business so that we can use that data to make a not just within the contact center and the CX experience, but there's a lot of wider benefits we see from our customers, not just in operationally, but also wider across their um, you know, across their organization as well.
00:08:40.080 --> 00:08:53.120
And I think that's really where we see a lot of the um unexpected benefits from our customers is when they really truly adopt this, there's a lot more wider benefits than just the operational savings and customer experience as well.
00:08:53.519 --> 00:08:56.320
It's interesting you said that recently as well about the new product.
00:08:56.720 --> 00:09:01.600
I mean, that was another really clear example for BlueCrest on the use of interaction analytics.
00:09:01.759 --> 00:09:06.000
We launched a new product, a product for our customers uh back in April 2025.
00:09:06.080 --> 00:09:11.919
We launched Health Monitor, which is a subscription service that brings customers back every year for an assessment.
00:09:12.000 --> 00:09:17.200
And the initial launch figures um and um conversion rates weren't what we expected.
00:09:17.360 --> 00:09:26.159
Uh, it's a very good product, very strong product, and and it it the importance of regular health monitoring is a really key differentiator for BlueQuest.
00:09:26.240 --> 00:09:29.759
And we weren't seeing the take-up levels that we we anticipated.
00:09:30.080 --> 00:09:35.679
Um, and our advisors were already assuring us that that they were offering it to customers on every call.
00:09:35.840 --> 00:09:44.000
Um, and we were getting a good response online um and a good response from our mail jobs, but not so much from the calls coming into the contact center.
00:09:44.080 --> 00:10:12.480
And interaction analytics enabled us to actually drill down into those calls and actually see that it was only being mentioned on an under 10% of relevant calls, but aged behavior was dictating that they were sticking to their RAL scripts, and and you get that with with uh sort of our advisors where that they're used to selling a product or used to discussing products in particular ways, um, and that enabled us to target training at those particular agents that weren't mentioning this new product, and we saw considerable increase of that.
00:10:12.639 --> 00:10:18.080
The same week we we actually fed that back to the to the team leaders and the contact center merger.
00:10:18.159 --> 00:10:23.440
Uh, it jumped up to 40%, it's now closer to the 60-70% mark.
00:10:23.519 --> 00:10:32.320
Uh, and we have um uh nearly 4,000 health monitor subscriptions now in place, and that insight wouldn't have been available before without interaction analytics.
00:10:32.639 --> 00:10:40.879
We would have we were only sampling a very small number of calls, and the cool and the calls that we were sampling weren't highlighting the scope of the problem in the contact centre.
00:10:42.720 --> 00:10:44.879
Yeah, yeah, it's another really incredible result.
00:10:45.120 --> 00:10:57.200
I find it I think one of the parts of the the collab particularly interesting is obviously you guys are are based in the healthcare sector, and we know how vitally important accuracy is in a sector like that.
00:10:57.279 --> 00:11:03.200
You know, summary can become part of a customer record, so getting it wrong carries that real consequence.
00:11:03.919 --> 00:11:09.360
I was just wondering why do you think human review was kind of a non-negotiable for you guys?
00:11:09.440 --> 00:11:16.240
And how have you made that oversight workable without losing some of those efficiency gains that you've talked through?
00:11:16.559 --> 00:11:16.960
Yeah.
00:11:18.080 --> 00:11:21.519
Again, Bluecrust, this won't be Bluecrest specific.
00:11:21.759 --> 00:11:27.039
I think any regulated healthcare company um does require accurate note-taking.
00:11:27.120 --> 00:11:36.799
Uh, some of the language that that's used, particularly some of the testing, uh, some of the medical conditions that we that we test for are quite complicated for AI to pick apart upon.
00:11:36.960 --> 00:11:42.559
Um, that didn't detract me from launching, uh going with auto-summary from launch, though.
00:11:42.799 --> 00:11:49.840
The fact that the human is in the loop um right from the very beginning um means that that risk is minimized.
00:11:50.000 --> 00:11:56.320
It's about empowering your agents and being really honest with them and bringing them then on that journey with you.
00:11:56.399 --> 00:12:03.279
So we made it very, very clear to our advisors on the phone that this was an expert experimental technology for us.
00:12:03.440 --> 00:12:17.519
We um we had done everything we could with prompting and uh and testing and running many iterations to make sure that we were getting as accurate auto summaries as possible, and that's been an ongoing process throughout since launch.
00:12:17.600 --> 00:12:26.960
Um, but we wanted to make sure we were taking them on that journey with us and empowering them to give us continuous feedback on the accuracy of that summary.
00:12:27.039 --> 00:12:37.840
So obviously, to start off with that the the um the accuracy wasn't as strong as it is now, and through iteration, through feedback, we've we've improved that and it's gone from strength to strength.
00:12:38.000 --> 00:12:55.440
But having that human in the loop, having that that advisor that either clicks yes to save um or amends it before they do it means that you can take that take that step and get autosummary in place from day one and trust your advisors, they're the experts, they know what notes should be going on to records.
00:12:55.679 --> 00:13:11.840
The time saving, even if it takes them 10, 20, 30 seconds to amend spellings, to correct wording, to remove, I don't know, maybe a vulnerability that's been captured in in inaccurately, um, is still a considerable time saving.
00:13:12.000 --> 00:13:23.919
We went our our average after call work time prior to auto summary was over three and a half minutes, and that primarily was due to the the detail uh of the notes that were required.
00:13:24.080 --> 00:13:31.840
Um prior to um the uh sorry post-launch, um we uh we we went under two and a half minutes.
00:13:31.919 --> 00:13:43.919
We now just sit just above two minutes, so there still is a period of after call work, um, but it's been reduced significantly and resulted in a a considerable resource saving.
00:13:44.159 --> 00:13:53.360
Um we um we didn't reply, we didn't obviously aim to make people redundant and we allowed natural attrition to reduce the headcount.
00:13:53.440 --> 00:14:00.799
Um, but that auto summary has reduced the cost for the contact center considerably um through that that significant time saving.
00:14:01.039 --> 00:14:05.759
The other thing that auto summary really drives is the consistency and format as well.
00:14:06.000 --> 00:14:23.120
Obviously, people write in different ways, um, and um as much as you can guide and train them to follow uh to your your advisors to follow templates and to word things in specific ways, auto summary automatically uh enforces that template and that style of language.
00:14:23.279 --> 00:14:30.639
So um it was it was win real win win really on efficiency and uh improvements in structural accuracy.
00:14:32.320 --> 00:14:40.559
Yeah, I think that's a really kind of great overview of being uh really thorough, really measured when it comes to these uh AI implementations.
00:14:40.639 --> 00:14:44.559
And actually leads me on to something I wanted to speak to you about, Reese.
00:14:44.720 --> 00:14:51.200
You know, we know that the market is is full of these agentic AI solutions and claims and bells and whistles right now.
00:14:51.279 --> 00:15:01.600
But what struck me about Bluecrest's story is I think it it's it's embedded in analytics, in summaries, in getting that clear view of what customers want.
00:15:02.720 --> 00:15:09.440
Not to be rude about it, but it's almost kind of the unglamorous foundation, I guess, for contact centers.
00:15:09.759 --> 00:15:12.799
I was just wondering if you could maybe speak to that.
00:15:12.879 --> 00:15:19.120
Is that an area that you think a lot of other contact centers perhaps overlook when they're looking to implement this sort of technology?
00:15:19.759 --> 00:15:20.159
100%.
00:15:20.559 --> 00:15:31.360
I think you know, looking after the healthcare vertical here at NICE, we have a number of, you know, it's a highly regulated industry, and being able to have a roadmap um and get quick wins to an organization is is a non-negotiable.
00:15:31.679 --> 00:15:42.480
And what's so great about working with customers like Bluecrest and here at Nice is that AI is embedded across all of our platform, not just within conversational AI and other elements.
00:15:42.720 --> 00:15:47.360
You know, there's a lot of efficiencies that can be found within the platform as well.
00:15:47.519 --> 00:15:55.120
And I think what you know is so great with you doing it in a measured fashion as well is the analytics can now quantify those automation opportunities.
00:15:55.200 --> 00:16:12.480
So it's not who shouts the loudest or you know, what we think is uh you know the biggest um return on our investment, but we can quantify that with our conversations and we can see which elements and actions are prime for automation and where we can really um you know get the quick wins for the organization.
00:16:12.720 --> 00:16:31.120
I think that's where uh Bluecrest have done really well and been able to roll that out so that we are ready for that next phase of you know conversation lay and self-service because they already know which elements are good candidates and they've already got the skeletons built from analytics to then go into nice cognity to automate and self-serve as well.
00:16:31.279 --> 00:16:45.519
So I think it's similar to what Russell has said, it's about that optimization and making it a key part of your organization and making the AI to be as accurate as possible and to be able to then piggyback off that um with the roadmap in the future phases as well.
00:16:48.320 --> 00:16:50.639
Yeah, thanks, Reese, and thanks, Russell, as well.
00:16:50.720 --> 00:16:53.039
It was uh it's a really, really great chat.
00:16:53.120 --> 00:17:04.240
I always like these opportunities to speak to people who have kind of that first hand experience and real practical examples of how this technology has been implemented kind of correctly and effectively.
00:17:04.319 --> 00:17:06.640
And I think, yeah, I think the audience will get a great deal from it.
00:17:06.720 --> 00:17:08.799
So thank you very much for your time today.
00:17:09.279 --> 00:17:09.759
Brilliant.
00:17:10.000 --> 00:17:10.559
Thank you so much.
00:17:12.480 --> 00:17:13.039
Great.
00:17:13.200 --> 00:17:16.720
I did also want to just quickly thank our viewers as well for tuning in.
00:17:16.880 --> 00:17:23.440
If you enjoyed this, please do like and subscribe to the channel and head on over to cxtstoday.com for more stories like these.
00:17:23.599 --> 00:17:26.319
Until next time, thanks for watching.
00:00:08.400 --> 00:00:10.640
Hello and welcome to CX Today.
00:00:10.800 --> 00:00:15.679
I'm Reese Fisher, Associate Editor, and today I'm very fortunate to be joined by two people.
00:00:15.759 --> 00:00:22.320
I'm here with Russell White, the transformation lead at Bluecrest, and Rhys Harper, the account executive at Nice.
00:00:22.719 --> 00:00:23.679
Thanks for joining me, guys.
00:00:23.760 --> 00:00:25.199
How are you both doing today?
00:00:26.399 --> 00:00:27.199
Really good, thank you.
00:00:27.359 --> 00:00:28.559
It's a good question to meet you.
00:00:29.679 --> 00:00:30.239
Yeah, great.
00:00:30.320 --> 00:00:33.520
It's great to be on here and uh have a good conversation today.
00:00:34.079 --> 00:00:35.439
Yeah, yeah, absolutely.
00:00:35.520 --> 00:00:51.280
You know, we're gonna be discussing uh a recent collaboration between Bluecrest and NICE, you know, and really I guess using that as a bit of a lens to dig into the relationship between contact centres and customer data, which has been a little fraught at times, I think.
00:00:51.359 --> 00:00:58.159
So we're gonna look at you know where contact centres get it right, where they get it wrong, the role AI plays, the role humans play.
00:00:58.320 --> 00:01:01.280
So yeah, I think it's gonna be a really, really interesting chat.
00:01:01.359 --> 00:01:05.359
And I guess off the back of that, Russell, I wanted to start with you.
00:01:06.159 --> 00:01:19.519
You know, some of the guess specifics of the collab with NICE, one of the big changes was Bluecrest moving from reviewing a sample of calls to analyzing every interaction.
00:01:19.760 --> 00:01:27.439
What did you maybe discover once that you could see that entire picture that you couldn't um that you were maybe missing before?
00:01:29.120 --> 00:01:48.079
Um, so yeah, uh obviously Bluecrest has uh implemented the the nice CX1 system, and as part of that package, we uh purchased interaction analytics, which enables us to uh review every single call, uh categorize those calls, uh, and also use the topic topic AI functionality within the system.
00:01:48.239 --> 00:01:57.519
We're really able to uh um capture information that we weren't able to capture at such a wide scale before, particularly with customer intent.
00:01:57.680 --> 00:02:02.719
So uh our contact center is massively important to us at BlueCrest Wellness.
00:02:02.799 --> 00:02:06.079
The contact centre really is the front door to preventative care.
00:02:06.159 --> 00:02:10.800
Every call we we answer is potentially a health problem found early.
00:02:11.039 --> 00:02:40.240
And knowing why our customers are contacting us and being able to segment that data and analyse that data um opens up a wealth of opportunity for um automation flows and and how we can um uh provide a better service by really splitting those calls out to the to the right people um and having the right automation flows or right processes in place to handle those interactions.
00:02:42.080 --> 00:02:42.879
Yeah, yeah.
00:02:43.039 --> 00:02:45.280
I guess I wanted to come to you next, Rhys.
00:02:45.599 --> 00:02:54.080
Obviously, Russell, you touched on the obvious being able to see all this information is one thing, but then actually acting on that, what you're being able to do off the back of that is obviously the really important next step.
00:02:54.159 --> 00:03:06.400
But you know, I was reading the report around this and I saw that some of the findings were used to improve access, uh reduce after call work, and kind of shape how your advisors handle those customer needs, which you've touched on there, Russell.
00:03:06.639 --> 00:03:16.319
Reese, what do you think separates a contact center that collects more insights from one that actually changes how it operates because of those insights?
00:03:16.719 --> 00:03:17.039
Definitely.
00:03:17.120 --> 00:03:21.439
So I think you obviously here at NICE, we run customer service for the enterprise.
00:03:21.520 --> 00:03:29.039
And what the what's been really great about working with Russell and the Bluecrest team is that the technology is able to deliver that for our customers.
00:03:29.199 --> 00:03:34.560
But it's ultimately about what you do with that information and how you build that into the people and the process.
00:03:34.719 --> 00:03:38.479
And I think you know, since we engaged and worked with Russell, we were always very clear.
00:03:38.639 --> 00:04:08.400
Well, Russell was very clear with us about what we were looking to achieve as part of this project and what you know the end goal was, um, and being able to continually optimize and use that to the technology and the analytics to be able to make informed decisions and further guide that roadmap moving forwards has been something, you know, as you were seeing with some of the um outcomes that Bluecrest has been able to achieve has really been based on that fundamental approach from the beginning to make the technology um work for the organization.
00:04:08.879 --> 00:04:09.199
Yeah.
00:04:09.599 --> 00:04:16.800
I think for me, yeah, our disposition has always told us like the outcome of the call, but not really why people were calling.
00:04:16.879 --> 00:04:25.680
And we guided to the outcome on the call, but having that that campaign attribution built around the the actual intent is very different.
00:04:25.839 --> 00:04:29.360
Um, and it gave us a real clear reporting picture of the customer.
00:04:29.600 --> 00:04:34.079
I mean, the the first thing you find is that the anecdote and reality are two different things.
00:04:34.240 --> 00:04:45.199
What advisors and teammates believe he's had now calls in good faith is not often what the transit trips show, um, which we'll we can talk about a little bit more further further later on.
00:04:45.360 --> 00:04:48.800
But um, and the other thing really is where the time goes.
00:04:48.959 --> 00:05:09.120
If you if you think about at the beginning of of every call, the actual the um the process of of customer identification and idea verification um is really uh you can really drill down onto how long that process is actually taking and and the different ways people handle it and unifying the approach on every call to maximize efficiency.
00:05:09.439 --> 00:05:15.120
Um so yeah, it it really was a big improvement for us having that information available.
00:05:16.959 --> 00:05:18.639
Yeah, yeah, I can imagine.
00:05:18.800 --> 00:05:25.439
You touch on the perhaps the the reality versus maybe some internal assumptions you may have had beforehand.
00:05:25.759 --> 00:05:33.920
Could you give us maybe an example where customer conversations pointed Bluecrest in a different direction from one perhaps that you guys expected?
00:05:35.439 --> 00:06:02.720
Um we weren't aware of really the the volume of calls that that where customers were reaching out with us because of technical difficulties, so it really enabled our uh IT and development teams to uh drill down and and concentrate on those um those um journeys through our website, through our well um our wellness app and identify um common issues that that customers were contacting us for.
00:06:02.959 --> 00:06:16.000
The probably the misconception side of things, um, a really good example of that is that um we uh launched self-serve um for our customers to be able to reschedule their appointments online.
00:06:16.480 --> 00:06:32.240
Uh, and obviously the business uh methodology behind that was it would there would be uh an um process improvement, so it's better for the customers just being able to reschedule online, but also that that would detract uh call volume into the contact center and be an efficiency gain for us.
00:06:32.319 --> 00:06:36.000
So it's win for the customer and win for our efficiency as well, really.
00:06:36.319 --> 00:06:44.879
Um, and so we launched self-serve on on reschedules and anticipated a drop in call volume to the contact center that didn't happen.
00:06:44.959 --> 00:07:00.240
Uh, and interaction analytics and and using AI and those tools really enabled us to um visualize the fact that the the intent, the reason people were uh the rescheduled intent was still just as high as it was prior to that self-surf.
00:07:00.319 --> 00:07:11.759
And in fact, by launching the self-surf, what we'd really enabled our customers was to do was to give them more flexibility in the bit in rescheduling, which just meant we got a higher volume of reschedules.
00:07:11.920 --> 00:07:20.240
So the business misconception was that this would be a huge efficiency gain for us, and actually it wasn't it, but obviously the benefit was still there for the customer.
00:07:21.439 --> 00:07:24.480
Yeah, yeah, that's uh yeah, that's a great example.
00:07:24.720 --> 00:07:31.040
Um, Reese, I wanted to come back to you again, picking out some more of Bluecrest's kind of results from the collab.
00:07:31.279 --> 00:07:38.160
You know, we've seen uh sort of better access, lower after call work, and more relevant customer conversations.
00:07:38.560 --> 00:07:50.560
I guess from your perspective at NICE, when organizations do deploy the interaction analytics that Russell's been talking about, where do they maybe often get the value calculation wrong?
00:07:50.959 --> 00:08:01.920
I think, you know, similar to what Russell said earlier, is not going in there, you know, with a hypothesis or, you know, for what you think is in the data and the analytics give you the answer.
00:08:02.079 --> 00:08:39.919
And I think, you know, what we we see with a lot of our customers is that changes on a week-to-week basis, and being able to have that enterprise level insight and uh changing on a regular basis to be able to guide those business decisions, like Russell mentioned, you know, whether that's a change in something we're offering to a customer or it's a change in the competitive landscape, being able to use that, and I know that Russell does this quite a lot, is feeding that back to the business so that we can use that data to make a not just within the contact center and the CX experience, but there's a lot of wider benefits we see from our customers, not just in operationally, but also wider across their um, you know, across their organization as well.
00:08:40.080 --> 00:08:53.120
And I think that's really where we see a lot of the um unexpected benefits from our customers is when they really truly adopt this, there's a lot more wider benefits than just the operational savings and customer experience as well.
00:08:53.519 --> 00:08:56.320
It's interesting you said that recently as well about the new product.
00:08:56.720 --> 00:09:01.600
I mean, that was another really clear example for BlueCrest on the use of interaction analytics.
00:09:01.759 --> 00:09:06.000
We launched a new product, a product for our customers uh back in April 2025.
00:09:06.080 --> 00:09:11.919
We launched Health Monitor, which is a subscription service that brings customers back every year for an assessment.
00:09:12.000 --> 00:09:17.200
And the initial launch figures um and um conversion rates weren't what we expected.
00:09:17.360 --> 00:09:26.159
Uh, it's a very good product, very strong product, and and it it the importance of regular health monitoring is a really key differentiator for BlueQuest.
00:09:26.240 --> 00:09:29.759
And we weren't seeing the take-up levels that we we anticipated.
00:09:30.080 --> 00:09:35.679
Um, and our advisors were already assuring us that that they were offering it to customers on every call.
00:09:35.840 --> 00:09:44.000
Um, and we were getting a good response online um and a good response from our mail jobs, but not so much from the calls coming into the contact center.
00:09:44.080 --> 00:10:12.480
And interaction analytics enabled us to actually drill down into those calls and actually see that it was only being mentioned on an under 10% of relevant calls, but aged behavior was dictating that they were sticking to their RAL scripts, and and you get that with with uh sort of our advisors where that they're used to selling a product or used to discussing products in particular ways, um, and that enabled us to target training at those particular agents that weren't mentioning this new product, and we saw considerable increase of that.
00:10:12.639 --> 00:10:18.080
The same week we we actually fed that back to the to the team leaders and the contact center merger.
00:10:18.159 --> 00:10:23.440
Uh, it jumped up to 40%, it's now closer to the 60-70% mark.
00:10:23.519 --> 00:10:32.320
Uh, and we have um uh nearly 4,000 health monitor subscriptions now in place, and that insight wouldn't have been available before without interaction analytics.
00:10:32.639 --> 00:10:40.879
We would have we were only sampling a very small number of calls, and the cool and the calls that we were sampling weren't highlighting the scope of the problem in the contact centre.
00:10:42.720 --> 00:10:44.879
Yeah, yeah, it's another really incredible result.
00:10:45.120 --> 00:10:57.200
I find it I think one of the parts of the the collab particularly interesting is obviously you guys are are based in the healthcare sector, and we know how vitally important accuracy is in a sector like that.
00:10:57.279 --> 00:11:03.200
You know, summary can become part of a customer record, so getting it wrong carries that real consequence.
00:11:03.919 --> 00:11:09.360
I was just wondering why do you think human review was kind of a non-negotiable for you guys?
00:11:09.440 --> 00:11:16.240
And how have you made that oversight workable without losing some of those efficiency gains that you've talked through?
00:11:16.559 --> 00:11:16.960
Yeah.
00:11:18.080 --> 00:11:21.519
Again, Bluecrust, this won't be Bluecrest specific.
00:11:21.759 --> 00:11:27.039
I think any regulated healthcare company um does require accurate note-taking.
00:11:27.120 --> 00:11:36.799
Uh, some of the language that that's used, particularly some of the testing, uh, some of the medical conditions that we that we test for are quite complicated for AI to pick apart upon.
00:11:36.960 --> 00:11:42.559
Um, that didn't detract me from launching, uh going with auto-summary from launch, though.
00:11:42.799 --> 00:11:49.840
The fact that the human is in the loop um right from the very beginning um means that that risk is minimized.
00:11:50.000 --> 00:11:56.320
It's about empowering your agents and being really honest with them and bringing them then on that journey with you.
00:11:56.399 --> 00:12:03.279
So we made it very, very clear to our advisors on the phone that this was an expert experimental technology for us.
00:12:03.440 --> 00:12:17.519
We um we had done everything we could with prompting and uh and testing and running many iterations to make sure that we were getting as accurate auto summaries as possible, and that's been an ongoing process throughout since launch.
00:12:17.600 --> 00:12:26.960
Um, but we wanted to make sure we were taking them on that journey with us and empowering them to give us continuous feedback on the accuracy of that summary.
00:12:27.039 --> 00:12:37.840
So obviously, to start off with that the the um the accuracy wasn't as strong as it is now, and through iteration, through feedback, we've we've improved that and it's gone from strength to strength.
00:12:38.000 --> 00:12:55.440
But having that human in the loop, having that that advisor that either clicks yes to save um or amends it before they do it means that you can take that take that step and get autosummary in place from day one and trust your advisors, they're the experts, they know what notes should be going on to records.
00:12:55.679 --> 00:13:11.840
The time saving, even if it takes them 10, 20, 30 seconds to amend spellings, to correct wording, to remove, I don't know, maybe a vulnerability that's been captured in in inaccurately, um, is still a considerable time saving.
00:13:12.000 --> 00:13:23.919
We went our our average after call work time prior to auto summary was over three and a half minutes, and that primarily was due to the the detail uh of the notes that were required.
00:13:24.080 --> 00:13:31.840
Um prior to um the uh sorry post-launch, um we uh we we went under two and a half minutes.
00:13:31.919 --> 00:13:43.919
We now just sit just above two minutes, so there still is a period of after call work, um, but it's been reduced significantly and resulted in a a considerable resource saving.
00:13:44.159 --> 00:13:53.360
Um we um we didn't reply, we didn't obviously aim to make people redundant and we allowed natural attrition to reduce the headcount.
00:13:53.440 --> 00:14:00.799
Um, but that auto summary has reduced the cost for the contact center considerably um through that that significant time saving.
00:14:01.039 --> 00:14:05.759
The other thing that auto summary really drives is the consistency and format as well.
00:14:06.000 --> 00:14:23.120
Obviously, people write in different ways, um, and um as much as you can guide and train them to follow uh to your your advisors to follow templates and to word things in specific ways, auto summary automatically uh enforces that template and that style of language.
00:14:23.279 --> 00:14:30.639
So um it was it was win real win win really on efficiency and uh improvements in structural accuracy.
00:14:32.320 --> 00:14:40.559
Yeah, I think that's a really kind of great overview of being uh really thorough, really measured when it comes to these uh AI implementations.
00:14:40.639 --> 00:14:44.559
And actually leads me on to something I wanted to speak to you about, Reese.
00:14:44.720 --> 00:14:51.200
You know, we know that the market is is full of these agentic AI solutions and claims and bells and whistles right now.
00:14:51.279 --> 00:15:01.600
But what struck me about Bluecrest's story is I think it it's it's embedded in analytics, in summaries, in getting that clear view of what customers want.
00:15:02.720 --> 00:15:09.440
Not to be rude about it, but it's almost kind of the unglamorous foundation, I guess, for contact centers.
00:15:09.759 --> 00:15:12.799
I was just wondering if you could maybe speak to that.
00:15:12.879 --> 00:15:19.120
Is that an area that you think a lot of other contact centers perhaps overlook when they're looking to implement this sort of technology?
00:15:19.759 --> 00:15:20.159
100%.
00:15:20.559 --> 00:15:31.360
I think you know, looking after the healthcare vertical here at NICE, we have a number of, you know, it's a highly regulated industry, and being able to have a roadmap um and get quick wins to an organization is is a non-negotiable.
00:15:31.679 --> 00:15:42.480
And what's so great about working with customers like Bluecrest and here at Nice is that AI is embedded across all of our platform, not just within conversational AI and other elements.
00:15:42.720 --> 00:15:47.360
You know, there's a lot of efficiencies that can be found within the platform as well.
00:15:47.519 --> 00:15:55.120
And I think what you know is so great with you doing it in a measured fashion as well is the analytics can now quantify those automation opportunities.
00:15:55.200 --> 00:16:12.480
So it's not who shouts the loudest or you know, what we think is uh you know the biggest um return on our investment, but we can quantify that with our conversations and we can see which elements and actions are prime for automation and where we can really um you know get the quick wins for the organization.
00:16:12.720 --> 00:16:31.120
I think that's where uh Bluecrest have done really well and been able to roll that out so that we are ready for that next phase of you know conversation lay and self-service because they already know which elements are good candidates and they've already got the skeletons built from analytics to then go into nice cognity to automate and self-serve as well.
00:16:31.279 --> 00:16:45.519
So I think it's similar to what Russell has said, it's about that optimization and making it a key part of your organization and making the AI to be as accurate as possible and to be able to then piggyback off that um with the roadmap in the future phases as well.
00:16:48.320 --> 00:16:50.639
Yeah, thanks, Reese, and thanks, Russell, as well.
00:16:50.720 --> 00:16:53.039
It was uh it's a really, really great chat.
00:16:53.120 --> 00:17:04.240
I always like these opportunities to speak to people who have kind of that first hand experience and real practical examples of how this technology has been implemented kind of correctly and effectively.
00:17:04.319 --> 00:17:06.640
And I think, yeah, I think the audience will get a great deal from it.
00:17:06.720 --> 00:17:08.799
So thank you very much for your time today.
00:17:09.279 --> 00:17:09.759
Brilliant.
00:17:10.000 --> 00:17:10.559
Thank you so much.
00:17:12.480 --> 00:17:13.039
Great.
00:17:13.200 --> 00:17:16.720
I did also want to just quickly thank our viewers as well for tuning in.
00:17:16.880 --> 00:17:23.440
If you enjoyed this, please do like and subscribe to the channel and head on over to cxtstoday.com for more stories like these.
00:17:23.599 --> 00:17:26.319
Until next time, thanks for watching.
