WEBVTT
00:00:09.039 --> 00:00:10.800
Hello and welcome to CX Today.
00:00:11.119 --> 00:00:17.039
One of the most significant shifts happening in customer experience technology right now is the ongoing shake-up in the CCAS market.
00:00:17.199 --> 00:00:27.519
You know, vendor consolidation, rapid AI innovation, changing customer expectations, and increasing pressure to modernize are all creating new challenges and opportunities for enterprise buyers.
00:00:27.760 --> 00:00:34.000
So to help us make sense of what this means in practice, I'm joined by Martin Taylor, co-founder and deputy CEO of Content Giro.
00:00:34.399 --> 00:00:35.679
Thank you for joining us, Martin.
00:00:36.159 --> 00:00:37.280
Pleasure to be here.
00:00:37.840 --> 00:00:39.359
So let's start with the big picture.
00:00:39.439 --> 00:00:42.399
You know, the CCAS market seems to be changing faster than ever.
00:00:42.799 --> 00:00:44.079
What's happening right now?
00:00:44.159 --> 00:00:47.200
You know, and why should buyers really be playing paying close attention?
00:00:47.679 --> 00:00:52.719
I suppose we're now in about stage three, I suppose, of the age of AI.
00:00:53.200 --> 00:01:01.359
Um and obviously as as CCAS uh vendors and people in the the CX space, we're right at the forefront of all of that.
00:01:01.679 --> 00:01:06.879
Because as we know, CX is the ideal place to introduce AI and scale it.
00:01:07.519 --> 00:01:12.959
So we've really got uh two major aspects that that the CCAS players are looking at.
00:01:13.120 --> 00:01:16.239
One is augmenting the human agent.
00:01:16.480 --> 00:01:34.079
So that's things like translating, transcribing, summarizing, uh, and then a lot of those uh, I suppose, offline features such as automated quality management, uh further automation around the workforce management, and so on, to increase efficiency.
00:01:34.239 --> 00:02:04.000
Um of course the other side is increasing the scope of automation, uh, and that's what's leading us from the kind of stage two gen AI into the stage three uh agentic AI, where you can have much more of an extensive dialogue to get to uh the bottom of a query and and hopefully resolve it all within the automated environment, but also the collecting of data for the human.
00:02:04.239 --> 00:02:08.400
So it that's really the the background at the moment.
00:02:08.560 --> 00:02:15.199
It's this uh transition, I suppose, from uh the augmenting into the the automating.
00:02:15.599 --> 00:02:24.159
Of course, why automation is important is that the human aspect to the contact centre is well over 90% of the cost.
00:02:24.400 --> 00:02:24.719
Sure.
00:02:24.960 --> 00:02:36.879
So you know Gartner, their stat is that I think the contact centre uh industry is worth$420 billion, which that's a fantastically big number.
00:02:37.360 --> 00:02:43.280
Uh obviously, not if you're Elon Musk, but for everyone else it's a big market.
00:02:43.599 --> 00:02:49.280
But then if we consider what's the the tech component of that market, it's a much smaller number.
00:02:49.840 --> 00:03:06.719
Uh and uh a stat I like from just a long time ago now, 2024 from Opus, uh, was that the tech component is only 2.4% of spend, 97.6% uh was in the the human component.
00:03:07.840 --> 00:03:14.400
But that by 2032, 58% of that human component would have been automated.
00:03:15.199 --> 00:03:24.159
So that I suppose increase in the in the total share that goes to to tech is great, obviously, if you're in in tech.
00:03:24.479 --> 00:03:29.280
And actually by last year, that uh 2.4% had become 4.6%.
00:03:30.000 --> 00:03:31.680
So we can see it's increasing.
00:03:32.400 --> 00:03:56.000
Uh and when I was at the NVIDIA GTC conference in San Jose back in March, uh Jensen Huang's keynote uh pretty much opened with the the customer uh service market, as he called it, uh will shortly be worth$35 billion a year, which is a significant uplift, around a quadrupling of the market size.
00:03:56.479 --> 00:04:01.840
So it's both uh a great opportunity, uh, but also a threat.
00:04:01.919 --> 00:04:11.840
If you're not keeping up, then that share of the uplift, which would be the the majority of the value, uh, would be taken by new entrants.
00:04:12.080 --> 00:04:20.879
So, of course, in CCAS uh we've got years now of understanding the business processes that we seek to automate.
00:04:21.040 --> 00:04:27.199
And of course, yeah, we we've been in this automation phase for for many years before AI came along.
00:04:27.759 --> 00:04:35.360
We like to look at someone like UK Power Networks, where we automate 94% of all of the inquiries completely.
00:04:35.680 --> 00:04:38.480
Um that's really pre-AI.
00:04:39.199 --> 00:04:45.279
So CCAS are good at doing this, uh, and you know, we're we're gonna be doing more and more of it.
00:04:45.439 --> 00:04:47.839
So that's really the the scene as we see it.
00:04:48.160 --> 00:04:48.560
Sure.
00:04:48.879 --> 00:04:54.240
And then within that context, you know, you've got obviously a lot of changes happening and consolidation as well.
00:04:54.399 --> 00:05:01.519
And you know, for for customers, it kind of might feel like moving platforms is this major effort that they're not quite sure how to tackle.
00:05:01.680 --> 00:05:11.199
So, you know, if an enterprise is is running on you know a legacy contact centre platform, um, it what is the the cost, as it were, for them to stay where they are?
00:05:11.600 --> 00:05:17.519
Yeah, I suppose it's like that old analogy of the frog in a slowly heating pan of water.
00:05:17.600 --> 00:05:21.439
Uh and I I've never known anyone actually test the theory.
00:05:21.600 --> 00:05:29.439
Does the frog boil or yeah, though supposedly if it it moves suddenly, it will jump out, otherwise it will slowly boil to death.
00:05:29.680 --> 00:05:34.879
So the the temperature in the pan is is getting uncomfortably warm now.
00:05:35.279 --> 00:05:49.920
Uh, and you know, people are worrying not just that they might not be keeping up, because obviously, let's remember, customer experience is the single biggest differentiator and source of competitive advantage for organizations.
00:05:50.240 --> 00:06:02.879
You've got this increasingly fickle consumer uh who will, yeah, we're told, often move off after years of brand loyalty, they'll be gone after one bad experience.
00:06:03.279 --> 00:06:10.399
So their expectations are set by the best consumer experience that they've ever come across.
00:06:10.480 --> 00:06:18.399
Uh so whoever you are, uh whatever type of organization, that's the consumer expectation that they're coming to the party with.
00:06:19.120 --> 00:06:34.800
At the same time, whilst uh perhaps most organizations have migrated their contact centers of the cloud, the big ones, uh large enterprises, governments, they're the ones who tend not to have made that migration.
00:06:35.040 --> 00:06:46.000
So actually, another Gartner stat is that 70% of contact center workers are working in a legacy on-premises environment today, uh, which might surprise many.
00:06:46.160 --> 00:06:53.439
So that's actually the work environment of the majority of people who work in these contact centers.
00:06:53.680 --> 00:07:02.399
So clearly that's because there are a lot of processes that would need to be changed if you are to migrate.
00:07:02.800 --> 00:07:07.680
So I think you've got a lot of people who are afraid of making that leap.
00:07:08.000 --> 00:07:14.160
In the early days of cloud, it was quite easy to say, well, uh, this cloud isn't safe.
00:07:14.319 --> 00:07:15.120
It's the internet.
00:07:15.439 --> 00:07:22.160
Uh, I can't move my data there, it's sensitive, we've got regulatory concerns, there's safety to worry about.
00:07:22.319 --> 00:07:27.680
Uh, and you could get away with that as an IT manager for for some years.
00:07:28.079 --> 00:07:32.240
But uh cloud has long been enterprise grade now.
00:07:32.560 --> 00:07:36.079
So I think those arguments are have gone away.
00:07:36.240 --> 00:07:43.839
Uh, but you know, let's face it, you've got some complex migrations, and we we do a lot of those at Content Guru.
00:07:44.240 --> 00:07:54.480
So there was a big one we did last year, and uh it was a large national service, uh critical infrastructure, et cetera.
00:07:54.720 --> 00:08:00.560
Um, and it was also the amalgamation of eight organizations into one.
00:08:00.959 --> 00:08:09.759
So uh that that some MA, as it were, had taken place over some years, but you know, we needed to consolidate all that infrastructure.
00:08:10.000 --> 00:08:18.480
And we needed to do a big data migration and data cleanse to a new customer data platform all at once.
00:08:18.800 --> 00:08:26.879
So that's that's a typical scenario for a large organization with embedded legacy infrastructure.
00:08:27.040 --> 00:08:35.200
Uh, but they want to not just offer a good service, they want to attract and retain the best people to deliver that service.
00:08:35.360 --> 00:08:39.279
Uh and yeah, they want to be working in a modern environment too.
00:08:39.440 --> 00:08:39.759
Yeah.
00:08:40.080 --> 00:08:44.240
So there are many, many good reasons to be doing it.
00:08:44.399 --> 00:08:49.519
Not least because when the CEO says, where's your your AI then?
00:08:49.759 --> 00:09:05.679
Uh, if you're not already modernized and in the cloud and you've got this omni-data and omni-channel environment already in place, you're simply not going to be in a position to deliver this promised land of AI, the the C level.
00:09:05.840 --> 00:09:12.240
And actually, beyond them, the board are demanding that you deliver and all the efficiency gains that they're wanting to see.
00:09:12.399 --> 00:09:14.799
Uh never mind replacing the people.
00:09:15.360 --> 00:09:22.320
Um how about just making them, say, one-third more efficient in an already heavily optimized work environment?
00:09:22.559 --> 00:09:25.679
And a lot of the time that's what our customers are asking for.
00:09:25.840 --> 00:09:28.080
We're saying, look, we've got increasing demand.
00:09:28.159 --> 00:09:32.480
Uh, we don't want to just uh people our way out of it.
00:09:32.720 --> 00:09:35.600
Uh we need to be more efficient and effective.
00:09:35.919 --> 00:09:38.639
And we want to bring together the silos.
00:09:38.879 --> 00:09:41.360
So it's we're not just talking about voice calls, of course.
00:09:41.519 --> 00:09:48.559
Many people think omni-channel's a done deal, but yeah, the majority of large organizations are still only multi-channel.
00:09:48.720 --> 00:09:55.519
So things like the social, the email, uh, will be and the chat will be happening on different platforms.
00:09:55.679 --> 00:09:58.159
So you've got different experiences going on.
00:09:58.399 --> 00:10:01.519
You're not amalgamating that data effectively.
00:10:01.600 --> 00:10:11.600
Uh, and yeah, so these are all challenges that the the the the the people who want these things to take place uh are uh having to consider.
00:10:12.320 --> 00:10:13.120
Yeah, exactly.
00:10:13.200 --> 00:10:26.080
And you mentioned that that pressure um, you know, to to be seen to be doing something with AI, um, but also, you know, in the context of the current Care Shakeup, you know, what does that mean for organizations that want to adopt a genetic AI?
00:10:26.639 --> 00:10:32.159
Yes, well, agentic, I suppose, uh learning the lessons of generative.
00:10:32.639 --> 00:10:38.559
So uh many people plowed into generative with great gusto.
00:10:38.720 --> 00:10:39.840
Some went too early.
00:10:39.919 --> 00:10:51.679
We all remember Klarna uh from 23, I think it was, where they said, right, we're gonna get rid of everybody, uh, we're doing this deal with OpenAI, and it's all gonna be agentic.
00:10:52.159 --> 00:10:58.720
And then you know, hurriedly they're having to hire everyone back because you know the customers didn't like it, it didn't work properly.
00:10:58.879 --> 00:11:05.200
So some have gone too soon uh in the agentic, uh in the regenerative, rather.
00:11:05.360 --> 00:11:15.519
Uh so IBM uh put out a stat that only 25% of Gen AI projects had generated ROI.
00:11:16.080 --> 00:11:22.799
Uh and actually that that goes with a lot of experiences that that we've seen around the industry.
00:11:23.120 --> 00:11:37.279
MIT put out an even more pessimistic uh note back in uh August of 2025 that 95% of Gen AI projects were failing because they weren't delivering ROI.
00:11:37.679 --> 00:11:43.039
So, in terms of a genetic, you've got to establish what is it we're actually trying to achieve here?
00:11:43.200 --> 00:11:45.120
What are the objectives?
00:11:45.360 --> 00:11:53.600
So, you know, a lot of the time people in the previous round of AI just plowed on because you know we've been told to do AI.
00:11:54.080 --> 00:11:55.840
Now it's a more thoughtful approach.
00:11:56.000 --> 00:11:58.559
So what are we actually trying to achieve here?
00:11:58.799 --> 00:12:10.320
Uh and also baselining the existing service so that the ROI as a delta between the baseline and the new way of doing things can be demonstrated.
00:12:10.559 --> 00:12:20.639
One of our customers is a city council, for example, and we did our AI uh generative demonstration.
00:12:20.799 --> 00:12:22.480
I loved it, it was great.
00:12:22.639 --> 00:12:32.960
Um we talked about where to introduce it, and initially they said we'll do it in this new service, uh, which is waste food waste collection.
00:12:33.120 --> 00:12:46.080
Uh and whilst that was great, because it was a new service, we argued that it wouldn't be possible to definitively prove that the AI way was better than doing it any other way.
00:12:46.240 --> 00:12:51.840
So in the end we applied it to the switchboard because that was well baselined.
00:12:52.159 --> 00:12:59.600
So we knew how many calls, how long, where they went, cost per call, cost per route, and so on.
00:12:59.840 --> 00:13:12.399
So when you then apply the regenerative version, and a lot of it is then contained uh and resolved, yeah, it's it's very straightforward to make that business case.
00:13:14.399 --> 00:13:14.879
Sure.
00:13:15.200 --> 00:13:26.240
And um, you know, when buyers are looking at all kind of all the options laid before them, you know, there's a lot of competing claims from vendors, you know, it can be like a complex process to choose a replacement platform.
00:13:26.320 --> 00:13:32.639
So, you know, when they begin this evaluation process, what would you suggest would be kind of the non-negotiable things they need to consider?
00:13:33.679 --> 00:13:39.360
I would look at the vendor's experience, I think, first.
00:13:39.679 --> 00:13:54.000
Uh so everyone can do a demo, uh, and it's not that difficult to put the the customer's website in as the source of information, and it looks like a great, it could pretty much do everything out of the box.
00:13:54.480 --> 00:13:58.159
The reality is is a lot more complex than that.
00:13:58.399 --> 00:14:16.240
So I would be looking at do does this vendor really understand what it means to operate the business processes that I'm seeking to automate, uh, or is it we've just come to it as a kind of idea and we're doing some prompt engineering.
00:14:16.480 --> 00:14:22.480
So experience, uh security, data sovereignty, that's a big factor now.
00:14:22.879 --> 00:14:34.080
People want to know where is my data and my customers' data being processed, uh by whom, what kind of uh jurisdiction uh is it subject to?
00:14:34.240 --> 00:14:42.720
Could it be seized and taken overseas uh according to some primary legislation affecting the vendor?
00:14:43.039 --> 00:14:45.840
So these are items to consider.
00:14:46.399 --> 00:14:54.000
Um also understanding uh the processes that you're looking to replace.
00:14:54.159 --> 00:14:58.720
So people always underestimate the the cost of how they do it today.
00:14:59.039 --> 00:15:03.679
And then, of course, you can't have AI without good data.
00:15:03.919 --> 00:15:09.600
So I always talk of the the cycle of data discovery, validation, and deployment.
00:15:09.840 --> 00:15:16.240
So they need to look at the automation in the frame of data first.
00:15:16.720 --> 00:15:21.759
Is my data in a state uh to do this agentic automation?
00:15:22.240 --> 00:15:29.440
Um you might well need a customer data platform in place to orchestrate your many systems of record.
00:15:29.679 --> 00:15:36.960
I think the days where everyone said, well, we're gonna get a CRM and everything is gonna be on the CRM, it's just not realistic now.
00:15:37.120 --> 00:15:42.559
Uh most organizations have multiple core systems of record, often multiple CRMs.
00:15:43.200 --> 00:15:57.360
So if you gather and orchestrate that data within a customer data platform, that is the environment where your agentic AI will succeed and get its information and deposit its information.
00:15:57.519 --> 00:16:08.480
And then the existing systems of record that your workers in your many different departments use day to day, those will be automatically and securely updated in the background.
00:16:08.960 --> 00:16:11.039
So look beyond the demo, really.
00:16:11.120 --> 00:16:17.759
Had they done this sort of thing before in our sector, what's their record of complex migrations?
00:16:18.080 --> 00:16:25.120
Because you don't want to be taking reckless experimental risks with your customers.
00:16:25.279 --> 00:16:28.720
Because as I just explained, we're a fickle bunch nowadays.
00:16:29.200 --> 00:16:30.720
Yeah, exactly.
00:16:30.879 --> 00:16:40.559
And you touched on orchestration, um, but particularly you know, moving into a gentic and having multiple AI agents, you know, doing multiple things, sometimes overlapping, working together.
00:16:40.720 --> 00:16:44.240
So, how does AI orchestration factor into a migration decision?
00:16:44.639 --> 00:16:47.120
Yeah, it's a bit like data orchestration.
00:16:47.360 --> 00:17:05.519
So whilst there's no single CRM to rule the world, uh similarly uh in terms of AI, different AI vendors with LLMs, small language models and things are continually leapfrogging one another, pulling ahead.
00:17:05.599 --> 00:17:10.960
It's a very fast-moving space with new entrants coming along at regular intervals.
00:17:11.680 --> 00:17:32.880
So therefore, I think the the job of your provider of the communication side of things is to overlay all of the uh leading model makers so that you've got access to the best models, the most appropriate models, and also the horizon scanning is taking place for you.
00:17:33.039 --> 00:17:45.759
Uh and yeah, we often find with our customers that we're replacing models, we're replacing engines in the background the whole time as the new uh and better products come along.
00:17:46.319 --> 00:17:55.359
So don't nail your colours to one mast because it may be today's leader, but it probably won't be tomorrow's leader.
00:17:55.680 --> 00:17:57.440
That's what I would advise there.
00:17:57.759 --> 00:17:59.039
Yeah, absolutely.
00:17:59.279 --> 00:18:08.640
Um so you know, we've covered a lot of ground here, but if you could leave enterprise buyers with one practical piece of advice, you know, as they navigate this period of change in the CCAS market, what would it be?
00:18:09.119 --> 00:18:16.160
I think it would be think you know widely about what it is you want to achieve.
00:18:16.559 --> 00:18:19.119
So it's a C-level decision, really.
00:18:19.359 --> 00:18:24.720
Often you'll find the requirements are being set too low down in the organization.
00:18:25.119 --> 00:18:26.799
The ambitions aren't there.
00:18:26.880 --> 00:18:31.920
They're just wanting to replace an existing process and modernize it a little bit.
00:18:32.559 --> 00:18:33.359
Think big.
00:18:33.599 --> 00:18:36.799
So think of the objectives you're looking to achieve.
00:18:37.119 --> 00:18:43.039
Think of baselining your existing services uh and understanding them properly.
00:18:43.279 --> 00:18:48.240
Um and just don't be swayed by a single flashy demo.
00:18:48.400 --> 00:18:54.799
Uh think through do these people understand uh my business processes, my sector?
00:18:55.039 --> 00:18:56.319
Do I have experience?
00:18:56.559 --> 00:19:00.480
Do I have the security accreditations that I'll be looking for?
00:19:00.559 --> 00:19:05.839
Um and then you know, proceed with caution and excitement at the same time.
00:19:06.400 --> 00:19:07.440
Yeah, absolutely.
00:19:07.599 --> 00:19:11.119
I think that that's a great place to leave it for our viewers to think about.
00:19:11.279 --> 00:19:16.960
So thank you, Martin, for sharing your insights and you know, helping us to understand what this market transition means.
00:19:17.519 --> 00:19:18.720
Absolute pleasure.
00:19:19.119 --> 00:19:20.720
And thanks to everyone for watching.
00:19:20.799 --> 00:19:27.920
Um, if you'd like to learn more about how organizations can approach CCAS modernization and AR adoption, visit contentguru.com.
00:19:28.079 --> 00:19:33.440
And for more videos, interviews, and news articles, head on over to cxtoday.com.
00:19:33.519 --> 00:19:35.759
And also you can find us on LinkedIn.
00:19:36.079 --> 00:19:38.880
So thanks for watching, and we'll see you next time.
00:00:09.039 --> 00:00:10.800
Hello and welcome to CX Today.
00:00:11.119 --> 00:00:17.039
One of the most significant shifts happening in customer experience technology right now is the ongoing shake-up in the CCAS market.
00:00:17.199 --> 00:00:27.519
You know, vendor consolidation, rapid AI innovation, changing customer expectations, and increasing pressure to modernize are all creating new challenges and opportunities for enterprise buyers.
00:00:27.760 --> 00:00:34.000
So to help us make sense of what this means in practice, I'm joined by Martin Taylor, co-founder and deputy CEO of Content Giro.
00:00:34.399 --> 00:00:35.679
Thank you for joining us, Martin.
00:00:36.159 --> 00:00:37.280
Pleasure to be here.
00:00:37.840 --> 00:00:39.359
So let's start with the big picture.
00:00:39.439 --> 00:00:42.399
You know, the CCAS market seems to be changing faster than ever.
00:00:42.799 --> 00:00:44.079
What's happening right now?
00:00:44.159 --> 00:00:47.200
You know, and why should buyers really be playing paying close attention?
00:00:47.679 --> 00:00:52.719
I suppose we're now in about stage three, I suppose, of the age of AI.
00:00:53.200 --> 00:01:01.359
Um and obviously as as CCAS uh vendors and people in the the CX space, we're right at the forefront of all of that.
00:01:01.679 --> 00:01:06.879
Because as we know, CX is the ideal place to introduce AI and scale it.
00:01:07.519 --> 00:01:12.959
So we've really got uh two major aspects that that the CCAS players are looking at.
00:01:13.120 --> 00:01:16.239
One is augmenting the human agent.
00:01:16.480 --> 00:01:34.079
So that's things like translating, transcribing, summarizing, uh, and then a lot of those uh, I suppose, offline features such as automated quality management, uh further automation around the workforce management, and so on, to increase efficiency.
00:01:34.239 --> 00:02:04.000
Um of course the other side is increasing the scope of automation, uh, and that's what's leading us from the kind of stage two gen AI into the stage three uh agentic AI, where you can have much more of an extensive dialogue to get to uh the bottom of a query and and hopefully resolve it all within the automated environment, but also the collecting of data for the human.
00:02:04.239 --> 00:02:08.400
So it that's really the the background at the moment.
00:02:08.560 --> 00:02:15.199
It's this uh transition, I suppose, from uh the augmenting into the the automating.
00:02:15.599 --> 00:02:24.159
Of course, why automation is important is that the human aspect to the contact centre is well over 90% of the cost.
00:02:24.400 --> 00:02:24.719
Sure.
00:02:24.960 --> 00:02:36.879
So you know Gartner, their stat is that I think the contact centre uh industry is worth$420 billion, which that's a fantastically big number.
00:02:37.360 --> 00:02:43.280
Uh obviously, not if you're Elon Musk, but for everyone else it's a big market.
00:02:43.599 --> 00:02:49.280
But then if we consider what's the the tech component of that market, it's a much smaller number.
00:02:49.840 --> 00:03:06.719
Uh and uh a stat I like from just a long time ago now, 2024 from Opus, uh, was that the tech component is only 2.4% of spend, 97.6% uh was in the the human component.
00:03:07.840 --> 00:03:14.400
But that by 2032, 58% of that human component would have been automated.
00:03:15.199 --> 00:03:24.159
So that I suppose increase in the in the total share that goes to to tech is great, obviously, if you're in in tech.
00:03:24.479 --> 00:03:29.280
And actually by last year, that uh 2.4% had become 4.6%.
00:03:30.000 --> 00:03:31.680
So we can see it's increasing.
00:03:32.400 --> 00:03:56.000
Uh and when I was at the NVIDIA GTC conference in San Jose back in March, uh Jensen Huang's keynote uh pretty much opened with the the customer uh service market, as he called it, uh will shortly be worth$35 billion a year, which is a significant uplift, around a quadrupling of the market size.
00:03:56.479 --> 00:04:01.840
So it's both uh a great opportunity, uh, but also a threat.
00:04:01.919 --> 00:04:11.840
If you're not keeping up, then that share of the uplift, which would be the the majority of the value, uh, would be taken by new entrants.
00:04:12.080 --> 00:04:20.879
So, of course, in CCAS uh we've got years now of understanding the business processes that we seek to automate.
00:04:21.040 --> 00:04:27.199
And of course, yeah, we we've been in this automation phase for for many years before AI came along.
00:04:27.759 --> 00:04:35.360
We like to look at someone like UK Power Networks, where we automate 94% of all of the inquiries completely.
00:04:35.680 --> 00:04:38.480
Um that's really pre-AI.
00:04:39.199 --> 00:04:45.279
So CCAS are good at doing this, uh, and you know, we're we're gonna be doing more and more of it.
00:04:45.439 --> 00:04:47.839
So that's really the the scene as we see it.
00:04:48.160 --> 00:04:48.560
Sure.
00:04:48.879 --> 00:04:54.240
And then within that context, you know, you've got obviously a lot of changes happening and consolidation as well.
00:04:54.399 --> 00:05:01.519
And you know, for for customers, it kind of might feel like moving platforms is this major effort that they're not quite sure how to tackle.
00:05:01.680 --> 00:05:11.199
So, you know, if an enterprise is is running on you know a legacy contact centre platform, um, it what is the the cost, as it were, for them to stay where they are?
00:05:11.600 --> 00:05:17.519
Yeah, I suppose it's like that old analogy of the frog in a slowly heating pan of water.
00:05:17.600 --> 00:05:21.439
Uh and I I've never known anyone actually test the theory.
00:05:21.600 --> 00:05:29.439
Does the frog boil or yeah, though supposedly if it it moves suddenly, it will jump out, otherwise it will slowly boil to death.
00:05:29.680 --> 00:05:34.879
So the the temperature in the pan is is getting uncomfortably warm now.
00:05:35.279 --> 00:05:49.920
Uh, and you know, people are worrying not just that they might not be keeping up, because obviously, let's remember, customer experience is the single biggest differentiator and source of competitive advantage for organizations.
00:05:50.240 --> 00:06:02.879
You've got this increasingly fickle consumer uh who will, yeah, we're told, often move off after years of brand loyalty, they'll be gone after one bad experience.
00:06:03.279 --> 00:06:10.399
So their expectations are set by the best consumer experience that they've ever come across.
00:06:10.480 --> 00:06:18.399
Uh so whoever you are, uh whatever type of organization, that's the consumer expectation that they're coming to the party with.
00:06:19.120 --> 00:06:34.800
At the same time, whilst uh perhaps most organizations have migrated their contact centers of the cloud, the big ones, uh large enterprises, governments, they're the ones who tend not to have made that migration.
00:06:35.040 --> 00:06:46.000
So actually, another Gartner stat is that 70% of contact center workers are working in a legacy on-premises environment today, uh, which might surprise many.
00:06:46.160 --> 00:06:53.439
So that's actually the work environment of the majority of people who work in these contact centers.
00:06:53.680 --> 00:07:02.399
So clearly that's because there are a lot of processes that would need to be changed if you are to migrate.
00:07:02.800 --> 00:07:07.680
So I think you've got a lot of people who are afraid of making that leap.
00:07:08.000 --> 00:07:14.160
In the early days of cloud, it was quite easy to say, well, uh, this cloud isn't safe.
00:07:14.319 --> 00:07:15.120
It's the internet.
00:07:15.439 --> 00:07:22.160
Uh, I can't move my data there, it's sensitive, we've got regulatory concerns, there's safety to worry about.
00:07:22.319 --> 00:07:27.680
Uh, and you could get away with that as an IT manager for for some years.
00:07:28.079 --> 00:07:32.240
But uh cloud has long been enterprise grade now.
00:07:32.560 --> 00:07:36.079
So I think those arguments are have gone away.
00:07:36.240 --> 00:07:43.839
Uh, but you know, let's face it, you've got some complex migrations, and we we do a lot of those at Content Guru.
00:07:44.240 --> 00:07:54.480
So there was a big one we did last year, and uh it was a large national service, uh critical infrastructure, et cetera.
00:07:54.720 --> 00:08:00.560
Um, and it was also the amalgamation of eight organizations into one.
00:08:00.959 --> 00:08:09.759
So uh that that some MA, as it were, had taken place over some years, but you know, we needed to consolidate all that infrastructure.
00:08:10.000 --> 00:08:18.480
And we needed to do a big data migration and data cleanse to a new customer data platform all at once.
00:08:18.800 --> 00:08:26.879
So that's that's a typical scenario for a large organization with embedded legacy infrastructure.
00:08:27.040 --> 00:08:35.200
Uh, but they want to not just offer a good service, they want to attract and retain the best people to deliver that service.
00:08:35.360 --> 00:08:39.279
Uh and yeah, they want to be working in a modern environment too.
00:08:39.440 --> 00:08:39.759
Yeah.
00:08:40.080 --> 00:08:44.240
So there are many, many good reasons to be doing it.
00:08:44.399 --> 00:08:49.519
Not least because when the CEO says, where's your your AI then?
00:08:49.759 --> 00:09:05.679
Uh, if you're not already modernized and in the cloud and you've got this omni-data and omni-channel environment already in place, you're simply not going to be in a position to deliver this promised land of AI, the the C level.
00:09:05.840 --> 00:09:12.240
And actually, beyond them, the board are demanding that you deliver and all the efficiency gains that they're wanting to see.
00:09:12.399 --> 00:09:14.799
Uh never mind replacing the people.
00:09:15.360 --> 00:09:22.320
Um how about just making them, say, one-third more efficient in an already heavily optimized work environment?
00:09:22.559 --> 00:09:25.679
And a lot of the time that's what our customers are asking for.
00:09:25.840 --> 00:09:28.080
We're saying, look, we've got increasing demand.
00:09:28.159 --> 00:09:32.480
Uh, we don't want to just uh people our way out of it.
00:09:32.720 --> 00:09:35.600
Uh we need to be more efficient and effective.
00:09:35.919 --> 00:09:38.639
And we want to bring together the silos.
00:09:38.879 --> 00:09:41.360
So it's we're not just talking about voice calls, of course.
00:09:41.519 --> 00:09:48.559
Many people think omni-channel's a done deal, but yeah, the majority of large organizations are still only multi-channel.
00:09:48.720 --> 00:09:55.519
So things like the social, the email, uh, will be and the chat will be happening on different platforms.
00:09:55.679 --> 00:09:58.159
So you've got different experiences going on.
00:09:58.399 --> 00:10:01.519
You're not amalgamating that data effectively.
00:10:01.600 --> 00:10:11.600
Uh, and yeah, so these are all challenges that the the the the the people who want these things to take place uh are uh having to consider.
00:10:12.320 --> 00:10:13.120
Yeah, exactly.
00:10:13.200 --> 00:10:26.080
And you mentioned that that pressure um, you know, to to be seen to be doing something with AI, um, but also, you know, in the context of the current Care Shakeup, you know, what does that mean for organizations that want to adopt a genetic AI?
00:10:26.639 --> 00:10:32.159
Yes, well, agentic, I suppose, uh learning the lessons of generative.
00:10:32.639 --> 00:10:38.559
So uh many people plowed into generative with great gusto.
00:10:38.720 --> 00:10:39.840
Some went too early.
00:10:39.919 --> 00:10:51.679
We all remember Klarna uh from 23, I think it was, where they said, right, we're gonna get rid of everybody, uh, we're doing this deal with OpenAI, and it's all gonna be agentic.
00:10:52.159 --> 00:10:58.720
And then you know, hurriedly they're having to hire everyone back because you know the customers didn't like it, it didn't work properly.
00:10:58.879 --> 00:11:05.200
So some have gone too soon uh in the agentic, uh in the regenerative, rather.
00:11:05.360 --> 00:11:15.519
Uh so IBM uh put out a stat that only 25% of Gen AI projects had generated ROI.
00:11:16.080 --> 00:11:22.799
Uh and actually that that goes with a lot of experiences that that we've seen around the industry.
00:11:23.120 --> 00:11:37.279
MIT put out an even more pessimistic uh note back in uh August of 2025 that 95% of Gen AI projects were failing because they weren't delivering ROI.
00:11:37.679 --> 00:11:43.039
So, in terms of a genetic, you've got to establish what is it we're actually trying to achieve here?
00:11:43.200 --> 00:11:45.120
What are the objectives?
00:11:45.360 --> 00:11:53.600
So, you know, a lot of the time people in the previous round of AI just plowed on because you know we've been told to do AI.
00:11:54.080 --> 00:11:55.840
Now it's a more thoughtful approach.
00:11:56.000 --> 00:11:58.559
So what are we actually trying to achieve here?
00:11:58.799 --> 00:12:10.320
Uh and also baselining the existing service so that the ROI as a delta between the baseline and the new way of doing things can be demonstrated.
00:12:10.559 --> 00:12:20.639
One of our customers is a city council, for example, and we did our AI uh generative demonstration.
00:12:20.799 --> 00:12:22.480
I loved it, it was great.
00:12:22.639 --> 00:12:32.960
Um we talked about where to introduce it, and initially they said we'll do it in this new service, uh, which is waste food waste collection.
00:12:33.120 --> 00:12:46.080
Uh and whilst that was great, because it was a new service, we argued that it wouldn't be possible to definitively prove that the AI way was better than doing it any other way.
00:12:46.240 --> 00:12:51.840
So in the end we applied it to the switchboard because that was well baselined.
00:12:52.159 --> 00:12:59.600
So we knew how many calls, how long, where they went, cost per call, cost per route, and so on.
00:12:59.840 --> 00:13:12.399
So when you then apply the regenerative version, and a lot of it is then contained uh and resolved, yeah, it's it's very straightforward to make that business case.
00:13:14.399 --> 00:13:14.879
Sure.
00:13:15.200 --> 00:13:26.240
And um, you know, when buyers are looking at all kind of all the options laid before them, you know, there's a lot of competing claims from vendors, you know, it can be like a complex process to choose a replacement platform.
00:13:26.320 --> 00:13:32.639
So, you know, when they begin this evaluation process, what would you suggest would be kind of the non-negotiable things they need to consider?
00:13:33.679 --> 00:13:39.360
I would look at the vendor's experience, I think, first.
00:13:39.679 --> 00:13:54.000
Uh so everyone can do a demo, uh, and it's not that difficult to put the the customer's website in as the source of information, and it looks like a great, it could pretty much do everything out of the box.
00:13:54.480 --> 00:13:58.159
The reality is is a lot more complex than that.
00:13:58.399 --> 00:14:16.240
So I would be looking at do does this vendor really understand what it means to operate the business processes that I'm seeking to automate, uh, or is it we've just come to it as a kind of idea and we're doing some prompt engineering.
00:14:16.480 --> 00:14:22.480
So experience, uh security, data sovereignty, that's a big factor now.
00:14:22.879 --> 00:14:34.080
People want to know where is my data and my customers' data being processed, uh by whom, what kind of uh jurisdiction uh is it subject to?
00:14:34.240 --> 00:14:42.720
Could it be seized and taken overseas uh according to some primary legislation affecting the vendor?
00:14:43.039 --> 00:14:45.840
So these are items to consider.
00:14:46.399 --> 00:14:54.000
Um also understanding uh the processes that you're looking to replace.
00:14:54.159 --> 00:14:58.720
So people always underestimate the the cost of how they do it today.
00:14:59.039 --> 00:15:03.679
And then, of course, you can't have AI without good data.
00:15:03.919 --> 00:15:09.600
So I always talk of the the cycle of data discovery, validation, and deployment.
00:15:09.840 --> 00:15:16.240
So they need to look at the automation in the frame of data first.
00:15:16.720 --> 00:15:21.759
Is my data in a state uh to do this agentic automation?
00:15:22.240 --> 00:15:29.440
Um you might well need a customer data platform in place to orchestrate your many systems of record.
00:15:29.679 --> 00:15:36.960
I think the days where everyone said, well, we're gonna get a CRM and everything is gonna be on the CRM, it's just not realistic now.
00:15:37.120 --> 00:15:42.559
Uh most organizations have multiple core systems of record, often multiple CRMs.
00:15:43.200 --> 00:15:57.360
So if you gather and orchestrate that data within a customer data platform, that is the environment where your agentic AI will succeed and get its information and deposit its information.
00:15:57.519 --> 00:16:08.480
And then the existing systems of record that your workers in your many different departments use day to day, those will be automatically and securely updated in the background.
00:16:08.960 --> 00:16:11.039
So look beyond the demo, really.
00:16:11.120 --> 00:16:17.759
Had they done this sort of thing before in our sector, what's their record of complex migrations?
00:16:18.080 --> 00:16:25.120
Because you don't want to be taking reckless experimental risks with your customers.
00:16:25.279 --> 00:16:28.720
Because as I just explained, we're a fickle bunch nowadays.
00:16:29.200 --> 00:16:30.720
Yeah, exactly.
00:16:30.879 --> 00:16:40.559
And you touched on orchestration, um, but particularly you know, moving into a gentic and having multiple AI agents, you know, doing multiple things, sometimes overlapping, working together.
00:16:40.720 --> 00:16:44.240
So, how does AI orchestration factor into a migration decision?
00:16:44.639 --> 00:16:47.120
Yeah, it's a bit like data orchestration.
00:16:47.360 --> 00:17:05.519
So whilst there's no single CRM to rule the world, uh similarly uh in terms of AI, different AI vendors with LLMs, small language models and things are continually leapfrogging one another, pulling ahead.
00:17:05.599 --> 00:17:10.960
It's a very fast-moving space with new entrants coming along at regular intervals.
00:17:11.680 --> 00:17:32.880
So therefore, I think the the job of your provider of the communication side of things is to overlay all of the uh leading model makers so that you've got access to the best models, the most appropriate models, and also the horizon scanning is taking place for you.
00:17:33.039 --> 00:17:45.759
Uh and yeah, we often find with our customers that we're replacing models, we're replacing engines in the background the whole time as the new uh and better products come along.
00:17:46.319 --> 00:17:55.359
So don't nail your colours to one mast because it may be today's leader, but it probably won't be tomorrow's leader.
00:17:55.680 --> 00:17:57.440
That's what I would advise there.
00:17:57.759 --> 00:17:59.039
Yeah, absolutely.
00:17:59.279 --> 00:18:08.640
Um so you know, we've covered a lot of ground here, but if you could leave enterprise buyers with one practical piece of advice, you know, as they navigate this period of change in the CCAS market, what would it be?
00:18:09.119 --> 00:18:16.160
I think it would be think you know widely about what it is you want to achieve.
00:18:16.559 --> 00:18:19.119
So it's a C-level decision, really.
00:18:19.359 --> 00:18:24.720
Often you'll find the requirements are being set too low down in the organization.
00:18:25.119 --> 00:18:26.799
The ambitions aren't there.
00:18:26.880 --> 00:18:31.920
They're just wanting to replace an existing process and modernize it a little bit.
00:18:32.559 --> 00:18:33.359
Think big.
00:18:33.599 --> 00:18:36.799
So think of the objectives you're looking to achieve.
00:18:37.119 --> 00:18:43.039
Think of baselining your existing services uh and understanding them properly.
00:18:43.279 --> 00:18:48.240
Um and just don't be swayed by a single flashy demo.
00:18:48.400 --> 00:18:54.799
Uh think through do these people understand uh my business processes, my sector?
00:18:55.039 --> 00:18:56.319
Do I have experience?
00:18:56.559 --> 00:19:00.480
Do I have the security accreditations that I'll be looking for?
00:19:00.559 --> 00:19:05.839
Um and then you know, proceed with caution and excitement at the same time.
00:19:06.400 --> 00:19:07.440
Yeah, absolutely.
00:19:07.599 --> 00:19:11.119
I think that that's a great place to leave it for our viewers to think about.
00:19:11.279 --> 00:19:16.960
So thank you, Martin, for sharing your insights and you know, helping us to understand what this market transition means.
00:19:17.519 --> 00:19:18.720
Absolute pleasure.
00:19:19.119 --> 00:19:20.720
And thanks to everyone for watching.
00:19:20.799 --> 00:19:27.920
Um, if you'd like to learn more about how organizations can approach CCAS modernization and AR adoption, visit contentguru.com.
00:19:28.079 --> 00:19:33.440
And for more videos, interviews, and news articles, head on over to cxtoday.com.
00:19:33.519 --> 00:19:35.759
And also you can find us on LinkedIn.
00:19:36.079 --> 00:19:38.880
So thanks for watching, and we'll see you next time.
