37: Amazon Connect Customer EMEA Community Event

WEBVTT
00:00:02.577 --> 00:00:05.467
Welcome to ACP, the Amazon Connect podcast.
00:00:06.668 --> 00:00:10.378
This is the show that focuses on Amazon Connect and related technologies.
00:00:11.028 --> 00:00:18.557
I'm your host, Tom Morgan, and I'm joined as usual by my co-host, AWS solution architect and contact center consultant, Alex Baker.
00:00:19.457 --> 00:00:24.007
Find out more about Cloud Interact by visiting us at cloudinteract.io.
00:00:34.539 --> 00:00:41.259
It's time for another ACP, and a special one this time as you find us on the 6:00 AM train to London.
00:00:41.609 --> 00:00:42.548
I'm here with Alex.
00:00:42.569 --> 00:00:43.988
Alex is commuting with me.
00:00:44.048 --> 00:00:44.679
Good morning.
00:00:45.039 --> 00:00:45.609
Good morning.
00:00:45.668 --> 00:00:46.069
Yes.
00:00:46.109 --> 00:00:47.798
Yeah, nice early start for us this morning.
00:00:48.209 --> 00:00:55.348
The reason we're both sitting on a train together is that we are headed to the Amazon Connect Customer Community event.
00:00:55.628 --> 00:00:57.698
Now, it's a two-day event in London.
00:00:57.988 --> 00:00:59.639
Alex was there yesterday.
00:00:59.709 --> 00:01:01.439
I am joining him for today.
00:01:01.939 --> 00:01:16.528
Alex, nine years ago when Amazon Connect first started and you first got involved with it, did you ever imagine that you would be sitting on a train heading to a Amazon Connect customer, regional event, one of many being held around the world?
00:01:17.388 --> 00:01:18.899
No No, I guess not.
00:01:18.938 --> 00:01:25.118
And yeah, thinking back to then, it was definitely something new and exciting, but really in its infancy.
00:01:25.409 --> 00:01:33.099
And I guess it's impressive the changes that have been made over those nine years and the sort of constant velocity of change as well is really impressive.
00:01:33.519 --> 00:01:47.578
A- and I've mentioned it quite a few times before, but the fact that there was that sort of game-changing moment when you realized that you could just go and set up a contact center yourself and go and experiment with it I think that always made it stand out for me.
00:01:47.638 --> 00:01:49.909
But yes, not necessarily.
00:01:50.009 --> 00:01:53.778
But yeah, pleased to be here, and it's been an interesting first day at the event yesterday.
00:01:54.379 --> 00:01:55.319
Yeah, absolutely.
00:01:55.358 --> 00:02:01.099
And so yesterday was all technical day, and today is more business track focused.
00:02:01.418 --> 00:02:07.379
So I guess, I know you didn't get to go to a lot of the sessions'cause we had a stand as well and you were on the stand.
00:02:07.688 --> 00:02:08.658
But what was the...
00:02:08.709 --> 00:02:10.079
the general kind of feeling?
00:02:10.088 --> 00:02:13.508
What was the atmosphere like during that technical day and your thoughts
00:02:13.588 --> 00:02:14.949
yeah, great atmosphere.
00:02:14.998 --> 00:02:25.239
It, it was, I think because it was the technical track day, there was a focus o- on allowing the more technically minded people to go and find out more about the product and do some learning.
00:02:25.248 --> 00:02:34.389
So there were a couple of really good workshop sessions around setting up Agentic AI and also around automating evaluations and quality monitoring.
00:02:34.679 --> 00:02:38.278
But yes, all in all, it was a really good, well-run day.
00:02:38.639 --> 00:02:46.618
A packed agenda, and it looks like today's even more it's it's starting at 9:00, and it runs all the way through, I think, till 6:00 with session after session.
00:02:46.628 --> 00:02:50.449
So hopefully we will get a chance to duck in and see a few of those today.
00:02:50.808 --> 00:02:51.278
Excellent.
00:02:51.278 --> 00:02:52.899
Definitely getting our money's worth.
00:02:53.199 --> 00:02:53.658
All right.
00:02:53.658 --> 00:02:56.818
So we're gonna have a couple more coffees, I think, to get us ready.
00:02:57.028 --> 00:03:01.109
And then we're just gonna, we're gonna be there throughout the day, try and give you a bit of a sense of the day.
00:03:04.907 --> 00:03:07.426
So it's a couple of days on from the event, and we're back.
00:03:07.467 --> 00:03:12.016
We've had a bit of time to think about it and a bit of distance to reflect on the event.
00:03:12.396 --> 00:03:22.437
So this episode is going to be for you if you didn't attend, I think, and you want to really understand, what happened there and what mattered and why it's important.
00:03:22.467 --> 00:03:24.497
Let's start, I think, with biggest takeaways.
00:03:24.516 --> 00:03:31.417
So for me, I think the most interesting thing of the whole event was that, gosh, there was a lot of AI.
00:03:31.646 --> 00:03:39.796
If you came to this hoping to have a conversation and hear some sessions about call flows and channels and queues, you would-- a bit disappointed, I think.
00:03:40.116 --> 00:03:49.887
This was all about AI and Amazon Connect, and that was probably the biggest, If we d- if we needed it, the biggest reality check, I think, of where Amazon Connect is really focusing its efforts right now.
00:03:50.477 --> 00:03:50.947
Yeah.
00:03:51.067 --> 00:03:52.276
It makes sense, doesn't it?
00:03:52.326 --> 00:04:01.317
In terms of the whole, the re-- almost rebranding of Amazon Connect as the agentic AI service within AWS.
00:04:01.776 --> 00:04:02.586
But yeah, agree.
00:04:02.616 --> 00:04:08.526
I guess I probably wasn't thinking I'd go there and not hear anything about AI, but yes, it was completely front and center, wasn't it?
00:04:08.866 --> 00:04:09.417
Unsurprisingly.
00:04:10.028 --> 00:04:10.429
Yes.
00:04:10.479 --> 00:04:10.989
Yeah, absolutely.
00:04:11.068 --> 00:04:11.808
Yeah, and you're right.
00:04:11.808 --> 00:04:19.028
Yeah, of course there was gonna be AI, but yeah, I was surprised that it was purely AI, and that seems to be the only thing that matters right now.
00:04:19.288 --> 00:04:21.249
And I know that's not actually true.
00:04:21.288 --> 00:04:25.449
If you look at engineering effort and stuff, all of the other stuff is important and keeps evolving.
00:04:25.749 --> 00:04:32.048
But in terms of what AWS are talking about right now, it's all AI all the time, and not just AWS, all the partners as well.
00:04:32.588 --> 00:04:32.889
Yes.
00:04:32.899 --> 00:04:34.689
What was your biggest takeaway from-
00:04:35.583 --> 00:04:37.293
AI themed as well.
00:04:37.312 --> 00:04:46.673
Again, perhaps unsurprisingly, but there was a lot around the sort of the increasing move towards AI agents acting as colleagues or coworkers.
00:04:47.052 --> 00:05:01.033
So there, there was quite a good bit in one of the keynotes around looking at it as an activity where I do it versus let's do it together versus you do it for me with your AI agent or agents.
00:05:01.452 --> 00:05:02.773
Yeah, possibly even agents.
00:05:02.783 --> 00:05:09.583
So a single agent may be orchestrating multiple other agents to go and do things for you.
00:05:09.583 --> 00:05:17.132
But it's moving away from maybe a year and a bit ago, it was the sort of the fairly-- the rigid almost chatbot format, wasn't it?
00:05:17.132 --> 00:05:19.132
Y- your ChatGPT type thing.
00:05:19.502 --> 00:05:28.812
Whereas now it's becoming much more of a sophisticated sort of hookup of different agents doing different things and acting as a almost a co-collaborator with you.
00:05:29.656 --> 00:05:30.416
Yeah, definitely.
00:05:30.476 --> 00:05:31.627
Yeah, that was really interesting.
00:05:38.810 --> 00:05:39.740
We've just arrived.
00:05:39.771 --> 00:05:47.190
We arrived a little bit early to set up our stand,'cause we are one of the exhibitors here today, and I'm just unpacking a few bits and pieces.
00:05:47.190 --> 00:05:50.141
But the thing that really struck me is two things really.
00:05:50.500 --> 00:05:56.930
The first this is held at the Ministry of Sound, so it's everything you think it's gonna be It's like being in a nightclub, and there's some...
00:05:56.930 --> 00:05:57.600
There's a DJ.
00:05:57.641 --> 00:05:58.240
There's music.
00:05:58.240 --> 00:05:59.701
It's relatively loud.
00:05:59.980 --> 00:06:01.850
There's so high energy, which is really nice.
00:06:01.911 --> 00:06:04.310
There's a massive glitter ball above my head.
00:06:04.591 --> 00:06:06.620
There is cool lighting everywhere.
00:06:06.701 --> 00:06:08.670
There are there's some really cool stuff on the wall.
00:06:09.110 --> 00:06:14.581
But there's also breakfast pastries and coffee and yogurts and granola and all those sort of things that you'd have at a conference.
00:06:14.581 --> 00:06:17.110
So it's this really interesting juxtaposition.
00:06:17.110 --> 00:06:24.120
I really like whoever organizes certainly the UK Amazon Connect events that we've been to, because they all have this kind of similar vibe.
00:06:24.151 --> 00:06:27.021
They're super high energy, modern slightly edgy.
00:06:27.050 --> 00:06:32.281
But it makes for a great atmosphere I think, and much more interesting than just a bland conference venue.
00:06:33.281 --> 00:06:40.750
The other interesting thing that I've noticed is how many AWS people are here, and just listening to conversations and also people I know, there's people from the UK.
00:06:40.771 --> 00:06:42.120
There's people from the US as well.
00:06:42.120 --> 00:06:43.930
There's a whole bunch of US people that have flown over.
00:06:44.221 --> 00:06:51.641
I think that's a really strong signal as well about this event, and so really looking forward to catching up with some of them as well as the day progresses.
00:06:55.488 --> 00:06:59.538
As I said, lots of AWS people came in from the US.
00:06:59.879 --> 00:07:01.788
Why do you think that matters?
00:07:01.788 --> 00:07:04.788
Like, why do you think that's important, an event like this?
00:07:06.603 --> 00:07:15.202
It just shows the massive focus that Amazon Con-Connect is getting from the wider AWS organization at the moment.
00:07:15.742 --> 00:07:22.083
It's-- I think they've said a few times over the last couple of years it's one of the biggest growing services.
00:07:22.512 --> 00:07:33.353
We had that headline again within probably the, or certainly the last year where it exceeded the billion-dollar a year run rate or billion-dollar run rate anyway.
00:07:34.052 --> 00:07:38.793
But yeah I think it's, it just shows that it's really important to, to AWS, I think.
00:07:39.000 --> 00:07:39.670
Yeah, definitely.
00:07:39.690 --> 00:07:49.880
I think there's definitely a thing you can use to find out or to work out, and it's not rocket science, but work out how big organizations probably tech org-- well, definitely tech organizations.
00:07:49.880 --> 00:07:54.560
This is probably applicable to other organizations as well, but I've noticed it with tech organizations.
00:07:54.870 --> 00:08:06.920
You can really tell where their focus is commercially by how much budget they get for things like events, for swag, for branded, whatever it is and for events and all that kind of stuff.
00:08:06.980 --> 00:08:10.100
And that's really obvious, and it's really obvious when it's cut as well.
00:08:10.430 --> 00:08:12.651
You can really tell what's important and what's not.
00:08:13.021 --> 00:08:17.690
And yeah, there was-- there's a lot of Amazon Connect people, there was a lot of Amazon Connect swag.
00:08:18.071 --> 00:08:20.471
Amazon Connect is having its time at the moment, I think.
00:08:20.540 --> 00:08:26.591
And that's-- I think that's-- that sends some signals about, where Amazon Connect sits inside AWS right now
00:08:27.338 --> 00:08:35.379
I guess it, it's probably a shrewd move if you think back historically, how often people do change their contact center stack.
00:08:35.908 --> 00:08:42.749
It's a pretty involved, pr-pretty big decision to make, and it's not like people tend to do it every year or two.
00:08:42.749 --> 00:08:48.278
It tends to be maybe five, seven, 10, 10-year change cycles.
00:08:48.708 --> 00:09:01.619
So them investing a lot upfront and trying to get as many people onto the platform is probably going to s- set them in good stead for the, yeah the next, five to 10 years in terms of ongoing revenue with some pretty big enterprises.
00:09:02.133 --> 00:09:02.822
Yeah, definitely.
00:09:04.072 --> 00:09:09.972
Let's talk a bit through the sort of technical day which was the day before the one w-we were both at.
00:09:10.023 --> 00:09:11.403
I know you were there, but I wasn't there.
00:09:11.682 --> 00:09:14.143
So it was a bit more technical, and there were some workshops as well.
00:09:14.442 --> 00:09:19.523
An agentic AI setup one and a automatic evaluation, like a QM-based one.
00:09:19.932 --> 00:09:25.202
I know you managed to get your head into both of them, not completely, but you had other things to do.
00:09:25.202 --> 00:09:29.673
But from what you saw, what was being taught in those sessions, in those workshops?
00:09:29.773 --> 00:09:31.822
What, if people were going there, what were they getting out of it?
00:09:32.736 --> 00:09:33.366
Yeah.
00:09:33.417 --> 00:09:52.437
And one thing, and I have-- I've mentioned the workshops before on the podcast, and I am a big fan of them, and they always come across as pretty slick when they're being run in a big event like this,'cause you've got I don't know how many people were, but, maybe 100 people, say, in the evaluation forms one.
00:09:53.006 --> 00:09:55.197
It's, it all seems to run pretty well.
00:09:55.197 --> 00:09:58.596
They set up environments for everyone, which are just throwaway environments.
00:09:58.596 --> 00:10:03.767
You get all the collateral so that you can run the workshop on the day, but you can also...
00:10:04.017 --> 00:10:11.037
you basically, you'll end up going home with a CloudFormation template to rebuild whatever you've built in the workshop.
00:10:11.466 --> 00:10:16.567
And it just, it always works, and it's well presented by the presenters.
00:10:16.986 --> 00:10:22.937
They take time to have somebody going around and checking people are okay with it and that they're not having issues.
00:10:23.287 --> 00:10:34.836
So yeah I think it was, it's a really good, it's a really good opportunity to go and just learn and immerse yourself in the whatever element of the product it was for a couple of hours.
00:10:35.336 --> 00:10:40.917
I think it's interesting that it was those two sessions that they chose to run.
00:10:41.177 --> 00:10:45.366
Thinking about all the things they could have, the workshops they could have chosen to run, they chose those two.
00:10:45.746 --> 00:11:06.586
I think it makes maybe, AWS think that there's a friction right now for technical people in standing agents up because they're relatively new, so getting people upskilled in how to do that, and then also knowing whether they're any good once they're live, doing that evaluation part, which is it's slightly unglamorous, honestly, compared to, setting up the evaluations, but it's so important, and we know that from our work as well.
00:11:06.917 --> 00:11:20.187
And not only is it important because it's important to get things right and know when your AI agent is not getting things right, but it's a huge confidence builder as well, and I think that's holding, that's holding a lot of people back, is the fear of what if it goes wrong?
00:11:20.187 --> 00:11:22.326
What if it, tells my customers bad things?
00:11:22.726 --> 00:11:24.397
That's where evaluations can really help.
00:11:24.397 --> 00:11:28.456
So it's interesting to look at, the choices they made in setting up those workshops
00:11:29.182 --> 00:11:30.533
Yeah, it's a good point, actually.
00:11:30.582 --> 00:11:39.123
And I suppose one of my takeaways was that and like you say I hold my hands up and say I did only poke my head in the door of the AI agent one.
00:11:39.643 --> 00:11:46.432
I just thought it was important to mention that, yes, it is possible to set up your own agent in a short time, like in the workshop.
00:11:46.972 --> 00:11:53.293
And yes, of course, it will be extremely capable, but do definitely bear in mind that it isn't just a one-and-done thing.
00:11:53.623 --> 00:12:11.393
We're finding it so important to continuously monitor and track what the agent is doing and monitor your statistics a-a-around how it's, tool use, for example, how it's dealing with the queries, what types of queries is it not able to deal with.
00:12:11.462 --> 00:12:18.903
And that is more of an ongoing iterative process rather than a, I'll just set it up and and, let it crack on really.
00:12:19.592 --> 00:12:20.113
Yeah.
00:12:20.113 --> 00:12:27.763
And that, that can be a bit of a problem with some of those workshops sometimes, is that the, in practice it's a little bit harder than they make it sound.
00:12:27.832 --> 00:12:32.283
And not always technically, but for some of those reasons that you say, is that y- you're not one and done.
00:12:32.633 --> 00:12:38.232
There is a whole bunch of things to do, but it when you're in the workshop in the moment it makes it seem very simple.
00:12:38.232 --> 00:12:40.452
You've just stepped through it, you've done it inside of an hour.
00:12:41.293 --> 00:12:44.373
Of course, you could just go back and do this and have something working inside of an hour.
00:12:44.413 --> 00:12:52.253
So it's, that's one of those areas where the workshops I think are really helpful, but you do need to position them against reality sometimes, and I think that's important
00:12:53.471 --> 00:12:53.780
Yeah.
00:12:53.780 --> 00:12:54.191
Agreed.
00:12:54.250 --> 00:12:54.421
Yeah.
00:12:54.530 --> 00:13:05.760
A-and like I say, I'm big fan of them, and I do think they and those ones in particular were run in a very slick way, and you do get something pretty cool and tangible to take away from it.
00:13:05.760 --> 00:13:07.990
But yeah, j-I just thought it was important to mention that.
00:13:08.041 --> 00:13:18.520
It's like it, it's and things like the business change aspect of it yes, it's great to go and do the technical setup, but there are a whole load of other things that are important, is what I'm saying.
00:13:19.144 --> 00:13:23.333
Yeah, and that's related to something else we heard during the conference actually, which we'll play now.
00:13:30.091 --> 00:13:32.441
We're well into the event now mid-morning.
00:13:32.451 --> 00:13:34.591
We've been through the keynote and a couple of sessions.
00:13:34.650 --> 00:13:44.341
Alex, I thought the keynote was really interesting for kind of talking more generally about how we work with agents and with AI and not everything is fully autonomous.
00:13:44.341 --> 00:13:50.051
It's more about working out where AI makes the most sense and how to evolve your organizations to deal with it.
00:13:50.051 --> 00:13:51.091
I thought that was interesting.
00:13:52.311 --> 00:13:52.890
Yeah.
00:13:52.900 --> 00:13:54.821
And the- there was some interesting takeaways, weren't there?
00:13:54.980 --> 00:14:01.740
For instance, the prediction o- of AI agents being more sort of part of your org chart going forward.
00:14:02.100 --> 00:14:08.500
And the another interesting one was that you may have an AI boss sooner than you have an AI colleague I thought was quite interesting.
00:14:08.500 --> 00:14:08.990
Yeah.
00:14:08.990 --> 00:14:11.750
That, that definitely there was a ripple of something around the room there.
00:14:11.801 --> 00:14:22.470
I'm not sure if it was nervous laughter or what, so we had lots of conversations at the demo stand, some really good conversations with people.
00:14:22.700 --> 00:14:25.120
But we did manage to get some sessions as well, as you heard.
00:14:25.490 --> 00:14:31.090
And I wanna get your take on this idea around, we could have an AI boss before we have an AI colleague.
00:14:31.740 --> 00:14:39.899
Is that realistic or is it something you say at a conference to get everybody, make sure everyone's listening and, get some social media interest?
00:14:40.967 --> 00:14:47.798
Yeah, th-th-that was, and I think that it came up again in the keynote from Matias Anduraga.
00:14:48.038 --> 00:14:50.677
Apologies Matias if I've pronounced your name incorrectly.
00:14:50.677 --> 00:14:52.947
But yeah he had some great points, didn't he?
00:14:52.998 --> 00:15:04.807
That around having an AI boss before an AI colleague, it sounds from what I'm hearing,on places like LinkedIn that yes that's definitely possible and definitely happening.
00:15:05.278 --> 00:15:27.607
I can't remember which organiz-ization it was, but they categorized people into sort of builders and so measurers was another category, and measurers was, I guess your sort of broadly middle management, and it was that area that was being cut back in and you can imagine, yes, probably being replaced with some degree of AI.
00:15:27.607 --> 00:15:32.238
So yeah, it sounds like it's definitely possible and most likely happening in places.
00:15:33.076 --> 00:15:33.365
Yeah.
00:15:33.416 --> 00:15:49.855
It's fascinating really, and it's-- I can't work out if it's deliberately provocative or whether it's a reaction to, the kind of top-down, you hear it from less educated or just less aware, CEOs and high-level managers are like,"Oh, we're gonna replace all coding jobs," and stuff like that.
00:15:50.056 --> 00:15:56.255
Whether it's a kind of reaction to that, it's actually, we're probably gonna replace all the management jobs first if that's all right because it turns out that might be easier.
00:15:56.615 --> 00:15:57.686
It's fascinating really.
00:15:57.686 --> 00:16:02.815
I'm reminded actually of somebody-- something somebody said to me like very early in my career.
00:16:03.086 --> 00:16:14.323
and I was a very junior dev and I was grumbling about a development manager because I didn't think he was very good And a more senior developer pulled me aside and was like,"He's probably paid less than you are." I was like,"Really?
00:16:14.384 --> 00:16:18.328
But he's the development manager." He's"Yeah, but he's just managing people." Just managing people is a whole job, right?
00:16:18.328 --> 00:16:21.639
It's not-- But he's not super technical, and he was quite technical, but he's not a developer.
00:16:22.349 --> 00:16:23.099
And so there was that...
00:16:23.149 --> 00:16:28.938
For me, that was a real like light bulb moment as a, inexperienced developer railing against the world.
00:16:29.109 --> 00:16:39.318
You don't think like that, but actually, yes, it's a different, it's a, just because you're managing a bunch of people you're not necessarily the best version of the best of all those people, it's a different discipline.
00:16:39.548 --> 00:16:40.349
Yeah, really interesting.
00:16:40.777 --> 00:16:52.876
A-and we've I'm probably just thinking about what order our podcasts are coming up in and don't wanna d- any sort of hit any spoiler alerts, but we have had discussions around Amazon Connect Talent recently.
00:16:53.376 --> 00:17:06.767
And the alm- almost like a glimpse into the sort of dystopian future of recruitment was quite an eye-opener there, where you can potentially just a-automate the whole recruitment process.
00:17:06.767 --> 00:17:11.356
It, yeah you can definitely see that the whole AI boss thing is not a stretch of the
00:17:11.759 --> 00:17:22.578
No, and you've got to imagine Mateus is coming from that place of having had those conversations for six months and having spoken to enterprises about this for three months probably to get soundings on the ground.
00:17:22.578 --> 00:17:26.798
So actually to bring it out in a public place at a conference, yeah, you're right.
00:17:26.798 --> 00:17:27.969
Maybe it's sooner than we think.
00:17:28.538 --> 00:17:29.439
Yeah, really interesting.
00:17:29.608 --> 00:17:36.669
I came into this discussion thinking I'm not sure I'd buy that, but actually now that we've now that we've chatted it through, I'm like actually maybe I do.
00:17:37.028 --> 00:17:37.949
But it's, yeah.
00:17:38.169 --> 00:17:38.459
Okay.
00:17:39.009 --> 00:17:39.459
Interesting.
00:17:39.576 --> 00:17:53.856
other one I thought, a-again from Matthias's his keynote but mentioned in, in a couple of other places, a couple of other presentations, but the fact that your future customer might be a piece of code as well, I thought that was quite interesting.
00:17:53.856 --> 00:17:59.346
So not only is your boss potentially going to be AI, but also your customer is going to be AI.
00:17:59.747 --> 00:18:10.616
And it might look at the resources that your company has available, and it might make a pretty snap judgment on whether or not it will be your customer, and it will be that straightforward and simple.
00:18:10.866 --> 00:18:12.796
Yeah, machine customers was the term.
00:18:12.923 --> 00:18:17.084
there was a speaker towards the end of the day called Clare Muscutt from Women in CX.
00:18:17.453 --> 00:18:19.443
Really good session, but introduced that term.
00:18:19.443 --> 00:18:20.384
I'd not heard that before.
00:18:20.384 --> 00:18:26.844
But it's, the idea that, yes, you will have-- on behalf of your u- your users, they will be using their own AI.
00:18:26.864 --> 00:18:30.923
So actually o-one of your customers might be an AI on behalf of a user.
00:18:30.973 --> 00:18:36.513
So this idea of machine customers and making the experience good for them because they will choose...
00:18:36.703 --> 00:18:38.413
they, they don't have loyalty particularly.
00:18:38.644 --> 00:18:40.864
They will choose the most efficient mechanism.
00:18:41.064 --> 00:18:49.094
So if you've got a terrible onboarding experience or whatever it is that's hostile to, to agents, then that's not gonna go down very well.
00:18:50.036 --> 00:18:50.316
Yeah.
00:18:51.215 --> 00:18:52.215
Yeah, it's really interesting.
00:18:53.105 --> 00:19:19.346
O-one, one other point that I found mildly amusing as well, and it's another one of Matthias', but the fact that he suggested kids will look at the browser the way we look at the fax machine, and that it resonated with me'cause I, I remember as a a sort of junior consultant going'round and trying to trying to plug fax machines into analog gateways and just being perplexed by why the hell would anyone use these things these days.
00:19:19.976 --> 00:19:22.665
Not to say that people don't still use them'cause we do.
00:19:22.705 --> 00:19:24.016
We see that every now and then.
00:19:24.016 --> 00:19:32.056
But yeah, I just thought it was amusing that thinking about our kids, they, they might have the same level of perplexity looking at a browser and thinking,"What's this?
00:19:32.086 --> 00:19:32.516
What's this for?"
00:19:33.711 --> 00:19:34.291
Absolutely.
00:19:34.321 --> 00:19:46.692
I'm already noticing my browser use has dropped off considerably, and I expect, as not to cross over different technologies, but as Siri gets its game on, this year and through next year that'll drop even more.
00:19:46.922 --> 00:19:47.932
So yeah, really interesting.
00:19:51.279 --> 00:19:54.589
I've managed to grab Simon, our CEO and leader.
00:19:54.680 --> 00:20:02.359
Alex and I were just talking about your your talk just now, and I said to Alex, it really feels like AI agents are having their moment at the moment.
00:20:02.859 --> 00:20:04.369
Yeah, absolutely, Tom.
00:20:04.690 --> 00:20:08.059
We're seeing that in our business over the last probably three to six months now.
00:20:08.460 --> 00:20:17.480
Whereas maybe if you have 100 deals, 80 would be migrations, 20 would be AI agents six months ago, and that ratio's pretty much flipped, hasn't it?
00:20:17.480 --> 00:20:17.843
Yeah.
00:20:17.843 --> 00:20:18.204
Yeah,
00:20:18.204 --> 00:20:18.567
definitely.
00:20:18.567 --> 00:20:21.597
Yeah, it's been s- loads of and loads of talk as well actually afterwards.
00:20:21.597 --> 00:20:22.938
I thought that was really interesting.
00:20:22.958 --> 00:20:26.758
There's loads of interest, which is, it's a risk at an event like this, right?
00:20:26.758 --> 00:20:29.087
'Cause we're sponsoring it and, that's expensive.
00:20:29.137 --> 00:20:31.978
But actually I think in those conversations, that's where the value is as well, right?
00:20:32.397 --> 00:20:38.748
Yeah, and I think- the industry's been talking about AI for, since ChatGPT, which was November'23?
00:20:38.748 --> 00:20:39.488
Yeah.
00:20:39.488 --> 00:20:40.253
Yeah, so- It
00:20:40.253 --> 00:20:41.288
feels like a lifetime ago, right?
00:20:41.498 --> 00:20:41.728
Yeah.
00:20:41.807 --> 00:20:49.448
So there's- they've had a lot of time to think about it, and now, as I said in the presentation, you need- you're probably at the stage of should we do it?
00:20:49.448 --> 00:20:50.218
Could we do it?
00:20:50.228 --> 00:20:51.178
How do we do it?
00:20:51.498 --> 00:20:57.657
I think in six months' time, you're gonna get pressure from everywhere going,"We have to do it," because it's real.
00:20:58.067 --> 00:20:59.248
It's a real impact.
00:20:59.798 --> 00:21:05.867
Customer experience is one of the best Obviously coding is a, is an amazing area, and, we've all seen that over the last few months.
00:21:06.208 --> 00:21:09.137
But customer experience is one of the first business ones that should go.
00:21:09.448 --> 00:21:17.028
And again, not cutting staff, but actually increasing capabilities to be able to talk to your customers and service your customers as well.
00:21:17.038 --> 00:21:17.877
Yeah, definitely.
00:21:17.877 --> 00:21:18.018
Yeah,
00:21:18.018 --> 00:21:18.397
yeah.
00:21:18.597 --> 00:21:26.688
It, yeah it's mad in a way to think even a couple of years ago you wouldn't imagine that Amazon Connect would have their own regional series of road shows, right?
00:21:26.688 --> 00:21:27.788
Of multi-day events.
00:21:28.317 --> 00:21:28.528
Yeah.
00:21:28.528 --> 00:21:29.438
There's two things there, isn't there?
00:21:29.438 --> 00:21:41.067
There's the one where Amazon Connect was this amazing service amongst 200 services which, let's be honest, probably got a little bit lost in the conversations with AWS account managers and their customers.
00:21:41.428 --> 00:21:44.567
And now it's right at the front number one, let's go.
00:21:44.567 --> 00:21:45.347
So that's great.
00:21:45.678 --> 00:21:48.038
But then the second thing is the technology is there now, right?
00:21:48.038 --> 00:21:54.258
Whereas, we I know we've had some discussions, quite passionate internal discussions about, is the agent good enough?
00:21:54.337 --> 00:21:55.688
Is the voice good enough?
00:21:56.028 --> 00:21:57.028
And I think we're there now.
00:21:57.077 --> 00:21:57.667
Is it perfect?
00:21:57.857 --> 00:22:01.647
Probably not, but it's absolutely good enough to implement in the wild.
00:22:01.948 --> 00:22:04.988
I think people get really hung up about the quality of the voice, right?
00:22:04.988 --> 00:22:05.887
Is it human enough?
00:22:05.928 --> 00:22:07.018
Is it realistic enough?
00:22:07.018 --> 00:22:09.188
And it's yeah, but can you understand it?
00:22:09.208 --> 00:22:10.188
Can it understand you?
00:22:10.228 --> 00:22:11.147
Can you get your job done?
00:22:11.248 --> 00:22:15.678
And we're also seeing, aren't we, Tom, we're seeing different requirements from different businesses.
00:22:15.738 --> 00:22:15.887
Yeah.
00:22:15.917 --> 00:22:19.988
So some businesses are saying,"We definitely want to, have an AI agent.
00:22:20.008 --> 00:22:28.897
We want it to sound like an AI agent, and declare it's an AI agent." And then we're also seeing businesses say,"Actually, we want it to sound as human as possible." Yeah.
00:22:29.087 --> 00:22:36.347
And, there are pros and cons of both, and we're the, we're there to help on both,'cause like I see, EU Act, that's g- is that gonna be delayed now
00:22:36.637 --> 00:22:36.867
or?
00:22:36.877 --> 00:22:36.897
Yeah.
00:22:36.917 --> 00:22:39.748
From what I understand, bits of it are s- are, I think, are still happening.
00:22:39.768 --> 00:22:44.377
The requirement to say that you're gonna be recording something, I think that still stands.
00:22:44.428 --> 00:22:44.788
Okay.
00:22:44.988 --> 00:22:50.028
And I think that's, I think declaring,"Hey, I'm an AI agent, I'm here to help," I don't think there's anything wrong with that.
00:22:51.740 --> 00:23:03.849
And finally, I think, I guess for people who are on the fence about jumping into this AI world, and also aren't necessarily at this event, but they're listening to this what's the next thing they should do?
00:23:03.900 --> 00:23:05.920
Is there value to them in coming to events like this?
00:23:05.970 --> 00:23:06.990
Is that, does that help?
00:23:06.990 --> 00:23:13.849
And'cause it feels like everyone here is talking that story, but not every- it's not as if everybody's got it all sorted and has all the answers, right?
00:23:14.700 --> 00:23:14.950
Yeah.
00:23:14.950 --> 00:23:25.640
So this is gonna be a difficult answer,'cause I'm gonna sound absolutely one thing we all know about Cloud Interact and this broadcast is we don't wanna do a sales pitch at all, but you genuinely need to go and speak to experts.
00:23:25.759 --> 00:23:25.849
Yeah.
00:23:25.970 --> 00:23:34.369
And you need to, and you need to ask your AWS ac- account team to get you the experts in this area, whether it's us or somebody else.
00:23:34.970 --> 00:23:43.529
But you do need to speak to them, because there are lots of little things that can go wrong, and lots of little things that you need to do that, that is, is to do it right.
00:23:43.539 --> 00:23:45.509
So e- engage with people.
00:23:45.630 --> 00:23:51.730
And I felt a bit of a dummy actually,'cause I kinda said,"Let's not do demos," and But actually, and you know better than anyone, Tom, don't you?
00:23:52.039 --> 00:23:58.569
With the AI that, with the, we use internally now, we can demo within one or two days, as opposed to 90 days.
00:23:58.569 --> 00:23:58.759
Yeah.
00:23:58.910 --> 00:23:59.009
Yeah.
00:23:59.009 --> 00:24:00.039
So we can show it people.
00:24:00.039 --> 00:24:07.230
So engage with us, we'll show it, we'll demonstrate it, and then we can get you in a 12-week process that actually gets you to production.
00:24:10.637 --> 00:24:16.548
Okay, so if you weren't at the event and you're thinking about whether you should go for the next event, hopefully we can help sway you.
00:24:17.087 --> 00:24:25.637
So if you didn't go, I think for me, the biggest thing that you missed was genuine interaction with important people.
00:24:25.718 --> 00:24:37.067
And that's, people from AWS, but people from the community as well partners, and people that can answer your questions and talk honestly and openly about, what's working and what's not.
00:24:37.067 --> 00:24:39.738
And I think that interaction is really important.
00:24:40.948 --> 00:24:41.498
Alex, what about you?
00:24:42.968 --> 00:24:43.528
Oh, I totally agree.
00:24:43.528 --> 00:24:48.038
And, w-without having compared notes, that was going to be my takeaway from it as well.
00:24:48.087 --> 00:24:56.607
We remote work most of the time, and it is always quite a treat these days to go out and especially to an event like that with...
00:24:57.147 --> 00:24:58.498
It was a really cool venue.
00:24:58.817 --> 00:25:02.137
It was-- it seemed incredibly well-organized and slick.
00:25:02.468 --> 00:25:16.788
But like you say, there were loads of great people there old faces, new people to talk to but all super interested in Amazon Connect and it, yeah, having a community around it is really important, I think
00:25:17.369 --> 00:25:18.230
Yeah, definitely.
00:25:18.430 --> 00:25:24.400
I think the other thing you probably missed if you didn't go was the real feel for the AI is here right now.
00:25:24.500 --> 00:25:36.750
And I think the audience was probably divided by people that are already doing this and are sharing their stories, and then a bunch of people that are not doing it and are starting to feel like they're being a bit left behind.
00:25:37.589 --> 00:25:38.430
That's the sense I got.
00:25:38.480 --> 00:25:38.930
I don't know about you
00:25:39.528 --> 00:25:41.067
Yeah yeah, good point actually.
00:25:41.317 --> 00:25:42.377
That's worth mentioning.
00:25:42.377 --> 00:25:59.788
There were a lot of people I thought that were there as Connect customers or users and probably having a bit of the,"Oh, what's next?" Or, how do we tackle agentic AI?" And I think they were there for that knowledge share and to get some inspiration from other people in the community.
00:26:00.281 --> 00:26:00.612
Yeah.
00:26:00.692 --> 00:26:01.501
Yeah, definitely.
00:26:01.721 --> 00:26:04.882
It's a good chance as well to see the state of the possible.
00:26:04.932 --> 00:26:06.481
Th- that's what the demos are good for.
00:26:06.481 --> 00:26:07.892
They're showing you what's possible.
00:26:08.192 --> 00:26:15.132
And I think the demos did a good job from everyone really of showing what's possible today, and I don't think there was anything in there that's unachievable.
00:26:15.332 --> 00:26:20.251
So they weren't, far out, future thinking, marketing mock-ups or anything like that.
00:26:20.251 --> 00:26:24.852
They were real demos using the technology because, yeah, because it's ready, as we keep saying.
00:26:24.902 --> 00:26:25.882
The technology is there.
00:26:25.932 --> 00:26:29.082
It's all the business change stuff that needs to happen, but that takes time.
00:26:29.152 --> 00:26:35.372
So I think for the next 12 months yeah, I think you need to have started this already, probably.
00:26:35.551 --> 00:26:40.201
I think if you wanna keep maintaining a competitive advantage would be my biggest takeaway from last week.
00:26:40.692 --> 00:26:44.372
Was yes, this stuff takes time, but everybody's started it.
00:26:44.382 --> 00:26:47.692
So if you haven't started it, don't panic, I wouldn't say.
00:26:47.971 --> 00:26:49.731
But it's, don't delay, probably.
00:26:49.842 --> 00:26:51.442
And there's, there's different ways to do that.
00:26:51.471 --> 00:26:57.422
Go find somebody talk to AWS, go find an expert to talk to But I think do make a start.
00:26:58.061 --> 00:27:09.951
If you've got any kind of thinking that AI and agentic AI could change how you work for the positive, and I think that's most industries, then you probably ought to have made a start by now
00:27:11.928 --> 00:27:12.587
Yeah, that's a good point
00:27:13.087 --> 00:27:19.567
We could talk about this all day, and I'm sure we will keep chatting to each other about the the good bits and the bad bits from last week.
00:27:19.698 --> 00:27:21.597
But it is time to bring this episode to an end.
00:27:21.597 --> 00:27:25.907
Thank you very much, Alex, for your time and your thoughts and also for spending the day.
00:27:26.298 --> 00:27:28.468
It was a really good day to to spend with you last week.
00:27:28.897 --> 00:27:30.218
Thank you everyone for listening.
00:27:30.347 --> 00:27:32.837
Be sure to subscribe in your favorite podcast player.
00:27:32.847 --> 00:27:33.978
That way you won't miss it.
00:27:34.577 --> 00:27:36.958
Whilst you're there, we'd love it if you would rate and review us.
00:27:36.988 --> 00:27:40.397
If you have colleagues that you think would benefit from this content, please let them know.
00:27:40.488 --> 00:27:46.807
To find out more about how Cloud Interact can help you on your contact center journey, visit cloudinteract.io.
00:27:47.198 --> 00:27:50.768
We're wrapping this call up now, and we'll connect with you next time.