AI Agents and the Future of Customer Service

Transcript

WEBVTT

00:00:01.080 --> 00:00:02.685
Welcome to the Ecosystem Show.

00:00:02.685 --> 00:00:05.160
We're thrilled to have you with us here.

00:00:05.160 --> 00:00:11.153
We challenge traditional mindsets and explore innovative approaches to maximizing the value of your software estate.

00:00:11.153 --> 00:00:13.523
We don't expect you to agree with everything.

00:00:13.523 --> 00:00:17.472
Challenge us, share your thoughts and let's grow together.

00:00:17.472 --> 00:00:20.548
Now let's dive in it's showtime.

00:00:20.548 --> 00:00:25.312
Welcome back to the Ecosystems podcast.

00:00:25.312 --> 00:00:26.344
Glad to have you with us.

00:00:26.344 --> 00:00:28.748
Sorry, it's not Ecosystems podcast.

00:00:28.748 --> 00:00:31.364
Ecosystems show it's the Ecosystems show.

00:00:31.666 --> 00:00:41.869
Yeah, yes, because YouTube prefer to call them shows rather than podcasts, because they feel like a show is video, includes video, and this is what this does.

00:00:42.299 --> 00:00:43.701
And who wouldn't want to see our charming faces?

00:00:43.654 --> 00:00:45.200
Exactly, exactly, exactly, includes video, and this is what this does.

00:00:45.200 --> 00:00:45.594
And who wouldn't want to see our charming faces?

00:00:45.594 --> 00:00:51.192
Exactly, exactly exactly, um, even my especially, especially now?

00:00:51.213 --> 00:00:59.008
that I'm, I'm back from, I'm back from from, um, uh, causing an airplane, an emergency diversion of an airplane.

00:00:59.008 --> 00:01:02.963
So, uh, I'm, I'm, wow, I'm looking well, if I do say so myself.

00:01:03.423 --> 00:01:11.253
So you, you, you on an air, on a on a flight, caused an emergency situation that had to divert and landed an airport.

00:01:12.480 --> 00:01:14.867
Yes, yes, wow, which I mean.

00:01:14.867 --> 00:01:15.367
It's not like.

00:01:15.367 --> 00:01:15.950
It's not like.

00:01:15.950 --> 00:01:19.974
It wasn't a goal, it wasn't a goal of mine and I was not part of.

00:01:19.974 --> 00:01:23.450
I was barely conscious, I was not part of the decision-making process.

00:01:24.379 --> 00:01:25.718
Yeah, wow, wow, yeah, wow, wow.

00:01:25.718 --> 00:01:27.183
So medical instance eh.

00:01:27.887 --> 00:01:33.530
Yes, yes, it was not the most fun I've ever had on a plane.

00:01:33.670 --> 00:01:39.850
I bet I bet Frightening really, but glad everything pulled through and you're okay.

00:01:39.850 --> 00:01:42.066
That's the good outcome of this.

00:01:42.066 --> 00:01:44.087
As you can see, there's only the two of us today.

00:01:44.087 --> 00:01:50.745
Our three colleagues weren't able to join us, but I'm sure they'll be back in the future.

00:01:51.367 --> 00:01:53.927
Not too long Is Will back in Scotland.

00:01:53.927 --> 00:01:57.468
That's just where we say Will is when he's missing.

00:01:58.540 --> 00:02:02.807
Right, right, I assume he's back in his manor.

00:02:06.724 --> 00:02:24.468
Do you remember when we did the show, kim, and we, for whatever reason, we recorded two episodes back to back on the same night, on the same day, and Will, not wanting it to appear that we had recorded them back to back, the rest of us just went and changed our clothes.

00:02:24.468 --> 00:02:30.947
Will took down the thing, the art on his wall, and swapped it out, and swapped it for different art.

00:02:30.947 --> 00:02:34.241
And when we talk about the art on Will's wall, what was it like?

00:02:34.241 --> 00:02:37.711
A red cow head, yeah.

00:02:38.479 --> 00:02:40.908
Hanging like he had shot a cow or something like that.

00:02:41.340 --> 00:02:43.609
It's what I would call eclectic art, right?

00:02:43.609 --> 00:02:49.568
Eclectic, I think, is the phrase that you would use for his artwork of choice.

00:02:49.568 --> 00:02:52.686
So I don't know.

00:02:52.686 --> 00:02:58.161
Just lately, oh, I tell you what a little personal story Today.

00:02:58.161 --> 00:03:02.451
In what's the time now?

00:03:02.451 --> 00:03:09.264
In my time it is eight, a bit after eight in the morning, at one o'clock.

00:03:09.264 --> 00:03:17.274
Today I will have my first food that I haven't eaten for seven days.

00:03:18.661 --> 00:03:19.664
So you did it, huh.

00:03:20.006 --> 00:03:20.448
I've done it.

00:03:20.448 --> 00:03:22.944
Yeah, Water only seven days.

00:03:22.944 --> 00:03:25.592
Dropped about five kgs in seven days.

00:03:27.061 --> 00:03:28.888
As an American, I don't know how much that is.

00:03:29.500 --> 00:03:30.906
Yeah, sorry, I don't know what it is.

00:03:31.260 --> 00:03:32.044
I guess that's good.

00:03:32.044 --> 00:03:32.967
I guess it's good.

00:03:33.762 --> 00:03:35.099
Let me just do the conversion.

00:03:35.099 --> 00:03:38.830
I think it's about two and a half kilograms to the pound.

00:03:38.830 --> 00:03:44.451
It is 11.2 pounds.

00:03:44.451 --> 00:03:51.485
There you go, there you go, good job.2 pounds.

00:03:51.485 --> 00:03:52.247
There you go, there you go, good job.

00:03:52.247 --> 00:03:53.111
11 pounds, nice one um.

00:03:53.132 --> 00:03:54.216
So yeah, I'm looking forward to my first meal.

00:03:54.235 --> 00:03:54.317
What?

00:03:54.317 --> 00:03:55.060
What's it gonna be an apple?

00:03:58.348 --> 00:04:00.432
not a, not a cheeseburger or something.

00:04:00.432 --> 00:04:01.294
It's gonna be an apple.

00:04:01.355 --> 00:04:03.741
No, no, no, no, no, no, because you gotta, you gotta.

00:04:03.741 --> 00:04:12.963
You know your body's gone, like your stomach particularly is, you know, uh has been operating for seven days as such, apart from you know detoxing.

00:04:12.963 --> 00:04:31.069
That's wild, and so yeah, well, actually I'll take a banana first, I'll wait about an hour, and then I'll have an apple, and I'll probably wait another hour and then I'll have a chicken and vegetable soup, and that just so your system can turn back on without you know blowing it out, type thing.

00:04:31.829 --> 00:04:33.812
So is this the first time that you've done this?

00:04:34.913 --> 00:04:36.334
No, I've done seven days.

00:04:36.334 --> 00:04:39.036
So it's 10 years ago when I did my last seven day fast.

00:04:39.036 --> 00:04:49.225
And then, as I, when I say fast, that for me, um, the way my brain runs is, I am very.

00:04:49.225 --> 00:04:57.173
It has to be no food whatsoever, no coffee, nothing, no liquid outside of water.

00:04:57.173 --> 00:04:58.983
So it's a water only fast.

00:04:58.983 --> 00:05:06.668
And in my past, about 19, 20 years ago, in one year, I did two 21-day fasts in one year.

00:05:06.668 --> 00:05:08.773
That's wild, that.

00:05:08.773 --> 00:05:09.454
What's wild?

00:05:09.454 --> 00:05:21.610
Water only um the first time I did a seven day fast, which is prior to that again, I was working, still full-time, and I was in kind of manual labor role, yet I was still able.

00:05:21.911 --> 00:05:34.584
Your body's amazing and what it can achieve and do and stuff, especially when you go over the about the three-day marking for males and and um, the difference between males and female fasting are worlds apart.

00:05:34.584 --> 00:05:39.019
There's no correlation or similarity that that you can do so.

00:05:39.019 --> 00:05:47.274
For example, the rule of thumb for females is that they should never fast the week prior to their period at all at all.

00:05:47.274 --> 00:06:08.281
It will just be because this is and it's interesting, I've learned you know, guys pretty much have one hormone as such, testosterone, running through the systems, um, where women have um, estrogen, testosterone and progesterone, I think the other one is um, and so fasting can really fuck up their hormones massively.

00:06:08.281 --> 00:06:13.822
So there's there's nowhere near the amount of studies done on female fasting as there are on males.

00:06:14.423 --> 00:06:17.408
Interesting, yeah, I crazy thing.

00:06:17.408 --> 00:06:18.711
I haven't felt hungry at all.

00:06:18.711 --> 00:06:19.452
Not not once.

00:06:19.452 --> 00:06:22.764
Not once have I thought, oh, I should have a craving or anything like that.

00:06:22.764 --> 00:06:29.814
And I think it's because for the last six months I've been doing intermittent fasting, which is seven days a week.

00:06:29.814 --> 00:06:36.750
I only eat after 12, um in the day and before eight, and at 8 pm at night I stop eating.

00:06:36.750 --> 00:06:39.514
So a 16 by 8, what that?

00:06:39.514 --> 00:06:41.624
What that's referred to, and so I think it made.

00:06:41.624 --> 00:06:50.487
This is probably the easiest fast I've had to do from a no kind of, you know, cravings or anything like that.

00:06:50.487 --> 00:06:52.812
It's been amazing awesome.

00:06:53.153 --> 00:07:03.971
Well, you, the the ecosystem show, come for the uh, come for the tech and stay for the armchair, medical or nutrition advice.

00:07:03.971 --> 00:07:17.603
Just just testimony, testimony right mark's experience is not an indicator or a recommendation, so it's uh yeah, so exactly, you know, as I say, do your research, do your research, do your research.

00:07:17.665 --> 00:07:27.985
So, um, I tell you what one word has been popping up everywhere at the moment for me and of course that means I'm kind of focusing on a new area and that word is agentic.

00:07:27.985 --> 00:07:34.142
So I had to understand what does this word mean, agentic and um?

00:07:34.142 --> 00:07:35.665
It's basically like.

00:07:35.665 --> 00:07:41.134
Actually, I'll read you the definition um from chat gpt um.

00:07:41.134 --> 00:07:46.494
It refers to being agent, like or having the characteristics of an agent.

00:07:46.494 --> 00:07:55.581
In a philosophical context, it often describes an individual's capacity to act independently and make their own free choices.

00:07:55.581 --> 00:08:08.045
In terms of ai and technology, it describes systems or processes that can act autonomously or make decisions without human intervention, effectively mimicking human agency.

00:08:08.045 --> 00:08:10.612
What do you think about this?

00:08:11.860 --> 00:08:13.122
Well, you know so.

00:08:13.122 --> 00:08:14.326
So a couple of thoughts.

00:08:14.326 --> 00:08:41.846
One, um, for some reason this fired in my brain when you were reading this, right, but you know, when you um, uh, we've had we, we humans, have had other humans acting as our agent for a long time, right, so people of a people of a certain age will remember the travel agent you know, remember that was, that was the only way you could book a flight for a very long time.

00:08:41.846 --> 00:08:52.792
Right, you had you had to call a travel agent because only they had the you know weird, uh, command line terminal that they could go book you a flight.

00:08:53.173 --> 00:08:56.408
Galileo, I think, was their, their, their back-end system is two.

00:08:56.408 --> 00:09:04.431
There's galileo, and I forget the other one, which is still the way all flights happen around the world is still in those two Amadeus and Sabre.

00:09:04.431 --> 00:09:06.014
Yes, there's three.

00:09:06.173 --> 00:09:19.250
Yes and well, and those were Amadeus and Sabre are, like you said, those are two, I know, still used by almost every airline on the planet to manage their flight operations, right?

00:09:19.250 --> 00:09:28.934
But so, anyway, you would have these travel agents and a human would hire a travel agent to act on their behalf to go book them, to go book them a flight.

00:09:28.934 --> 00:09:46.157
Or, if you are, we said, people of a certain age, well, people of a certain level of wealth and fame might hire an agent to represent them in a contract negotiation, say a sports agent or a talent agent, or you know what have you.

00:09:46.157 --> 00:09:50.746
So, you know, it's a kind of like co-pilot, right?

00:09:50.746 --> 00:10:02.990
We've taken this word that has real meaning in human to human contexts and we've grafted it onto this paradigm of artificial intelligence.

00:10:02.990 --> 00:10:09.327
And it's, it's, it's, you know, it's similar to how Windows, you know, we called, we called it Windows.

00:10:09.488 --> 00:10:26.716
or, you know, desktop operating systems have always had this metaphor of the folder and the desktop itself you know so so in some ways it's really just, you know, it's using concepts that are familiar as a way to describe what's happening in AI.

00:10:26.716 --> 00:10:34.192
So that was a completely in the moment thought that I had when you were reading that definition.

00:10:34.840 --> 00:11:00.385
Yeah, it's so valid and I think that I think by Q1 next year, I think this whole concept of agents is really going to sink in for people, and I don't know if you saw the announcements this week from um uh open ai, but they brought out real-time voice-to-voice um apis and yeah, it did.

00:11:00.606 --> 00:11:01.408
And you know what?

00:11:01.408 --> 00:11:16.113
What's totally epic about this, you literally will be able to create an agent that is your outbound calling agent and so, let's say, your airline mucks things up or whatever.

00:11:16.113 --> 00:11:21.131
You'll be able to say to your agent hey, can you put a call in and make a claim or do whatever.

00:11:21.131 --> 00:11:25.326
They'll call up that contact center as a human level voice.

00:11:25.326 --> 00:11:31.803
They will not be able to tell the difference and say, hey, listen, this is about my flight, here's my date, like they'll have all your details that you've given it.

00:11:31.803 --> 00:11:45.765
And when they go, oh no, sorry, we can't do anything about it, they'll refer to policy 5963 on their website and go well, according to your policy, here it says you, like they'll have all that detail, like we would never bother going and reading all that stuff, right?

00:11:46.606 --> 00:11:58.720
first of all, I just I've been having the weirdest feud with the, with my, with our property management company that we're you know we pay service fees to like condo fees.

00:11:58.720 --> 00:12:16.749
They have taken, uh, they, they took, uh, I'll you the details, but five months ago I spoke with them about solving a problem that they themselves have created and they finally got back to me yesterday and, lo and behold, they needed five months to actually make the problem worse.

00:12:17.309 --> 00:12:38.490
So I would love property management cable companies all of these hated industries right To just have an agent that calls and deals with them for me, and you know, and that agent has it, understands all the laws, the bylaws of your area, like how you can get action from them, and like the detail.

00:12:38.490 --> 00:12:42.530
And so you imagine that operator on the line, man, they're just going to not know what hits them.

00:12:42.530 --> 00:12:47.789
All of a sudden, these very knowledgeable people are calling in agents.

00:12:47.809 --> 00:12:57.461
No, all of a sudden, all of a sudden, the airline or the property management company, they're going to have their own agent and it's just two agents are going to duke it out.

00:12:57.461 --> 00:13:07.033
It's going to be like the weirdest, the weirdest pub trivia you've had, you know, like when two nerds get together and argue about the minute details of a problem.

00:13:07.600 --> 00:13:10.089
I actually think that will get to a solution quicker, right?

00:13:10.089 --> 00:13:16.293
Because what happens, right is that human nature is kick the can down the road, right?

00:13:16.293 --> 00:13:26.773
You got a problem Like I don't know if you've noticed, in the airlines, and particularly was prevalent when I was flying Virgin Australia and, you know, frequently in Australia.

00:13:26.773 --> 00:13:34.110
And so you'd go to the check-in desk and say, listen, can I upgrade my seat to XYZ?

00:13:34.110 --> 00:13:37.046
And they're like, oh, I'll go and ask in the lounge.

00:13:37.046 --> 00:13:37.827
They'll do that.

00:13:37.827 --> 00:13:40.591
So you go to the lounge, yeah, can you upgrade your seat?

00:13:40.591 --> 00:13:49.380
Oh, listen, when you go on the aircraft, ask the person that when you're boarding the aircraft and what they do is this kick down, like I'm not going to say no to you, but I'm going to kick it down.

00:13:49.380 --> 00:13:51.405
And it's the next person in the chain.

00:13:51.886 --> 00:14:08.254
and of course you get there and they're like, oh no, they should have done that for you back at the service desk and it's too late, right and it's like but the thing is, I think if you have this agentent interaction and their goal is a solution, they're not going to do any of that kicking the can down the road.

00:14:08.254 --> 00:14:09.662
They will.

00:14:09.662 --> 00:14:12.268
Hey, yes, you know the policy, I know the policy.

00:14:12.268 --> 00:14:12.929
We're going to do this.

00:14:12.929 --> 00:14:14.033
Bang Solved.

00:14:14.033 --> 00:14:16.349
I think it'll strip out all that crap.

00:14:28.460 --> 00:14:38.591
Yeah, you know, but I think it'll be very interesting to see how, in this and I want to, in, I do, I want to, I want to lose track of, kind of talking about how, the, how early days of these announcements that we've seen over the last, you know, month or so have have gone and you know how I, how we think they're playing.

00:14:38.591 --> 00:15:01.355
But before we get there, I do, I do think that it'll be interesting to see how the personalities of the humans to whom the agents are accountable, basically like the personalities of their creators or the personalities of their masters, what happens there?

00:15:01.355 --> 00:15:05.798
Right, and I'll give you, I'll give you, I'll give you a scenario.

00:15:05.798 --> 00:15:15.075
Right, there are, there are plenty of people in the world who you know, you are definitely one of them.

00:15:15.075 --> 00:15:23.687
You know, I have things, we all listen, we all have things that we really, we do want to control, right, like, for me, my, my weakness are slides.

00:15:23.687 --> 00:15:33.438
Right, I don't know one, I don't like anyone else's slides, but for the most part, I am very happy to delegate something.

00:15:33.458 --> 00:16:00.903
You know, if I think back, you know, if I think of my experience as as a manager and a supervisor of people in the workplace, right, the people who I work best with have always been you know, as a, you know as direct reports of mine have always been people who are super autonomous, super willing to sort of go do their own thing, because I am perfectly happy for my team to go do their own thing.

00:16:00.903 --> 00:16:08.950
Yeah, I really just care sort of about the result at the end and whether or not you've acted, you know, like a decent human being in order to get there.

00:16:08.950 --> 00:16:14.770
But I don't, I don't care where you are, I don't care what you do with your time, I don't care how you do.

00:16:14.770 --> 00:16:19.283
You know, I mean, I don't care about most of those things, but I think that there is a whole.

00:16:19.283 --> 00:16:32.956
So for me, the idea of turning an agent loose to go deal with things and to be autonomous in that way and to report back to me and to come back to me with the result, and then we'll decide if we like the result.

00:16:47.720 --> 00:17:02.274
Personality type that is going to be very uncomfortable with the idea of turning someone loose, whether it's an AI agent or a human agent, turning someone loose to go do the thing for them or on their behalf, and then add into that the layer of what if you are dissatisfied?

00:17:02.274 --> 00:17:07.586
What if the human master is dissatisfied with the outcome that the agent has produced?

00:17:07.586 --> 00:17:07.846
Right?

00:17:07.846 --> 00:17:13.107
So the airline's agent and your agent have added about the policy and they come to a conclusion.

00:17:13.107 --> 00:17:14.634
Well, is the human.

00:17:14.713 --> 00:17:17.964
If the human doesn't get the result they're looking for, are they going to be pissed?

00:17:17.964 --> 00:17:19.750
And I think the answer is yes.

00:17:20.519 --> 00:17:29.094
So here's the, here's the thing, and this, this comes, I don't know to a degree, to a maturity level in understanding the value of time.

00:17:29.094 --> 00:18:08.644
Right, and one of the modern-day philosophers I follow is a gentleman I can't remember his name, it's a very multi-long name, nivelle I think he goes by as his handle In fact that's his first name, neville and and what he says is that he worked up early on in his life the concept of money and time and how they interrelate so he said if it's going to take me two hours to go to a mall and buy something, but I could buy that online and have it shipped to my door, no brainer.

00:18:09.384 --> 00:18:11.108
Even if there's a career fee, no brainer.

00:18:11.108 --> 00:18:12.230
Right, and he go.

00:18:12.230 --> 00:18:15.226
He looked at every area of his life and said is the?

00:18:15.226 --> 00:18:28.688
He said like, let's say he go, has a product that fails and he realizes the time to get it corrected even he could get corrected is going to take out of his life, is going to be more than what he values.

00:18:28.688 --> 00:18:32.017
That product He'll like oh well, let it go.

00:18:32.017 --> 00:18:32.617
Yeah, who cares?

00:18:32.617 --> 00:18:33.118
Who cares?

00:18:33.118 --> 00:18:37.548
Because I value my time more importantly than my bank account.

00:18:37.548 --> 00:18:39.099
I have the funds to do it.

00:18:39.099 --> 00:18:40.726
I'd rather pay for it to be done.

00:18:40.787 --> 00:18:50.122
And he and he talked about his relationship with his mother and stuff, where she would ask him to do stuff and he would just get someone else to do it and pay them because it was a good use of his time.

00:18:50.122 --> 00:18:53.775
It was just, and I think that that's what that's going to come back to.

00:18:53.775 --> 00:18:54.915
Is that for me, like?

00:18:54.915 --> 00:18:59.818
I'm like, bring it on, because, whatever the answer is, I'm happy.

00:18:59.818 --> 00:19:09.284
I just didn't have to go through the process and stress and argument with somebody on a phone call to get the outcome that I am seeking is that?

00:19:09.284 --> 00:19:18.368
What if the airline said to their agent programming listen, make sure you do everything you can to not pay out.

00:19:18.368 --> 00:19:30.361
Right, you need to use tactics of deflection and avoidance and ambiguity and stuff Like the Talco in New Zealand Spark.

00:19:30.361 --> 00:19:36.142
They got a law case against them because they made their pricing so confusing on purpose to deceive the customers over time.

00:19:36.142 --> 00:19:37.003
It was.

00:19:37.665 --> 00:20:09.961
It was brought to the commission and they were found that they tactfully went out of their way to screw customers over on purpose, so it's ryanair, but as an electricity was it an electricity company telecom telecom a telecom, a telco so it was equivalent, equivalent to British Telecom, the same entity Government, originally government loan telecommunication became privatized and then went through this process of doing everything to screw their customers over, knowing they had no choice, there's nobody else they could go to.

00:20:10.674 --> 00:20:25.876
Long-time listeners or repeat listeners of the podcast will will, of course, appreciate that the eco or the show, the ecosystem show, takes a long running position that Ryanair is the worst company on the planet.

00:20:25.876 --> 00:20:35.603
So just just so we can get as we can get, that out there again, um classic classic, the worst, the worst so back to agents right now.

00:20:35.804 --> 00:20:59.663
Like I got chat gpt yesterday to make me a 40 hour course to go deep into agents and learn it and understand it and chat gpt make a course for you yeah, and like it, gives me, it goes out and does all the research and says I recommend you know, uh, you know, on monday in the morning, you, you're going to spend four hours, you're going to get hands-on, and I've got it doing it.

00:20:59.663 --> 00:21:02.183
I said I don't want just a single view of this.

00:21:02.183 --> 00:21:08.539
I don't want just Microsoft's view, I don't want just OpenAI's view, I want AWS's, I want Google's as well in that mix.

00:21:08.539 --> 00:21:11.728
And cheers, cheers.

00:21:11.728 --> 00:21:14.342
I hope to have one of those later today.

00:21:17.795 --> 00:21:18.257
And it will allow.

00:21:18.257 --> 00:21:21.405
Well, it is 8.30 at night on my end, Exactly exactly.

00:21:21.405 --> 00:21:23.000
We're 12 hours separate here.

00:21:23.623 --> 00:21:24.325
I understand.

00:21:24.325 --> 00:21:26.895
So yeah, going deep on that Now.

00:21:26.895 --> 00:21:35.268
Another thing announcement made this morning by OpenAI which I find intriguing is they brought out this idea of ChatGPT Canvas.

00:21:39.936 --> 00:21:41.000
So is that like co-pilot pages?

00:21:41.000 --> 00:21:41.319
It's identical.

00:21:41.319 --> 00:21:41.760
It's almost like.

00:21:41.760 --> 00:21:43.965
When I saw the interface I'm like it's identical.

00:21:44.105 --> 00:21:44.386
This is.

00:21:44.386 --> 00:21:45.796
I was like what the fuck?

00:21:45.855 --> 00:21:52.363
it's identical so I'm currently I am currently super annoyed with my co-pilot.

00:21:52.363 --> 00:22:03.967
So so the background, the background on this is that, um, we uh, several weeks back we started basically the cloud lighthouse policy.

00:22:03.967 --> 00:22:12.144
The company policy for whether you get a co-pilot for M three 65 license is if you ask for one and you shall have.

00:22:12.144 --> 00:22:16.130
So Mark has not asked for his yet, so he does not have one, correct.

00:22:16.130 --> 00:22:17.596
So anyway, I have one.

00:22:17.596 --> 00:22:36.509
But when I saw the co-pilot pages announcement a few weeks back, I thought to myself, because this is probably something that you and I should unpack the different ways that you and I use or look at AI in our personal, like in our own individual work life.

00:22:36.509 --> 00:22:45.421
Right, but so anyway, I saw this announcement for Copilot Pages and I thought now that that is useful.

00:22:45.421 --> 00:22:46.921
It's not earth shattering.

00:22:46.994 --> 00:22:49.045
It's not groundbreaking, but it's it's.

00:22:49.045 --> 00:22:52.659
It's massively useful and I cannot.

00:22:52.659 --> 00:22:57.749
So I cannot get it to appear in the Cloud Lighthouse tenant.

00:22:57.749 --> 00:23:07.178
And what annoys me right now is that I have asked Copilot how to make Copilot pages appear and it is not aware.

00:23:07.178 --> 00:23:10.646
It's not aware that that capability has not yet been offered to me.

00:23:10.935 --> 00:23:12.863
It's just the rollout ring right.

00:23:12.863 --> 00:23:14.059
It started mid-September.

00:23:14.059 --> 00:23:21.984
The US tenants would have got it first up, you know, in specific areas, and then it'll be their high usage, you know, customers.

00:23:22.935 --> 00:23:24.663
We're a US northeast.

00:23:24.934 --> 00:23:26.358
But do you have 100,000 users?

00:23:27.803 --> 00:23:29.105
in your tenant?

00:23:29.105 --> 00:23:29.606
No, we are.

00:23:30.458 --> 00:23:41.890
Microsoft got customers of that size on Copilot right, so I think it's those ones that probably get it first and then rolls out to ultimately the rats and mice and stuff.

00:23:41.890 --> 00:23:44.057
That are just two or three licenses around the world.

00:23:44.118 --> 00:23:48.798
But I tell you what Eventually, Microsoft's most valued partner will get the feature.

00:23:49.338 --> 00:23:50.540
Exactly exactly.

00:23:50.540 --> 00:24:00.925
So I find this intriguing and my question that jumps to mind is did Microsoft do a deal with OpenAI?

00:24:00.925 --> 00:24:05.063
Because when I look at the conversation thread, they've been working on this for about two years.

00:24:05.063 --> 00:24:36.883
Did Microsoft have a deal with OpenAI that they obviously will know way before the public about feature development and stuff based on their relationship, did they do a thing and said, hey, can we announce ours first and get that feature and we'll build on it and we'll put the, the Microsoft skin over the top, which, remember, satya, in his keynote, calls this UI for AI, which is a really, really it's this.

00:24:36.903 --> 00:24:45.655
Then, this week, once again, I've been doing a bunch of work in this space and it's really blowing my mind up large what ui for ai really means.

00:24:45.655 --> 00:25:09.424
Because I was doing some research and I googled a service now, um, sap, oracle, you know the Oracle's HR system, sap's HR system, workday Salesforce, maybe one other and I searched for what was there.

00:25:09.424 --> 00:25:16.383
So first I did a prompt that asked for what were the names of the AI tools that they had for each of those brands.

00:25:16.383 --> 00:25:23.624
It came back and then I searched those up, looking just for screenshots of their interfaces right Of their chat.

00:25:23.624 --> 00:25:30.267
So I ended up with a on a PowerPoint slide with maybe six of these different things.

00:25:30.914 --> 00:25:51.392
The problem with each product producing its own chat experience like this is that employees in the organization that are already overwhelmed by the 50 apps that they have to use each day have now got 50 different AI tools to interact just with the data inside that data set.

00:25:51.392 --> 00:25:52.717
It's not holistic.

00:25:52.717 --> 00:26:02.883
You want to have an interface that goes across all your data sets because this data set in this system might correlate to another data set in another system.

00:26:02.883 --> 00:26:18.166
But the model that all these point solutions have gone for is no, it's just about my data set, and I think that Microsoft has made an incredible tweak in in what sachia said there.

00:26:18.186 --> 00:26:30.885
Uh, the ui for ai is going to be so important because I think what's going to happen is in business, people are going to want one chat interface agent to their data on their desktop.

00:26:30.885 --> 00:26:33.695
They don't want 50 of them, right?

00:26:33.695 --> 00:26:40.703
Oh shit, that's in that system and it like well, you're just going to do context switching, app switching, just like we've been doing for years.

00:26:40.703 --> 00:27:10.127
With this model from Microsoft, you can use Copilot Studio, connect into all those data sets across your organization the workday, the service, now the whatever and when you prompt, you're going to have that context window across your entire data estate, and I think that was probably one of the smartest moves Microsoft have just made in their Wave 2 update.

00:27:12.318 --> 00:27:14.645
So I have lots of thoughts on this topic.

00:27:14.645 --> 00:27:21.126
Let me see if I can quickly share them in the time it takes for me to not forget what they are.

00:27:21.126 --> 00:27:31.525
Yes, so first of all, I have been surprised over the last gosh.

00:27:31.525 --> 00:27:33.406
Are we at two years now?

00:27:33.666 --> 00:27:35.147
Yeah, nearly two years November, it'll be two years.

00:27:35.147 --> 00:27:35.509
Yeah, yeah, yeah.

00:27:35.509 --> 00:27:37.250
Two years November, it'll be two years yeah.

00:27:37.471 --> 00:27:38.471
Chat, gpt Right.

00:27:38.471 --> 00:27:53.970
So so, for all of the advancements that we have made in the technology over the last two years, I have been very surprised at the kind of the stagnation of the user experience Right.

00:27:54.797 --> 00:27:57.285
Like and I still very much believe, right?

00:27:57.285 --> 00:28:20.683
You know, and anyone who's read my white papers or in blogs on this topic know that, and comparison that I like to use is that when the iPhone came out, for a few years, developers spent a couple of years trying to cram desktop applications into a smaller screen before realizing that the iPhone was a fundamentally different form factor and suited for different things.

00:28:20.683 --> 00:28:23.215
And then the same thing happened with the watch right.

00:28:23.215 --> 00:28:34.001
For the first couple of years after the apple watch came out, we had these iphone apps that were just like really, really teeny, tiny, tiny apps on our phone to use, right, right.

00:28:34.001 --> 00:28:48.808
So I'm very confident that we're going to see the world, the global developer community and user experience, designers, et cetera, et cetera.

00:28:48.808 --> 00:28:55.844
We're going to land on the UI for AI and we're going to probably have several right.

00:28:55.844 --> 00:29:02.145
This is not just like a one device type of thing, but I am surprised that it's taken so long.

00:29:02.734 --> 00:29:06.224
Yeah, but I think any UI has to be multimodal by default.

00:29:06.224 --> 00:29:08.560
Yes, right, just as a given.

00:29:39.967 --> 00:29:40.186
Yeah.

00:29:40.186 --> 00:29:42.468
So I think that that definitely is.

00:29:42.468 --> 00:30:17.807
One just observation is that I am surprised that the technology, the vendors of the companies making this technology and you know there are thousands and tens of thousands of software companies that are doing things let's focus on, um, let's focus on four of them, on the big four, and I'm going to set meta, facebook aside, because facebook and meta are such a fundamentally different proposition than microsoft, apple, google or or Alphabet and Amazon.

00:30:17.807 --> 00:30:19.049
Okay, yeah, yeah.

00:30:19.049 --> 00:30:41.723
So I think that the companies, so you basically have two ends of a spectrum here, at which one end you've always had sort of this best of breed philosophy, and I'm going to be generous and call it a philosophy that Salesforce, amazon, the Googles of the world have had.

00:30:41.723 --> 00:30:50.397
You know, like you should get G Suite or whatever it's called now, like Google for your office productivity, and you should get Amazon for your cloud infrastructure.

00:30:50.397 --> 00:31:39.382
You should get Salesforce for your business, you know, for your CRM or whatever, and you should piece all these things together right in a way that you know for your CRM or whatever, and you should piece all these things together right In a way that you know lets you use what they would argue the best of breed, and then on the other end of the spectrum, you have the Apple and the Microsoft approach, right, apple and Microsoft, of course, being the adults in the room that you know they've been around the longest, they're the most mature, they have kind of the best end to end, in my opinion, offering in their particular space, and I think that it is the two technologies that I'm the most you know that I think are very interesting in terms of making the first crack at this user experience thing are Copilot from Microsoft and Apple Intelligence from Apple, which hasn't even been released yet.

00:31:39.736 --> 00:32:09.210
And the reason I'm so interested in where Copilot and Apple Intelligence go over the next couple of years on this issue is because Apple and Microsoft are really the only two companies that have as far reaching a device and or software and or data ecosystem to be able to create that one window to artificial intelligence across all of your data.

00:32:09.210 --> 00:32:17.076
It simply is not within the realm of something that Google or Salesforce or Amazon and certainly Meta are even able to produce.

00:32:17.076 --> 00:32:20.546
So it'll be really interesting to see where Apple and Microsoft take this.

00:32:21.214 --> 00:32:34.037
Yeah, see, for me, I feel that Apple is going to have my personal data right and it'll have agency across my personal data, and then Microsoft will have agency across my business data.

00:32:34.037 --> 00:32:37.626
That's the kind of the view I take of it.

00:32:38.115 --> 00:32:52.744
I mean so, says so, probably would agree all of the many, many, many, many, many of us who have been rolling around with Outlook installed on their iPhone for years, ever since what was it?

00:32:52.744 --> 00:32:54.317
It was circa?

00:32:54.317 --> 00:33:00.288
It was around 2013 when Microsoft acquired Sunrise.

00:33:00.288 --> 00:33:04.727
It was Sunrise Calendar, which was a great calendar app, turned it into Outlook.

00:33:04.727 --> 00:33:08.164
So yeah, no, I mean for sure, for sure.

00:33:08.184 --> 00:33:19.900
Yeah, I just think it's going to be incredibly interesting times in that you know, this concept of agents is so far reaching and how.

00:33:19.900 --> 00:33:33.138
You know if you I've thought about it from a home automation perspective and and having it intelligent enough to know temperatures, know the lighting conditions and adjusting everything just automatically.

00:33:33.138 --> 00:33:39.857
Based on that, knowing that, you know, I've noticed that every monday you normally do your grocery order and have it delivered.

00:33:39.857 --> 00:33:47.517
You skipped it this Monday, but I know that you'll be annoyed later in the week that you didn't do it, so I'll just order it for you.

00:33:47.517 --> 00:33:55.061
I know all the favorite products and stuff you want and all of a sudden the truck arrives outside and I'm like fantastic.

00:33:56.329 --> 00:33:58.012
Well, how fabulous will be.

00:33:58.012 --> 00:34:12.376
Think of the quality, think of the quality, the improvement in the quality of the modern marriage, if only that, if only that use case can come to fruition well, and then you take that a step further.

00:34:12.436 --> 00:34:19.822
Right, I saw this week a tiktoker from australia where a guy has basically created a free chrome extension.

00:34:20.684 --> 00:34:24.936
And there's there's two main shopping brands in australia they're coles and they're woolworths.

00:34:25.237 --> 00:34:27.811
For your grocery shopping, those are the two brands.

00:34:28.492 --> 00:34:41.764
And what he did is that, before he announced this, he plotted for six months the price point on night, the like the 80 of products that people buy as staples every day in their life.

00:34:41.764 --> 00:35:07.380
And what he noticed is that there was this whole cyclical price that then would get discounted a week later and it was kind of like this wave effect of the price in this week it's discounted, the price in this week it's discounted, and this is for not perishable foods, but your hand wash, your laundry detergents and stuff like this.

00:35:07.380 --> 00:35:19.934
And so he then started monitoring across both brands and he noticed that there was an offset If you took the exact same product from two different stores, they would offset their discounting cycles between each other on this weekly scale.

00:35:19.934 --> 00:35:27.373
And so what he did is that this chrome extension allows you to go in and it can for any product you click on.

00:35:27.373 --> 00:35:32.996
It shows you the price point today and what it was last week and what it was the week before and the week before.

00:35:32.996 --> 00:35:36.554
You can see this plotted graph underneath and you can decide.

00:35:36.574 --> 00:35:39.019
You know what laundry detergent this week?

00:35:39.019 --> 00:35:40.023
It's 50 off.

00:35:40.023 --> 00:35:49.298
I'm going to buy two lots of it you know to do the next two weeks, because I know it'll be on discount again in two weeks time based on the pattern that is shown in the past.

00:35:49.719 --> 00:36:00.690
Now you add that into the agent system an agent that can extract, that can extrapolate and interpret that data without the mental load of sitting around pondering you?

00:36:00.710 --> 00:36:02.259
don't have to worry about it, it just handles it for you.

00:36:02.259 --> 00:36:03.967
Now here's another thing.

00:36:03.967 --> 00:36:06.757
Another Australian TikToker I follow.

00:36:06.757 --> 00:36:11.530
He is, and he's always helping you bring your household expenditure down.

00:36:11.530 --> 00:36:13.034
And what he noticed?

00:36:13.034 --> 00:36:24.760
He went down to get to the chemist to buy five products from the chemist I forget what they were and he has this app that scans a barcode and it said, oh, it's this price at this store, this store, this store.

00:36:24.760 --> 00:36:30.137
And once again he found a 50% variance for any scanned product.

00:36:31.039 --> 00:36:31.601
Fascinating.

00:36:31.730 --> 00:36:32.132
And he goes.

00:36:32.132 --> 00:36:34.273
This is amazing, like you.

00:36:34.273 --> 00:36:36.119
Now you hand that to an agent, right?

00:36:36.119 --> 00:36:37.451
Who's going to go?

00:36:37.451 --> 00:36:40.293
You know what I can order it, including the courier fees, etc.

00:36:40.293 --> 00:36:42.635
It's going to be cheaper to order it from.

00:36:42.635 --> 00:36:59.547
You know the online store of X, y or Z man, I tell you, I reckon e-commerce is going to go through the roof online buying of anything because it's going to find the best optimized deal price point for you and it'll purchase from there rather than your store of choice.

00:37:03.429 --> 00:37:03.469
Mm.

00:37:03.469 --> 00:37:05.938
Hmm, yeah, point for you, and it'll purchase from there rather than your store of choice.

00:37:05.938 --> 00:37:06.400
Yeah, now, now let's um.

00:37:06.400 --> 00:37:07.405
So I think in a future episode I would sorry.

00:37:07.405 --> 00:37:07.927
What were you saying?

00:37:07.947 --> 00:37:08.027
mark.

00:37:08.027 --> 00:37:16.831
I was gonna say customer loyalty will will be dead going forward, it'll kill customer loyalty um or customer perceived loyalty.

00:37:16.831 --> 00:37:23.101
You know that these programs try to induce you into yeah well, I think it'll kill.

00:37:23.360 --> 00:37:24.543
I think it's likely to.

00:37:24.543 --> 00:37:25.644
I don't think it'll kill.

00:37:25.644 --> 00:37:43.112
I know you love to say that things are dead.

00:37:43.112 --> 00:37:45.054
Customer loyalty programs probably across the board, but no more.

00:37:45.054 --> 00:37:48.260
No no industry more so than in highly commoditized industries right Like air travel.

00:37:48.280 --> 00:37:56.623
I actually don't think that air travel is nearly is is as highly commoditized, because air travel is air travel is is.

00:37:56.623 --> 00:37:59.556
First of all, it's very specific.

00:37:59.556 --> 00:38:06.876
It can be very specific to where you live, right to specific markets, totally to there.

00:38:06.876 --> 00:38:13.777
Really, there really really are benefits as someone who flies a lot, there are in, you know, huge.

00:38:13.777 --> 00:38:23.855
There can be huge benefits to sort of concentrating your miles, earning your points, earning in in an airline alliance, for example but what?

00:38:23.896 --> 00:38:43.456
what happened if your agent knew that, knew all the data of all the others, knew everything about how you could optimize a flight and like because I see on you know some folks online they know that whatever airline sometimes you can buy points and stuff right and and status credits and stuff like that, and it knows the price, like it could work all that stuff out and go.

00:38:43.456 --> 00:38:46.295
Hey, by the way, this month they've got this deal.

00:38:46.295 --> 00:38:53.155
If you go there, it means the next five flights you're going to do you're going to be business all the way because it's going to be so economical to do it.

00:38:53.155 --> 00:38:54.659
I think it's amazing.

00:38:54.659 --> 00:38:56.775
I can't wait to you know.

00:39:00.289 --> 00:39:12.244
So it's going to cause, it's going, this is going to cause, I think, significant changes in the patterns of customer loyalty and in the way that brands attract and retain that loyalty over time.

00:39:12.244 --> 00:39:14.829
The more commoditized it is Right.

00:39:14.829 --> 00:39:21.981
So like if I can buy a bottle of my favorite dish washing detergent and.

00:39:22.563 --> 00:39:26.731
I live in the UK, so at Tesco or at Sainsbury's, you know what I mean Like, like.

00:39:26.731 --> 00:39:31.643
So those really where the same brand is available for multiple retail providers.

00:39:31.643 --> 00:39:44.956
That's, I think you loyalty to those retail, those retail outlets, is out the window, to those retail, those retail outlets, is out the window.

00:39:44.956 --> 00:39:46.599
You know, I think that you know, probably you've got.

00:39:46.599 --> 00:39:48.362
You know, I really love the clothing brand Ascot.

00:39:48.362 --> 00:39:51.175
They're a Swedish brand and I just love the t-shirts.

00:39:51.175 --> 00:39:55.231
I love the t-shirts, I love the, the, the button down shirt, I just love them, right.

00:39:55.231 --> 00:39:58.838
But you only get them from that store, my Allbirds.

00:39:58.838 --> 00:40:01.590
You only get Allbirds shoes from Allbirds, right.

00:40:01.590 --> 00:40:13.250
So those sorts of those integrated, those integrated manufacturing, distribution and retail outlets are probably going to fare better.

00:40:13.250 --> 00:40:18.280
But you know they're, they're under, they're under pressure right now for a whole host of other reasons.

00:40:18.280 --> 00:40:18.822
So but?

00:40:18.902 --> 00:40:43.681
but you think, imagine your agent knows the style that you're into and it discovers because it's out there researching all the time the optimal shirt that you like and it finds five other brands yeah that allow you a bit of um uh diversity in in what you're wearing, but they're on point as to what you like to wear and you would never have gone out of your way to look at that because you're wearing, but they're on point as to what you like to wear and you would never have gone out of your way to look at that because you're happy with where you are.

00:40:43.681 --> 00:40:46.885
But all of a sudden now you're exposed to new and oh my gosh.

00:40:46.885 --> 00:40:58.434
So I think, even like, take Allbirds as an example, you know what, if there are three others that kind of suited a similar style, but understood your personality type, I just think it's still gonna open up.

00:40:58.434 --> 00:41:19.800
I think the one thing that will where loyalty will still stay is a good coffee shop and a good um, a you know a great, a great restaurant, a great wine bar, a great you know, cocktail, like I don't think those will go away in a hurry, like I think these other things I think will transform very, very rapidly.

00:41:19.800 --> 00:41:35.443
The other thing I think in 2025 people are going to realize how much ai, the need for them to get on board and understand and work with ai, as in much more in the population broadly outside of tech yeah, yeah, you know so.

00:41:35.623 --> 00:41:45.789
So I and I know we're, we're almost at time here, but there's a few themes, topics that I think would be really interesting to pick up on in a future episode.

00:41:45.789 --> 00:42:00.286
So one is and I alluded to this earlier your and my, the different ways and the different perspectives that you and I have on how we've integrated, or maybe not integrated, ai into our daily life and work.

00:42:00.286 --> 00:42:01.670
I think that's a really interesting topic.

00:42:01.670 --> 00:42:15.266
I would like to pick up on how this you know, agentification and this idea of agents has been rolled out and received over the last month, couple of months or so.

00:42:15.266 --> 00:42:23.282
I think that's interesting, or so I think that's interesting, and I also think that it's really interesting to think a little bit about how you know what this means.

00:42:23.282 --> 00:42:32.101
You know what agents mean from a regulatory perspective and how governments are going to deal with accountability.

00:42:32.523 --> 00:42:39.556
And I'm thinking about that example that you cited I forget who it was where you know, basically, the, the, what was it?

00:42:39.556 --> 00:42:39.797
The?

00:42:39.797 --> 00:42:53.518
The, the telco had instructed um, um had had instructed um, uh, the you know folks to to confuse their customers?

00:42:53.778 --> 00:42:55.585
Yeah, cause they could drive more revenue.

00:42:55.585 --> 00:43:00.019
Yeah, it's going to be interesting, cause, I mean, they were litigated against and that was way before AI.

00:43:00.019 --> 00:43:01.476
This is like 10 years ago.

00:43:01.476 --> 00:43:03.795
This type of thing happened.

00:43:03.795 --> 00:43:05.114
So yeah, interesting times, mate.

00:43:05.114 --> 00:43:06.119
It's been great talking.

00:43:06.119 --> 00:43:10.255
We're definitely over time, but until next time, ciao, ciao.

00:43:10.715 --> 00:43:11.016
See ya.

00:43:11.777 --> 00:43:13.563
Thanks for tuning in to the Ecosystem Show.

00:43:13.563 --> 00:43:19.737
We hope you found today's discussion insightful and thought-provoking, and maybe you had a laugh or two.

00:43:19.737 --> 00:43:25.704
Remember your feedback and challenges help us all grow, so don't hesitate to share your perspective.

00:43:25.704 --> 00:43:31.750
Stay connected with us for more innovative ideas and strategies to enhance your software estate.

00:43:31.750 --> 00:43:35.855
Until next time, keep pushing the boundaries and creating value.

00:43:35.855 --> 00:43:37.416
See you on the next episode.

AI Agents and the Future of Customer Service

Listen On

Recent Episodes

AI Agents and the Future of Customer Service

Join over 1500 Power Platform & Dynamics 365 pros for a quick, fun, and insightful 4-minute read each week!

Subscribe to Tech Trailblazer newsletter

Listen On

Recent Episodes