WEBVTT
00:00:00.000 --> 00:00:11.039
I'm Steve Harley, and today on Future Ready with Beckla, I'll be sitting down with Nico Charitan, Microsoft's Director of AI and Business Solutions for UK and I.
00:00:13.519 --> 00:00:18.879
We're going to be speaking about the challenges that businesses face in adopting generative AI technologies.
00:00:19.039 --> 00:00:26.000
We're going to be discussing the differences between Microsoft's co-pilot chat and co-pilot for Microsoft 365.
00:00:26.559 --> 00:00:34.960
We're also going to be discussing the mindset shifts that business leaders need to make in order to capture the value and promise of AI in their organisations.
00:00:35.200 --> 00:00:37.280
I found it a really fascinating discussion.
00:00:37.359 --> 00:00:39.119
I'm sure that you will too.
00:00:42.320 --> 00:00:44.240
Welcome to Future Ready with Beckler.
00:00:44.399 --> 00:00:52.240
I'm Steve Harley, and I'm joined today by Nico Charitan, Microsoft's Director of AI Business Solutions for UK and Ireland.
00:00:52.399 --> 00:00:53.119
Welcome, Nico.
00:00:53.359 --> 00:00:53.759
How are you?
00:00:54.000 --> 00:00:54.719
Thank you for having me.
00:00:54.880 --> 00:00:55.439
Very well.
00:00:55.679 --> 00:00:56.320
Brilliant.
00:00:56.479 --> 00:01:04.159
Well, we're here to talk today about AI and Microsoft's Copilot, which we're moving into wave free now.
00:01:04.400 --> 00:01:05.680
The momentum's been unreal.
00:01:05.760 --> 00:01:06.560
What's that been like?
00:01:06.879 --> 00:01:07.280
Incredible.
00:01:07.359 --> 00:01:08.079
Yeah, definitely.
00:01:08.159 --> 00:01:12.719
This technology since Generative AI showed up with Chat GPT.
00:01:12.959 --> 00:01:19.599
We've seen an amazing adoption of this technology and the fastest technology adoption in history, which is incredible.
00:01:19.680 --> 00:01:27.439
And I feel like we went from that moment where it was a bit of a novelty when people started to think, oh, this is amazing.
00:01:27.680 --> 00:01:30.480
I can get all these answers, but still hallucinated.
00:01:30.799 --> 00:01:33.040
So it was more of a gimmicky thing.
00:01:33.359 --> 00:01:40.959
We went through a phase of habit building because the models evolved, the quality of the output improved, and people started to embrace it in everyday life.
00:01:41.519 --> 00:01:45.760
And now we are getting into a phase where it's becoming agentic.
00:01:45.840 --> 00:01:48.719
So it's taking action and gets things done.
00:01:48.959 --> 00:01:53.120
That is super exciting because actual promise of AI is coming to fruition.
00:01:53.680 --> 00:02:02.879
I definitely, you know, if you look at the tools that have, you know, Microsoft have brought to market and others, but but the power of them is just unreal.
00:02:03.120 --> 00:02:08.000
What's that been like trying to kind of map out and navigate that go-to-market journey?
00:02:08.159 --> 00:02:11.120
What does the portfolio look like today for Microsoft?
00:02:11.439 --> 00:02:13.360
Yeah, it's uh it's it's been a journey.
00:02:13.520 --> 00:02:14.080
It's been a journey.
00:02:14.159 --> 00:02:21.599
We've been through a lot of change, but I feel like we have now a really good solid portfolio with mainly two offerings that are very clear.
00:02:21.759 --> 00:02:23.199
The first one is Copilot Chat.
00:02:23.599 --> 00:02:27.039
That is the one that is embedded into Microsoft 365 solutions today.
00:02:27.120 --> 00:02:30.240
So if you have one, you get these at no additional cost.
00:02:30.400 --> 00:02:32.319
This is more of a conversational AI.
00:02:32.479 --> 00:02:42.479
You can ask questions, browse the internet, gives you answers, but it's with that enterprise grade security, and it gets to interact with the documents that you're working on, which is fantastic.
00:02:42.639 --> 00:02:45.520
Now that's the entry point, I would say.
00:02:45.599 --> 00:02:50.159
And it's a great tool for learning, uh learning how to prompt and how to interact with AI.
00:02:50.879 --> 00:03:09.280
But then the real powerhouse is Microsoft 365 Copilot, where you get the full power of a tool that really gets to know you, understands your context, your data, everything that you have access to, and the way you do things to really take action on your behalf and give you something very polished.
00:03:09.439 --> 00:03:15.599
And that is that is the one that has uh additional licensing, but uh is the one that unlocks the full power of AI.
00:03:15.919 --> 00:03:20.159
It's really uh it's a really powerful tool, uh co-pilot with Microsoft 365.
00:03:20.240 --> 00:03:30.400
That that uh ability to be able to interact with your own data and ask it questions and it understands your context really does you know change the way that you utilize it.
00:03:30.560 --> 00:03:39.439
Um where do you see um organizations, um particularly SMEs and then the mid-market, in terms of uh adoption on that journey?
00:03:39.520 --> 00:03:50.159
Because you've got chat that's delivering out that AI literacy piece, but but but what's that uptake been like for uh for the the the full fat version, as it were?
00:03:50.479 --> 00:04:00.719
So, firstly, the adoption of AI is is kind of it's very interesting because a lot of businesses go through that intentionally, but a lot of AI adoption is happening unintentionally.
00:04:00.800 --> 00:04:03.599
And that shadow AI, it's a real thing.
00:04:03.680 --> 00:04:04.400
It's a real thing.
00:04:04.639 --> 00:04:11.840
When we look at Copilot, I kind of disclose numbers specifically, but the adoption of chat is really, really strong.
00:04:12.000 --> 00:04:18.240
Um, we see that our offering is AI for work, and this is a big differentiator, right?
00:04:18.399 --> 00:04:36.959
Um and and then we start to see those projects starting to come along with the paid version because very quickly you hit a wall once you get the habit built into working with this assistant that gives you a lot of information and gets things for you, you want to explore what's next.
00:04:37.040 --> 00:04:42.720
And and that is when you say, like, can can this tool actually do something extra for me?
00:04:43.199 --> 00:04:45.600
And that's when you get you you need more.
00:04:45.680 --> 00:04:50.800
And there are lots of differences between chat and pay that make it very rich and people seek.
00:04:51.120 --> 00:05:01.199
So, Nico, there's I think it's fair to say that there is uh a level of of confusion around you know the difference between copilot chat and copilot for Microsoft 365.
00:05:02.160 --> 00:05:06.800
What are the fundamental differences and and when do people choose one over the other?
00:05:07.360 --> 00:05:10.959
There are a few things that are that are different between the two.
00:05:11.120 --> 00:05:34.160
Firstly, is work IQ is a critical thing because we have in Microsoft 365 copilot a way to uh get compiler to know you, know your data, know your files, know the people you work with, know your context, the projects you're involved in, has access to your chats, your emails, and everything that you have within your business to make a lot more sense.
00:05:34.319 --> 00:05:43.600
So it is the context of the data and uh who you are, how you work, and the skills, access to skills and tools to get things done.
00:05:43.839 --> 00:05:49.519
Then you have the elements of the agents when copilot gets to take action on your behalf.
00:05:49.680 --> 00:06:15.199
So instead of just getting a summary of a document that you're reading, you get to actually ask copilot to produce a full analysis model for you in Excel, and you can give copilot the agency to go and start playing with that Excel file, creating the formulas, creating the pivot tables, creating the graphs and charts and different tasks and so on, give you whatever analysis insights has found.
00:06:15.600 --> 00:06:20.480
But at the end of that task, you have a full working document that you can use.
00:06:20.720 --> 00:06:23.680
It has it because it has the formulas, it has everything in it.
00:06:23.839 --> 00:06:28.560
And that's the same in PowerPoint, same in Words, same very powerful agency capabilities.
00:06:28.800 --> 00:06:33.360
You have out-of-the-box agents like researcher and analyst.
00:06:33.759 --> 00:06:42.319
They can take hours off your hands by going to the web and documents that you have within your business to pull together a lot of insights and packaging things up.
00:06:42.560 --> 00:06:50.639
And when you start to see these agentic capabilities in action, you start wondering how I can tap into these to automate automate a workflow.
00:06:51.120 --> 00:06:56.959
And and Copilot, the paid version of Copilot has that built-in agentic platform.
00:06:57.120 --> 00:07:05.759
So you can start building things with workflow or agent builder or app builder and copilot studio as well.
00:07:05.839 --> 00:07:19.759
So you can customize where these agents get information from, the type of responses they give, and you can start using it for answering HR questions or helping you with specific um process that you go through every week, etc.
00:07:20.319 --> 00:07:22.720
Um, and and finally the model choice.
00:07:22.879 --> 00:07:34.480
The fact that you have access to the latest technology from different AI labs to pick from, that's also something that users tend to love because you see these models taking advantage from each other.
00:07:34.879 --> 00:07:35.040
Yeah.
00:07:35.199 --> 00:07:42.639
So I mean, I I think the I suppose the the big thing that I'm I'm really interested in is how businesses are actually going to extract the value from it.
00:07:42.720 --> 00:07:51.199
Because you you speak a lot a bit about that that habit forming and that learning and that AI literacy, where we're talking about chat and understanding, you know, how do I prompt?
00:07:51.279 --> 00:07:53.360
How do I get a get a better result out?
00:07:53.600 --> 00:08:04.000
And then you've got that tool set we've co-pilot, Microsoft 365 co-pilot, um, where you've you you you you've kind of got to extract that value for the business.
00:08:04.079 --> 00:08:06.240
You know, how do we get that through to the bottom line?
00:08:06.319 --> 00:08:09.600
You know, that that feels like a challenge for businesses at the moment.
00:08:09.839 --> 00:08:10.720
Yeah, absolutely.
00:08:10.879 --> 00:08:16.560
And we see a lot of businesses, there are there are every business is different in the way they approach AI.
00:08:16.720 --> 00:08:17.839
Every project is different.
00:08:17.920 --> 00:08:19.199
It's very heterogeneous.
00:08:19.279 --> 00:08:26.240
But there are some patterns and we see where they become a blocker or a bottleneck, and it depends on where you are in the journey.
00:08:26.480 --> 00:08:29.680
The first VCs, those are in the early stages.
00:08:29.839 --> 00:08:36.960
They the business leaders understand that they need to do something with AI, they want to do something with AI, but they're overwhelmed with the information.
00:08:37.120 --> 00:08:38.000
They don't know where to start.
00:08:38.159 --> 00:08:54.240
So they're failing at that AI strategy, and that results in user-led BYO AI, shadow AI, and it's non-compliant, puts a lot of risk into businesses, and we get in the news all the time things that leak out unintendedly.
00:08:54.480 --> 00:09:19.279
The second pattern is where you have some people that jumped into the AI bandwagon, they started throwing uh projects at the tech, they love the tech, they got very passionate about it, and they hit a roadblock when they try to scale some of those experiments they've been working on because they quickly realize that their data is not in order, that the security is not what needs to be, the compliance is a big question mark.
00:09:19.600 --> 00:09:24.559
So they really quickly get to the point where, like, hey, we're not ready for it.
00:09:24.720 --> 00:09:28.399
And and they need to go back and that stalls the project.
00:09:28.639 --> 00:09:39.440
And then the third pattern that I see a lot is when those things have been done, maybe the business leaders have a strategy, they have been working on the compliance and security aspects of it.
00:09:39.600 --> 00:09:42.399
They didn't focus on the people, they didn't focus on the culture.
00:09:42.639 --> 00:09:50.240
And that is where the tools are there, they are very excited, but adoption is very low and it's not cutting through.
00:09:50.399 --> 00:09:55.440
And they start wondering what happened, and they underestimate that changed management in their people.
00:09:55.919 --> 00:09:56.960
That's really interesting.
00:09:57.279 --> 00:10:02.399
A lot of uh I've done some research in the background as part of some of my studies for my MBA course.
00:10:02.480 --> 00:10:24.559
And a lot of what I found when when you speak to to small businesses is that that that there is that loop, that there's that requirement for leadership to um make it make sense first so that people can understand what it is for, what's the purpose, why are we doing this, and how are we going to apply it so that they can kind of align what they do to generating that value.
00:10:24.799 --> 00:10:29.759
And then that's an interpretation piece then by by the organization, that's that cultural change.
00:10:29.840 --> 00:10:31.120
You know, what does it mean here?
00:10:31.279 --> 00:10:32.639
How do I utilize that?
00:10:32.799 --> 00:10:36.799
And then that kind of goes into that experimentation and that readiness and that development.
00:10:36.960 --> 00:10:40.320
And and what I've seen is that it that's that's a loop.
00:10:40.480 --> 00:10:45.759
And I think there's a there's a big uh there's a tendency to to look within the industry.
00:10:45.840 --> 00:10:54.639
We go, all right, well, we we do we have an AI strategy and then we we we fix governance and then we we'll do some training, and then everyone will adopt AI and we'll all get value.
00:10:54.799 --> 00:10:57.840
But that's not what's happening because this technology is changing so much.
00:10:57.919 --> 00:10:58.879
Are you seeing that as well?
00:10:59.120 --> 00:11:01.759
100% and I we're seeding waves.
00:11:02.000 --> 00:11:08.799
So this is why I usually recommend that they build a strategy that is firstly based on outcomes.
00:11:08.960 --> 00:11:09.919
What do you want to achieve?
00:11:10.159 --> 00:11:12.879
So, what do you want technology for, right?
00:11:13.039 --> 00:11:17.919
Do you want to reduce the backlog of uh service, customer service support tickets?
00:11:18.159 --> 00:11:21.039
Do you want to improve the accuracy of your forecasting?
00:11:21.279 --> 00:11:27.679
Do you want to uh speed up the delivery of those quotes to your customers on the projects that you are servicing?
00:11:27.840 --> 00:11:31.279
So all of those things need to be put at the top.
00:11:31.440 --> 00:11:36.000
It is a business strategy that uses AI to get there.
00:11:36.720 --> 00:11:37.600
That's the first thing.
00:11:37.759 --> 00:11:48.960
But at the same time, as you say, it's like a multifaceted approach to this strategy because you need to be thinking about that strategy top-down with vision, what do you want to achieve, leadership sponsorship.
00:11:49.120 --> 00:11:52.080
You have to have, you know, those measures of success.
00:11:52.240 --> 00:12:04.399
But at the same time, in parallel, you need to have that data and security, those projects that you've been postponing for years, that your IT leader were telling you, hey, we need to figure out this, and it was not a real problem.
00:12:04.720 --> 00:12:06.879
Now AI gives access to that information.
00:12:06.960 --> 00:12:10.960
You need to you need to bump those projects up and actually get them done.
00:12:11.120 --> 00:12:11.360
Yeah.
00:12:11.600 --> 00:12:15.279
And then you need to spackle all that, you know, culture piece.
00:12:15.360 --> 00:12:21.120
How do we get AI, entry-level AI, to the hands of our employees?
00:12:21.200 --> 00:12:23.120
Uh, this is where compiler chart comes in.
00:12:23.200 --> 00:12:33.200
You know, it's safe, secure, it doesn't tap into all this knowledge yet, but it helps them get that habit building, that prompt mastering, and all of that stuff.
00:12:33.360 --> 00:12:39.120
And then the experimentation that you talked about, once you have all of these things, or in parallel, I would say, start small.
00:12:39.440 --> 00:12:46.080
Understand what are the use cases by function, by role that you could tap into to realize value quickly.
00:12:46.240 --> 00:13:01.360
And this is where you you really see that people get excited because they identify the challenge, they fixed it or they solved it, they improved it with AI, and that triggers their imagination on oh my god, we can do all of this now, and that spreads out.
00:13:01.519 --> 00:13:05.759
So start small, try fast, and scale quickly.
00:13:08.000 --> 00:13:11.120
I mean, I recognize so much of what you say there, Nico.
00:13:11.360 --> 00:13:27.120
When you're when you're talking about the um the data and the the preparation for the technology, I mean, how many organizations took a Microsoft uh uh file server and dumped the data into SharePoint Online and didn't didn't know of the none of the compliance and none of the revisiting.
00:13:27.360 --> 00:13:39.279
We see that when we start to look at Copilot and all of a sudden, because it's powerful, because it can understand so much more and it's easier to surface something that's overshared, there's that compliance piece that people need to go through first.
00:13:39.360 --> 00:13:39.600
Yeah.
00:13:39.919 --> 00:13:42.639
And certainly we see that as a stage.
00:13:42.879 --> 00:13:50.720
But we also see people perhaps hesitant around which technology to adopt and when, um, particularly in a field that's moving so fast.
00:13:50.960 --> 00:13:57.360
And and we look at you know the the rise of Amthropic and and and and the announcement recently of Microsoft co-work.
00:13:57.519 --> 00:13:58.960
How's Microsoft approaching that?
00:13:59.120 --> 00:14:00.879
What's the message to people and leaders?
00:14:01.039 --> 00:14:04.320
You know, should they uh there's a temptation to weigh?
00:14:04.399 --> 00:14:05.759
Is that the right thing?
00:14:06.159 --> 00:14:08.159
Um, well, our strategy is very clear.
00:14:08.399 --> 00:14:14.639
We see in the market we have these AI labs that are leapfrogging each other with the latest innovations in their models, right?
00:14:14.960 --> 00:14:19.120
For us, the strategy is clear is multi-model is the way to go.
00:14:19.279 --> 00:14:31.919
We are a platform, and the way our engineering team has structured co-pilot is very smart because it operates independently of the model, which gives us the ability to switch models on the day.
00:14:32.080 --> 00:14:38.480
For example, the last model that OpenAI released to market, copilot embraced that model on the same day.
00:14:38.799 --> 00:14:41.759
And because of that, now we can also switch to different vendors.
00:14:41.919 --> 00:14:45.519
And now the multi-model strategy gives you choice.
00:14:45.679 --> 00:14:49.679
And this is one of the key differentiations of Microsoft 365, right?
00:14:49.759 --> 00:14:52.000
Um, copilot, the paid version of it.
00:14:52.240 --> 00:15:04.320
It gives you that model choice, but not only do you have the model choice because you know different different AI labs will have the advantage at a given time, we also can experiment with some extra features because we have multi-models.
00:15:04.399 --> 00:15:22.320
So we just launched um, for example, critique and counsel to two experiences within the researcher agents, where you get one of the models, OpenAI's model, running the research, and Claude from Anthropic challenging that research and trying to validate to make sure that it's sound.
00:15:22.480 --> 00:15:26.720
And the outputs based on on what you get are much better.
00:15:26.879 --> 00:15:30.480
I mean, are the best of their kind because there's different.
00:15:30.639 --> 00:15:43.200
Or you can trigger both models to give the research at the same time and compare side by side and and get, you know, what's the best of both and the human making a decision based on all the information available.
00:15:43.279 --> 00:15:45.200
And you don't need to be switching from one to the other.
00:15:45.360 --> 00:15:49.440
So that strategy of multimodal, that's where Microsoft goes.
00:15:49.600 --> 00:15:54.799
But at the same time, we have the element of we are a platform that you're using to run your business.
00:15:55.039 --> 00:16:03.519
So we are embedding all of this within the portfolio from Microsoft 365, the Office apps, you have Dynamics 365, Power Platform.
00:16:03.600 --> 00:16:10.320
It is embedded on everything that you already have natively, which gives you the ability to do things in the flow of work.
00:16:10.399 --> 00:16:14.240
You're not switching from one to the other, copy-pasting, downloading things.
00:16:14.399 --> 00:16:22.720
And that simplicity is going to become even more visible as you get the autonomy of those agents coming to life.
00:16:22.960 --> 00:16:32.799
And the third piece, and very important as well, is that enterprise creates security and governance and responsibility that you would expect from Microsoft, right?
00:16:32.960 --> 00:16:40.240
We are making sure that whatever copilot has access to is related to what the user has access to.
00:16:40.320 --> 00:16:44.159
So we have that uh level of permission and regularity.
00:16:44.240 --> 00:16:46.960
So you won't see the information you're not meant to see.
00:16:47.120 --> 00:16:52.080
Um, and also keeping the information of your business within the boundaries of your business.
00:16:52.320 --> 00:16:53.360
Very important, right?
00:16:53.519 --> 00:17:00.639
We see a lot of information leaking because employees are using consumer applications for getting work done.
00:17:00.879 --> 00:17:10.880
And that's something that with Microsoft 365 compiler and compiler chat, even you get that level of protection that your information is still within the boundaries of your tenant, your environment.
00:17:11.440 --> 00:17:13.599
That's really, really uh interesting there.
00:17:13.759 --> 00:17:20.240
You're talking about that change and that that that kind of sub agent, that multi-model approach.
00:17:20.400 --> 00:17:31.359
Um obviously the industry uh is awash with conversations about agentic AI and how that's going to change the structure of organizations.
00:17:31.599 --> 00:17:37.759
Um, I was really interested in uh the release of Agent 365 uh and the announcements there.
00:17:37.920 --> 00:17:42.880
Um how does Microsoft see that organization in the future?
00:17:42.960 --> 00:17:44.559
I think you call it the frontier firm.
00:17:45.440 --> 00:18:08.559
Yeah, so we see a lot of um this journey going into a place where the users that are spending a lot of time today doing things and tasks um on some of the artifacts like office documents and and the likes, they will become more strategic in the way they direct agents to get work done.
00:18:08.880 --> 00:18:19.680
So humans will be focusing on value tasks while agents will handle the process or the execution of some of those tasks, which is really interesting.
00:18:19.839 --> 00:18:28.799
Now, when we start looking at these, you have a lot of personalized agents created by the user or created uh out of the box for the user.
00:18:29.119 --> 00:18:40.400
And then you also have the process-oriented um agents that are coming into orchestrating multiple functions to get some output done on behalf of the company, not a user.
00:18:40.559 --> 00:18:43.440
All of those things will expand, right?
00:18:43.599 --> 00:18:50.480
We will be in a business that is full of digital colleagues or agents that are operating different tasks.
00:18:50.640 --> 00:18:52.160
And you'll have to govern that.
00:18:52.400 --> 00:19:09.039
As today, an IT organization is governing the people, and you have the identities for those people, the permissions for those people, what do they have access to, what do they know about the organization, what is the what is the tooling that they can use to execute different tasks?
00:19:09.200 --> 00:19:10.799
The same thing will happen with agents.
00:19:10.880 --> 00:19:30.960
And this is where Agent 365, soon to be released, comes to life, where you have a single pane of glass, a registry of all your agents, and you can govern them, you can monitor them, you can see the security aspects of it, and and gives you that confidence that things are not going out of control, and you have a single pane of glass for everything.
00:19:31.279 --> 00:19:34.079
I mean, that's that's a real shift that we're facing.
00:19:34.319 --> 00:19:43.440
Organizations being both um human and digital colleagues, as it were, and the management of those agents is is a is a huge change.
00:19:43.599 --> 00:19:53.279
Um, I think there's a there's a challenge that we I still don't think lots of businesses and organizations have have have got the handle of getting the most out of humans, let alone digital.
00:19:53.440 --> 00:20:01.839
So having tools like Agent 365 to be able to start that journey of understanding what their organization looks like and how they can manage it becomes really, really important.
00:20:02.640 --> 00:20:23.279
Uh if you were advising a CIO or a CEO today, what is the most important shift that people need to make uh in their mindset and their approach to make sure that they're going to you know realize the advantages of tools like Copilot and other AI technologies and drive value for their companies?
00:20:23.680 --> 00:20:30.799
I think specifically for adoption, firstly, you need to understand that you have a heterogeneous workforce.
00:20:31.039 --> 00:20:37.359
And there's research actually that talks about the different ways that different generations are using it.
00:20:37.920 --> 00:20:44.880
You have the boomers using it only if it is embedded within things that they trust.
00:20:45.119 --> 00:20:54.000
You have millennials being very pragmatic with it and expecting answers, so very much a you know finding information, retrieval.
00:20:54.160 --> 00:21:11.359
Um, you have millennials that are a lot more um used to planning things with agents, uh, but very task-oriented, and then you have generations uh after that, Gen Z and so on, that they are using as a life coach essentially.
00:21:11.519 --> 00:21:14.319
They do everything with AI first and different mentality.
00:21:14.559 --> 00:21:22.240
So understanding that you have to influence humans, that change management is critical.
00:21:22.799 --> 00:21:23.039
Right?
00:21:23.200 --> 00:21:37.920
So, how do you put in place a journey for them to understand the technology, to get used to um understanding what the technology can do for them, how to embrace it, what are the outputs that you're expecting?
00:21:38.000 --> 00:21:51.519
That vision we talked about before, super important because there is a real fear that AI will take my job instead of AI will remove the drudgery of my daily work so I can focus on value-added tasks.
00:21:51.759 --> 00:21:59.759
Like I see a lot of companies that started earlier in the journey trying to get the ROI by cutting costs and replacing humans.
00:22:00.000 --> 00:22:07.839
And arguably they're coming back into the oh, we failed because we should have thought about this differently.
00:22:08.000 --> 00:22:16.799
We should unlock growth, creativity, innovation with the use of AI to maintain our teams, but grow as a business exponentially.
00:22:16.960 --> 00:22:18.880
And this is where we see the benefit.
00:22:19.039 --> 00:22:23.519
But you need to have that vision, you need to have that clarity, and you need to bring people on a journey.
00:22:23.599 --> 00:22:23.759
Yeah.
00:22:24.160 --> 00:22:26.000
Change management is critical.
00:22:26.160 --> 00:22:36.400
How to use it, motivate them, show them, help them understand how to unpack use cases and what are the tools and empower people with the right tools.
00:22:36.480 --> 00:22:38.480
And that's another thing that is critical, right?
00:22:38.640 --> 00:22:48.640
Because you have the people at the top have a vision, but the people at the bottom are the ones that know what are the things that are time wasters where the bottlenecks are.
00:22:48.799 --> 00:23:01.519
So when you give them the tools with copilot or even low-code or no code, like copilot studio, that gives them the ability to create workload automation and and boost the productivity.
00:23:01.759 --> 00:23:09.279
And while in the end user productivity space, you might get a 10, 20% productivity gains, which is amazing.
00:23:09.759 --> 00:23:15.759
When you see the 10x return, it's when you automate a workflow and you redesign, you reshape that business process.
00:23:15.920 --> 00:23:18.799
You absolutely change the way you interact with your customers.
00:23:18.960 --> 00:23:21.920
You know, you you bend the curve at innovation and that's super rich.
00:23:22.160 --> 00:23:22.319
Yeah.
00:23:22.480 --> 00:23:23.519
When you change the business model.
00:23:24.000 --> 00:23:24.240
Absolutely.
00:23:24.400 --> 00:23:26.480
And that and that's the the the potential here, isn't it?
00:23:26.720 --> 00:23:26.960
Absolutely.
00:23:27.119 --> 00:23:28.880
But you've got to bring the people along along with you.
00:23:29.119 --> 00:23:29.519
100%.
00:23:29.920 --> 00:23:37.039
Because usually what you have is the automation of a workflow, it is um is going across people, right?
00:23:37.359 --> 00:23:41.440
I always think about that, the doorman replacement analogy, right?
00:23:41.599 --> 00:23:45.839
You you hire a consultant to tell you, hey, how can I cut costs?
00:23:46.079 --> 00:23:53.279
And in you're running a hotel, and they might say, you know what, fire the doorman and replace it for an automated door that opens.
00:23:53.519 --> 00:23:58.960
And suddenly you minimize the value of that human to opening a door.
00:23:59.279 --> 00:24:07.200
Well, that person is not just doing that, it's customer experience, is giving you status as a brand, is giving you security.
00:24:07.359 --> 00:24:12.720
There are so many other things that that person does by open but by opening the door.
00:24:12.880 --> 00:24:17.599
So it's a misconcept of how can I cut the full cost of this.
00:24:17.680 --> 00:24:20.559
So you need to be cautious with that.
00:24:20.799 --> 00:24:24.720
However, there are scenarios where it is very obvious, right?
00:24:25.039 --> 00:24:30.480
Support tickets or time spent on specific things, answering queries, very repetitive tasks.
00:24:30.640 --> 00:24:32.799
That's where a lot of businesses are starting, right?
00:24:33.359 --> 00:24:36.960
And and that that makes a lot of sense because there is a very clear cut.
00:24:37.279 --> 00:24:37.839
Fantastic.
00:24:37.920 --> 00:24:45.359
Um Nico, I could speak for another hour with you about AI adoption and the challenges that we see in the market.
00:24:45.519 --> 00:24:48.480
Um, but but sadly, we're gonna run out run out of time.
00:24:48.640 --> 00:24:50.880
Um, I just want to thank you very much for sharing your insights.
00:24:50.960 --> 00:24:53.200
It's been really, really good fun having you here.
00:24:53.279 --> 00:24:59.039
Um, and uh and I wish you all the best and uh and thank you for joining us here on uh Future Ready with Beckla.
00:24:59.200 --> 00:24:59.839
Thank you so much.
00:00:00.000 --> 00:00:11.039
I'm Steve Harley, and today on Future Ready with Beckla, I'll be sitting down with Nico Charitan, Microsoft's Director of AI and Business Solutions for UK and I.
00:00:13.519 --> 00:00:18.879
We're going to be speaking about the challenges that businesses face in adopting generative AI technologies.
00:00:19.039 --> 00:00:26.000
We're going to be discussing the differences between Microsoft's co-pilot chat and co-pilot for Microsoft 365.
00:00:26.559 --> 00:00:34.960
We're also going to be discussing the mindset shifts that business leaders need to make in order to capture the value and promise of AI in their organisations.
00:00:35.200 --> 00:00:37.280
I found it a really fascinating discussion.
00:00:37.359 --> 00:00:39.119
I'm sure that you will too.
00:00:42.320 --> 00:00:44.240
Welcome to Future Ready with Beckler.
00:00:44.399 --> 00:00:52.240
I'm Steve Harley, and I'm joined today by Nico Charitan, Microsoft's Director of AI Business Solutions for UK and Ireland.
00:00:52.399 --> 00:00:53.119
Welcome, Nico.
00:00:53.359 --> 00:00:53.759
How are you?
00:00:54.000 --> 00:00:54.719
Thank you for having me.
00:00:54.880 --> 00:00:55.439
Very well.
00:00:55.679 --> 00:00:56.320
Brilliant.
00:00:56.479 --> 00:01:04.159
Well, we're here to talk today about AI and Microsoft's Copilot, which we're moving into wave free now.
00:01:04.400 --> 00:01:05.680
The momentum's been unreal.
00:01:05.760 --> 00:01:06.560
What's that been like?
00:01:06.879 --> 00:01:07.280
Incredible.
00:01:07.359 --> 00:01:08.079
Yeah, definitely.
00:01:08.159 --> 00:01:12.719
This technology since Generative AI showed up with Chat GPT.
00:01:12.959 --> 00:01:19.599
We've seen an amazing adoption of this technology and the fastest technology adoption in history, which is incredible.
00:01:19.680 --> 00:01:27.439
And I feel like we went from that moment where it was a bit of a novelty when people started to think, oh, this is amazing.
00:01:27.680 --> 00:01:30.480
I can get all these answers, but still hallucinated.
00:01:30.799 --> 00:01:33.040
So it was more of a gimmicky thing.
00:01:33.359 --> 00:01:40.959
We went through a phase of habit building because the models evolved, the quality of the output improved, and people started to embrace it in everyday life.
00:01:41.519 --> 00:01:45.760
And now we are getting into a phase where it's becoming agentic.
00:01:45.840 --> 00:01:48.719
So it's taking action and gets things done.
00:01:48.959 --> 00:01:53.120
That is super exciting because actual promise of AI is coming to fruition.
00:01:53.680 --> 00:02:02.879
I definitely, you know, if you look at the tools that have, you know, Microsoft have brought to market and others, but but the power of them is just unreal.
00:02:03.120 --> 00:02:08.000
What's that been like trying to kind of map out and navigate that go-to-market journey?
00:02:08.159 --> 00:02:11.120
What does the portfolio look like today for Microsoft?
00:02:11.439 --> 00:02:13.360
Yeah, it's uh it's it's been a journey.
00:02:13.520 --> 00:02:14.080
It's been a journey.
00:02:14.159 --> 00:02:21.599
We've been through a lot of change, but I feel like we have now a really good solid portfolio with mainly two offerings that are very clear.
00:02:21.759 --> 00:02:23.199
The first one is Copilot Chat.
00:02:23.599 --> 00:02:27.039
That is the one that is embedded into Microsoft 365 solutions today.
00:02:27.120 --> 00:02:30.240
So if you have one, you get these at no additional cost.
00:02:30.400 --> 00:02:32.319
This is more of a conversational AI.
00:02:32.479 --> 00:02:42.479
You can ask questions, browse the internet, gives you answers, but it's with that enterprise grade security, and it gets to interact with the documents that you're working on, which is fantastic.
00:02:42.639 --> 00:02:45.520
Now that's the entry point, I would say.
00:02:45.599 --> 00:02:50.159
And it's a great tool for learning, uh learning how to prompt and how to interact with AI.
00:02:50.879 --> 00:03:09.280
But then the real powerhouse is Microsoft 365 Copilot, where you get the full power of a tool that really gets to know you, understands your context, your data, everything that you have access to, and the way you do things to really take action on your behalf and give you something very polished.
00:03:09.439 --> 00:03:15.599
And that is that is the one that has uh additional licensing, but uh is the one that unlocks the full power of AI.
00:03:15.919 --> 00:03:20.159
It's really uh it's a really powerful tool, uh co-pilot with Microsoft 365.
00:03:20.240 --> 00:03:30.400
That that uh ability to be able to interact with your own data and ask it questions and it understands your context really does you know change the way that you utilize it.
00:03:30.560 --> 00:03:39.439
Um where do you see um organizations, um particularly SMEs and then the mid-market, in terms of uh adoption on that journey?
00:03:39.520 --> 00:03:50.159
Because you've got chat that's delivering out that AI literacy piece, but but but what's that uptake been like for uh for the the the full fat version, as it were?
00:03:50.479 --> 00:04:00.719
So, firstly, the adoption of AI is is kind of it's very interesting because a lot of businesses go through that intentionally, but a lot of AI adoption is happening unintentionally.
00:04:00.800 --> 00:04:03.599
And that shadow AI, it's a real thing.
00:04:03.680 --> 00:04:04.400
It's a real thing.
00:04:04.639 --> 00:04:11.840
When we look at Copilot, I kind of disclose numbers specifically, but the adoption of chat is really, really strong.
00:04:12.000 --> 00:04:18.240
Um, we see that our offering is AI for work, and this is a big differentiator, right?
00:04:18.399 --> 00:04:36.959
Um and and then we start to see those projects starting to come along with the paid version because very quickly you hit a wall once you get the habit built into working with this assistant that gives you a lot of information and gets things for you, you want to explore what's next.
00:04:37.040 --> 00:04:42.720
And and that is when you say, like, can can this tool actually do something extra for me?
00:04:43.199 --> 00:04:45.600
And that's when you get you you need more.
00:04:45.680 --> 00:04:50.800
And there are lots of differences between chat and pay that make it very rich and people seek.
00:04:51.120 --> 00:05:01.199
So, Nico, there's I think it's fair to say that there is uh a level of of confusion around you know the difference between copilot chat and copilot for Microsoft 365.
00:05:02.160 --> 00:05:06.800
What are the fundamental differences and and when do people choose one over the other?
00:05:07.360 --> 00:05:10.959
There are a few things that are that are different between the two.
00:05:11.120 --> 00:05:34.160
Firstly, is work IQ is a critical thing because we have in Microsoft 365 copilot a way to uh get compiler to know you, know your data, know your files, know the people you work with, know your context, the projects you're involved in, has access to your chats, your emails, and everything that you have within your business to make a lot more sense.
00:05:34.319 --> 00:05:43.600
So it is the context of the data and uh who you are, how you work, and the skills, access to skills and tools to get things done.
00:05:43.839 --> 00:05:49.519
Then you have the elements of the agents when copilot gets to take action on your behalf.
00:05:49.680 --> 00:06:15.199
So instead of just getting a summary of a document that you're reading, you get to actually ask copilot to produce a full analysis model for you in Excel, and you can give copilot the agency to go and start playing with that Excel file, creating the formulas, creating the pivot tables, creating the graphs and charts and different tasks and so on, give you whatever analysis insights has found.
00:06:15.600 --> 00:06:20.480
But at the end of that task, you have a full working document that you can use.
00:06:20.720 --> 00:06:23.680
It has it because it has the formulas, it has everything in it.
00:06:23.839 --> 00:06:28.560
And that's the same in PowerPoint, same in Words, same very powerful agency capabilities.
00:06:28.800 --> 00:06:33.360
You have out-of-the-box agents like researcher and analyst.
00:06:33.759 --> 00:06:42.319
They can take hours off your hands by going to the web and documents that you have within your business to pull together a lot of insights and packaging things up.
00:06:42.560 --> 00:06:50.639
And when you start to see these agentic capabilities in action, you start wondering how I can tap into these to automate automate a workflow.
00:06:51.120 --> 00:06:56.959
And and Copilot, the paid version of Copilot has that built-in agentic platform.
00:06:57.120 --> 00:07:05.759
So you can start building things with workflow or agent builder or app builder and copilot studio as well.
00:07:05.839 --> 00:07:19.759
So you can customize where these agents get information from, the type of responses they give, and you can start using it for answering HR questions or helping you with specific um process that you go through every week, etc.
00:07:20.319 --> 00:07:22.720
Um, and and finally the model choice.
00:07:22.879 --> 00:07:34.480
The fact that you have access to the latest technology from different AI labs to pick from, that's also something that users tend to love because you see these models taking advantage from each other.
00:07:34.879 --> 00:07:35.040
Yeah.
00:07:35.199 --> 00:07:42.639
So I mean, I I think the I suppose the the big thing that I'm I'm really interested in is how businesses are actually going to extract the value from it.
00:07:42.720 --> 00:07:51.199
Because you you speak a lot a bit about that that habit forming and that learning and that AI literacy, where we're talking about chat and understanding, you know, how do I prompt?
00:07:51.279 --> 00:07:53.360
How do I get a get a better result out?
00:07:53.600 --> 00:08:04.000
And then you've got that tool set we've co-pilot, Microsoft 365 co-pilot, um, where you've you you you you've kind of got to extract that value for the business.
00:08:04.079 --> 00:08:06.240
You know, how do we get that through to the bottom line?
00:08:06.319 --> 00:08:09.600
You know, that that feels like a challenge for businesses at the moment.
00:08:09.839 --> 00:08:10.720
Yeah, absolutely.
00:08:10.879 --> 00:08:16.560
And we see a lot of businesses, there are there are every business is different in the way they approach AI.
00:08:16.720 --> 00:08:17.839
Every project is different.
00:08:17.920 --> 00:08:19.199
It's very heterogeneous.
00:08:19.279 --> 00:08:26.240
But there are some patterns and we see where they become a blocker or a bottleneck, and it depends on where you are in the journey.
00:08:26.480 --> 00:08:29.680
The first VCs, those are in the early stages.
00:08:29.839 --> 00:08:36.960
They the business leaders understand that they need to do something with AI, they want to do something with AI, but they're overwhelmed with the information.
00:08:37.120 --> 00:08:38.000
They don't know where to start.
00:08:38.159 --> 00:08:54.240
So they're failing at that AI strategy, and that results in user-led BYO AI, shadow AI, and it's non-compliant, puts a lot of risk into businesses, and we get in the news all the time things that leak out unintendedly.
00:08:54.480 --> 00:09:19.279
The second pattern is where you have some people that jumped into the AI bandwagon, they started throwing uh projects at the tech, they love the tech, they got very passionate about it, and they hit a roadblock when they try to scale some of those experiments they've been working on because they quickly realize that their data is not in order, that the security is not what needs to be, the compliance is a big question mark.
00:09:19.600 --> 00:09:24.559
So they really quickly get to the point where, like, hey, we're not ready for it.
00:09:24.720 --> 00:09:28.399
And and they need to go back and that stalls the project.
00:09:28.639 --> 00:09:39.440
And then the third pattern that I see a lot is when those things have been done, maybe the business leaders have a strategy, they have been working on the compliance and security aspects of it.
00:09:39.600 --> 00:09:42.399
They didn't focus on the people, they didn't focus on the culture.
00:09:42.639 --> 00:09:50.240
And that is where the tools are there, they are very excited, but adoption is very low and it's not cutting through.
00:09:50.399 --> 00:09:55.440
And they start wondering what happened, and they underestimate that changed management in their people.
00:09:55.919 --> 00:09:56.960
That's really interesting.
00:09:57.279 --> 00:10:02.399
A lot of uh I've done some research in the background as part of some of my studies for my MBA course.
00:10:02.480 --> 00:10:24.559
And a lot of what I found when when you speak to to small businesses is that that that there is that loop, that there's that requirement for leadership to um make it make sense first so that people can understand what it is for, what's the purpose, why are we doing this, and how are we going to apply it so that they can kind of align what they do to generating that value.
00:10:24.799 --> 00:10:29.759
And then that's an interpretation piece then by by the organization, that's that cultural change.
00:10:29.840 --> 00:10:31.120
You know, what does it mean here?
00:10:31.279 --> 00:10:32.639
How do I utilize that?
00:10:32.799 --> 00:10:36.799
And then that kind of goes into that experimentation and that readiness and that development.
00:10:36.960 --> 00:10:40.320
And and what I've seen is that it that's that's a loop.
00:10:40.480 --> 00:10:45.759
And I think there's a there's a big uh there's a tendency to to look within the industry.
00:10:45.840 --> 00:10:54.639
We go, all right, well, we we do we have an AI strategy and then we we we fix governance and then we we'll do some training, and then everyone will adopt AI and we'll all get value.
00:10:54.799 --> 00:10:57.840
But that's not what's happening because this technology is changing so much.
00:10:57.919 --> 00:10:58.879
Are you seeing that as well?
00:10:59.120 --> 00:11:01.759
100% and I we're seeding waves.
00:11:02.000 --> 00:11:08.799
So this is why I usually recommend that they build a strategy that is firstly based on outcomes.
00:11:08.960 --> 00:11:09.919
What do you want to achieve?
00:11:10.159 --> 00:11:12.879
So, what do you want technology for, right?
00:11:13.039 --> 00:11:17.919
Do you want to reduce the backlog of uh service, customer service support tickets?
00:11:18.159 --> 00:11:21.039
Do you want to improve the accuracy of your forecasting?
00:11:21.279 --> 00:11:27.679
Do you want to uh speed up the delivery of those quotes to your customers on the projects that you are servicing?
00:11:27.840 --> 00:11:31.279
So all of those things need to be put at the top.
00:11:31.440 --> 00:11:36.000
It is a business strategy that uses AI to get there.
00:11:36.720 --> 00:11:37.600
That's the first thing.
00:11:37.759 --> 00:11:48.960
But at the same time, as you say, it's like a multifaceted approach to this strategy because you need to be thinking about that strategy top-down with vision, what do you want to achieve, leadership sponsorship.
00:11:49.120 --> 00:11:52.080
You have to have, you know, those measures of success.
00:11:52.240 --> 00:12:04.399
But at the same time, in parallel, you need to have that data and security, those projects that you've been postponing for years, that your IT leader were telling you, hey, we need to figure out this, and it was not a real problem.
00:12:04.720 --> 00:12:06.879
Now AI gives access to that information.
00:12:06.960 --> 00:12:10.960
You need to you need to bump those projects up and actually get them done.
00:12:11.120 --> 00:12:11.360
Yeah.
00:12:11.600 --> 00:12:15.279
And then you need to spackle all that, you know, culture piece.
00:12:15.360 --> 00:12:21.120
How do we get AI, entry-level AI, to the hands of our employees?
00:12:21.200 --> 00:12:23.120
Uh, this is where compiler chart comes in.
00:12:23.200 --> 00:12:33.200
You know, it's safe, secure, it doesn't tap into all this knowledge yet, but it helps them get that habit building, that prompt mastering, and all of that stuff.
00:12:33.360 --> 00:12:39.120
And then the experimentation that you talked about, once you have all of these things, or in parallel, I would say, start small.
00:12:39.440 --> 00:12:46.080
Understand what are the use cases by function, by role that you could tap into to realize value quickly.
00:12:46.240 --> 00:13:01.360
And this is where you you really see that people get excited because they identify the challenge, they fixed it or they solved it, they improved it with AI, and that triggers their imagination on oh my god, we can do all of this now, and that spreads out.
00:13:01.519 --> 00:13:05.759
So start small, try fast, and scale quickly.
00:13:08.000 --> 00:13:11.120
I mean, I recognize so much of what you say there, Nico.
00:13:11.360 --> 00:13:27.120
When you're when you're talking about the um the data and the the preparation for the technology, I mean, how many organizations took a Microsoft uh uh file server and dumped the data into SharePoint Online and didn't didn't know of the none of the compliance and none of the revisiting.
00:13:27.360 --> 00:13:39.279
We see that when we start to look at Copilot and all of a sudden, because it's powerful, because it can understand so much more and it's easier to surface something that's overshared, there's that compliance piece that people need to go through first.
00:13:39.360 --> 00:13:39.600
Yeah.
00:13:39.919 --> 00:13:42.639
And certainly we see that as a stage.
00:13:42.879 --> 00:13:50.720
But we also see people perhaps hesitant around which technology to adopt and when, um, particularly in a field that's moving so fast.
00:13:50.960 --> 00:13:57.360
And and we look at you know the the rise of Amthropic and and and and the announcement recently of Microsoft co-work.
00:13:57.519 --> 00:13:58.960
How's Microsoft approaching that?
00:13:59.120 --> 00:14:00.879
What's the message to people and leaders?
00:14:01.039 --> 00:14:04.320
You know, should they uh there's a temptation to weigh?
00:14:04.399 --> 00:14:05.759
Is that the right thing?
00:14:06.159 --> 00:14:08.159
Um, well, our strategy is very clear.
00:14:08.399 --> 00:14:14.639
We see in the market we have these AI labs that are leapfrogging each other with the latest innovations in their models, right?
00:14:14.960 --> 00:14:19.120
For us, the strategy is clear is multi-model is the way to go.
00:14:19.279 --> 00:14:31.919
We are a platform, and the way our engineering team has structured co-pilot is very smart because it operates independently of the model, which gives us the ability to switch models on the day.
00:14:32.080 --> 00:14:38.480
For example, the last model that OpenAI released to market, copilot embraced that model on the same day.
00:14:38.799 --> 00:14:41.759
And because of that, now we can also switch to different vendors.
00:14:41.919 --> 00:14:45.519
And now the multi-model strategy gives you choice.
00:14:45.679 --> 00:14:49.679
And this is one of the key differentiations of Microsoft 365, right?
00:14:49.759 --> 00:14:52.000
Um, copilot, the paid version of it.
00:14:52.240 --> 00:15:04.320
It gives you that model choice, but not only do you have the model choice because you know different different AI labs will have the advantage at a given time, we also can experiment with some extra features because we have multi-models.
00:15:04.399 --> 00:15:22.320
So we just launched um, for example, critique and counsel to two experiences within the researcher agents, where you get one of the models, OpenAI's model, running the research, and Claude from Anthropic challenging that research and trying to validate to make sure that it's sound.
00:15:22.480 --> 00:15:26.720
And the outputs based on on what you get are much better.
00:15:26.879 --> 00:15:30.480
I mean, are the best of their kind because there's different.
00:15:30.639 --> 00:15:43.200
Or you can trigger both models to give the research at the same time and compare side by side and and get, you know, what's the best of both and the human making a decision based on all the information available.
00:15:43.279 --> 00:15:45.200
And you don't need to be switching from one to the other.
00:15:45.360 --> 00:15:49.440
So that strategy of multimodal, that's where Microsoft goes.
00:15:49.600 --> 00:15:54.799
But at the same time, we have the element of we are a platform that you're using to run your business.
00:15:55.039 --> 00:16:03.519
So we are embedding all of this within the portfolio from Microsoft 365, the Office apps, you have Dynamics 365, Power Platform.
00:16:03.600 --> 00:16:10.320
It is embedded on everything that you already have natively, which gives you the ability to do things in the flow of work.
00:16:10.399 --> 00:16:14.240
You're not switching from one to the other, copy-pasting, downloading things.
00:16:14.399 --> 00:16:22.720
And that simplicity is going to become even more visible as you get the autonomy of those agents coming to life.
00:16:22.960 --> 00:16:32.799
And the third piece, and very important as well, is that enterprise creates security and governance and responsibility that you would expect from Microsoft, right?
00:16:32.960 --> 00:16:40.240
We are making sure that whatever copilot has access to is related to what the user has access to.
00:16:40.320 --> 00:16:44.159
So we have that uh level of permission and regularity.
00:16:44.240 --> 00:16:46.960
So you won't see the information you're not meant to see.
00:16:47.120 --> 00:16:52.080
Um, and also keeping the information of your business within the boundaries of your business.
00:16:52.320 --> 00:16:53.360
Very important, right?
00:16:53.519 --> 00:17:00.639
We see a lot of information leaking because employees are using consumer applications for getting work done.
00:17:00.879 --> 00:17:10.880
And that's something that with Microsoft 365 compiler and compiler chat, even you get that level of protection that your information is still within the boundaries of your tenant, your environment.
00:17:11.440 --> 00:17:13.599
That's really, really uh interesting there.
00:17:13.759 --> 00:17:20.240
You're talking about that change and that that that kind of sub agent, that multi-model approach.
00:17:20.400 --> 00:17:31.359
Um obviously the industry uh is awash with conversations about agentic AI and how that's going to change the structure of organizations.
00:17:31.599 --> 00:17:37.759
Um, I was really interested in uh the release of Agent 365 uh and the announcements there.
00:17:37.920 --> 00:17:42.880
Um how does Microsoft see that organization in the future?
00:17:42.960 --> 00:17:44.559
I think you call it the frontier firm.
00:17:45.440 --> 00:18:08.559
Yeah, so we see a lot of um this journey going into a place where the users that are spending a lot of time today doing things and tasks um on some of the artifacts like office documents and and the likes, they will become more strategic in the way they direct agents to get work done.
00:18:08.880 --> 00:18:19.680
So humans will be focusing on value tasks while agents will handle the process or the execution of some of those tasks, which is really interesting.
00:18:19.839 --> 00:18:28.799
Now, when we start looking at these, you have a lot of personalized agents created by the user or created uh out of the box for the user.
00:18:29.119 --> 00:18:40.400
And then you also have the process-oriented um agents that are coming into orchestrating multiple functions to get some output done on behalf of the company, not a user.
00:18:40.559 --> 00:18:43.440
All of those things will expand, right?
00:18:43.599 --> 00:18:50.480
We will be in a business that is full of digital colleagues or agents that are operating different tasks.
00:18:50.640 --> 00:18:52.160
And you'll have to govern that.
00:18:52.400 --> 00:19:09.039
As today, an IT organization is governing the people, and you have the identities for those people, the permissions for those people, what do they have access to, what do they know about the organization, what is the what is the tooling that they can use to execute different tasks?
00:19:09.200 --> 00:19:10.799
The same thing will happen with agents.
00:19:10.880 --> 00:19:30.960
And this is where Agent 365, soon to be released, comes to life, where you have a single pane of glass, a registry of all your agents, and you can govern them, you can monitor them, you can see the security aspects of it, and and gives you that confidence that things are not going out of control, and you have a single pane of glass for everything.
00:19:31.279 --> 00:19:34.079
I mean, that's that's a real shift that we're facing.
00:19:34.319 --> 00:19:43.440
Organizations being both um human and digital colleagues, as it were, and the management of those agents is is a is a huge change.
00:19:43.599 --> 00:19:53.279
Um, I think there's a there's a challenge that we I still don't think lots of businesses and organizations have have have got the handle of getting the most out of humans, let alone digital.
00:19:53.440 --> 00:20:01.839
So having tools like Agent 365 to be able to start that journey of understanding what their organization looks like and how they can manage it becomes really, really important.
00:20:02.640 --> 00:20:23.279
Uh if you were advising a CIO or a CEO today, what is the most important shift that people need to make uh in their mindset and their approach to make sure that they're going to you know realize the advantages of tools like Copilot and other AI technologies and drive value for their companies?
00:20:23.680 --> 00:20:30.799
I think specifically for adoption, firstly, you need to understand that you have a heterogeneous workforce.
00:20:31.039 --> 00:20:37.359
And there's research actually that talks about the different ways that different generations are using it.
00:20:37.920 --> 00:20:44.880
You have the boomers using it only if it is embedded within things that they trust.
00:20:45.119 --> 00:20:54.000
You have millennials being very pragmatic with it and expecting answers, so very much a you know finding information, retrieval.
00:20:54.160 --> 00:21:11.359
Um, you have millennials that are a lot more um used to planning things with agents, uh, but very task-oriented, and then you have generations uh after that, Gen Z and so on, that they are using as a life coach essentially.
00:21:11.519 --> 00:21:14.319
They do everything with AI first and different mentality.
00:21:14.559 --> 00:21:22.240
So understanding that you have to influence humans, that change management is critical.
00:21:22.799 --> 00:21:23.039
Right?
00:21:23.200 --> 00:21:37.920
So, how do you put in place a journey for them to understand the technology, to get used to um understanding what the technology can do for them, how to embrace it, what are the outputs that you're expecting?
00:21:38.000 --> 00:21:51.519
That vision we talked about before, super important because there is a real fear that AI will take my job instead of AI will remove the drudgery of my daily work so I can focus on value-added tasks.
00:21:51.759 --> 00:21:59.759
Like I see a lot of companies that started earlier in the journey trying to get the ROI by cutting costs and replacing humans.
00:22:00.000 --> 00:22:07.839
And arguably they're coming back into the oh, we failed because we should have thought about this differently.
00:22:08.000 --> 00:22:16.799
We should unlock growth, creativity, innovation with the use of AI to maintain our teams, but grow as a business exponentially.
00:22:16.960 --> 00:22:18.880
And this is where we see the benefit.
00:22:19.039 --> 00:22:23.519
But you need to have that vision, you need to have that clarity, and you need to bring people on a journey.
00:22:23.599 --> 00:22:23.759
Yeah.
00:22:24.160 --> 00:22:26.000
Change management is critical.
00:22:26.160 --> 00:22:36.400
How to use it, motivate them, show them, help them understand how to unpack use cases and what are the tools and empower people with the right tools.
00:22:36.480 --> 00:22:38.480
And that's another thing that is critical, right?
00:22:38.640 --> 00:22:48.640
Because you have the people at the top have a vision, but the people at the bottom are the ones that know what are the things that are time wasters where the bottlenecks are.
00:22:48.799 --> 00:23:01.519
So when you give them the tools with copilot or even low-code or no code, like copilot studio, that gives them the ability to create workload automation and and boost the productivity.
00:23:01.759 --> 00:23:09.279
And while in the end user productivity space, you might get a 10, 20% productivity gains, which is amazing.
00:23:09.759 --> 00:23:15.759
When you see the 10x return, it's when you automate a workflow and you redesign, you reshape that business process.
00:23:15.920 --> 00:23:18.799
You absolutely change the way you interact with your customers.
00:23:18.960 --> 00:23:21.920
You know, you you bend the curve at innovation and that's super rich.
00:23:22.160 --> 00:23:22.319
Yeah.
00:23:22.480 --> 00:23:23.519
When you change the business model.
00:23:24.000 --> 00:23:24.240
Absolutely.
00:23:24.400 --> 00:23:26.480
And that and that's the the the potential here, isn't it?
00:23:26.720 --> 00:23:26.960
Absolutely.
00:23:27.119 --> 00:23:28.880
But you've got to bring the people along along with you.
00:23:29.119 --> 00:23:29.519
100%.
00:23:29.920 --> 00:23:37.039
Because usually what you have is the automation of a workflow, it is um is going across people, right?
00:23:37.359 --> 00:23:41.440
I always think about that, the doorman replacement analogy, right?
00:23:41.599 --> 00:23:45.839
You you hire a consultant to tell you, hey, how can I cut costs?
00:23:46.079 --> 00:23:53.279
And in you're running a hotel, and they might say, you know what, fire the doorman and replace it for an automated door that opens.
00:23:53.519 --> 00:23:58.960
And suddenly you minimize the value of that human to opening a door.
00:23:59.279 --> 00:24:07.200
Well, that person is not just doing that, it's customer experience, is giving you status as a brand, is giving you security.
00:24:07.359 --> 00:24:12.720
There are so many other things that that person does by open but by opening the door.
00:24:12.880 --> 00:24:17.599
So it's a misconcept of how can I cut the full cost of this.
00:24:17.680 --> 00:24:20.559
So you need to be cautious with that.
00:24:20.799 --> 00:24:24.720
However, there are scenarios where it is very obvious, right?
00:24:25.039 --> 00:24:30.480
Support tickets or time spent on specific things, answering queries, very repetitive tasks.
00:24:30.640 --> 00:24:32.799
That's where a lot of businesses are starting, right?
00:24:33.359 --> 00:24:36.960
And and that that makes a lot of sense because there is a very clear cut.
00:24:37.279 --> 00:24:37.839
Fantastic.
00:24:37.920 --> 00:24:45.359
Um Nico, I could speak for another hour with you about AI adoption and the challenges that we see in the market.
00:24:45.519 --> 00:24:48.480
Um, but but sadly, we're gonna run out run out of time.
00:24:48.640 --> 00:24:50.880
Um, I just want to thank you very much for sharing your insights.
00:24:50.960 --> 00:24:53.200
It's been really, really good fun having you here.
00:24:53.279 --> 00:24:59.039
Um, and uh and I wish you all the best and uh and thank you for joining us here on uh Future Ready with Beckla.
00:24:59.200 --> 00:24:59.839
Thank you so much.
