WEBVTT
00:00:04.480 --> 00:00:05.759
Hello, welcome to today's show.
00:00:05.839 --> 00:00:09.199
I'm Jacqueline Stockwell, CEO of Founder Leadership Through Data.
00:00:09.359 --> 00:00:12.800
I inspire and motivate information leaders across the world.
00:00:13.679 --> 00:00:22.000
Today I am with Kevin McDonald, and we're going to be talking about artificial intelligence, especially copiler and co-work.
00:00:22.239 --> 00:00:29.760
Kevin has been an MVP since 2021 and he's an absolute true guide for the modern workplace space.
00:00:30.079 --> 00:00:43.759
Now, Kevin has over a decade in the financial services, shaping digital workplaces before moving into consultancy, where he handles everything from high-level strategy to hands-on adoption.
00:00:44.000 --> 00:00:51.039
In his daily role, Kevin helps support organization strategies across Microsoft's various co-pilot endpoints.
00:00:51.359 --> 00:01:08.239
And he's deeply passionate about helping people unlock their potential using AI to deliver productivity whilst ensuring business understands what to use and when to use it and how to actually get measurable value out of their AI investments.
00:01:08.719 --> 00:01:09.840
Thank you very much for having me on here.
00:01:10.000 --> 00:01:10.480
It's a pleasure.
00:01:10.879 --> 00:01:15.120
We often hear the phrase your AI is only as good as your data.
00:01:15.359 --> 00:01:24.560
From an information management perspective, what is the biggest wake-up call organizations get when they turn on Microsoft Copilot for the first time?
00:01:24.879 --> 00:01:30.400
So I think the the first one when they turn it on, it's actually really easy to turn on, which is fantastic.
00:01:30.879 --> 00:01:38.719
Having been through some products where you have to go round and round in circles enabling things, it's quite refreshing just go dick, it's on.
00:01:38.879 --> 00:01:46.640
It's also one of the scariest things, because then people start using a bit and start investigating what you can get back from there.
00:01:46.879 --> 00:01:48.480
And that's that's the problem.
00:01:48.719 --> 00:02:01.359
It's all about permissions and that the fact that despite for years and years and years talking about governance and information architecture and making sure you're putting the right permissions on, very few organizations actually do it.
00:02:01.439 --> 00:02:08.080
I I remember years ago when I was not on the consulting side looking for a headset, I need to get a replacement one.
00:02:08.319 --> 00:02:12.960
And the top result came back with the head of service desk's one-to-one notes with his team.
00:02:13.120 --> 00:02:15.199
And they included some quite personal information.
00:02:15.360 --> 00:02:21.280
Thankfully I knew him, so uh messed him quite away and said, you really need to change the permissions not to give everyone from there.
00:02:21.439 --> 00:02:23.759
But it's things like that that come up.
00:02:24.000 --> 00:02:25.439
And this isn't a new problem.
00:02:25.520 --> 00:02:28.719
You've been able to do search and bring that for a long time.
00:02:29.039 --> 00:02:40.560
The problem now is those results are coming back into prompts, and you're not I I wouldn't necessarily see quickly if I wasn't checking the citation that it had come from someone's notes from there.
00:02:40.639 --> 00:02:48.800
So you can get information surfaced without it being obvious that you shouldn't see things, and that's that's kind of the risk of it that people are seeing.
00:02:48.960 --> 00:02:50.400
And how severe is it?
00:02:50.560 --> 00:02:54.159
Well, it's a classic consulting answer, but it it depends on that.
00:02:54.319 --> 00:03:00.960
Have organizations actually try to address some of their governance, have they put in place some life cycle for their data on that?
00:03:01.120 --> 00:03:09.599
And then you've got to balance that desire to kind of have the perfect data for AI with actually not being left behind.
00:03:09.840 --> 00:03:15.680
And I think in a very binary world that we live in too often, this is one of those nuanced things.
00:03:15.840 --> 00:03:21.759
You shouldn't either have the perfect data and go with AI or not have the perfect data and not use any AI.
00:03:21.919 --> 00:03:26.639
It's gotta be a scaling bit on there, and it can also be two-tone within that.
00:03:26.719 --> 00:03:30.240
And that's that's the kind of shock of people seeing that information.
00:03:30.400 --> 00:03:35.520
It's really how you handle that that the the top organizations will come out on top, really.
00:03:35.840 --> 00:03:45.599
Microsoft have introduced copilot endpoint acts across their entire stack, from Microsoft 365 to security and power platforms.
00:03:45.759 --> 00:03:50.960
How do you help organisations like map out these endpoints so they work together all seamlessly?
00:03:51.199 --> 00:03:58.800
So I I think the first thing I'd say is working out what challenge an organization is trying to solve within there.
00:03:59.039 --> 00:04:05.439
It's very easy in this kind of AI will solve everything, let's throw copilot everyone, that will that will solve it.
00:04:05.599 --> 00:04:11.919
Actually, when I speak to organizations, trying to get under what they're actually trying to achieve with it is really important.
00:04:12.080 --> 00:04:21.040
If you just go and have that conversation and they want to be more productive and things like that, very good and noble things, but there's usually more to it than that.
00:04:21.199 --> 00:04:24.160
So kind of understanding what they're looking to get.
00:04:24.399 --> 00:04:35.920
So if you're looking to be more effective at handling when you have a security incident, either preventing that or handling that faster, then something like security co-pilot's going to be where you're going with it.
00:04:36.000 --> 00:04:43.839
If you're looking to answer questions on your organization's content in a more effective way, M365 Copilot will be fantastic for that.
00:04:44.000 --> 00:04:51.519
If you're looking to pull information from different sources and actually take intelligent action on those, cowork is even better.
00:04:51.600 --> 00:04:55.839
And I'm sure we'll come back more onto co-work as the top topic at the moment.
00:04:56.160 --> 00:04:58.720
But all of these things have got different ways.
00:04:58.800 --> 00:05:04.720
They've all got different costs, both at running, implementing, making sure people use.
00:05:05.040 --> 00:05:10.160
So looking through the value that they bring from that is key as to whether you go from it.
00:05:10.240 --> 00:05:12.959
And again, a very hot topic with co-work on that.
00:05:13.120 --> 00:05:18.319
So in in terms of the actual endpoints and where people will use those, it's more about what they need.
00:05:18.560 --> 00:05:23.519
The one thing I would say is get this AI into where people live.
00:05:23.680 --> 00:05:30.160
It's why Satch and Adela's comment of co-pilot is the UI for AI love, because you can go into Teams.
00:05:30.240 --> 00:05:31.680
You have that single place.
00:05:31.839 --> 00:05:41.920
You don't have to kind of go, oh, hang on, I want to use AI for incident management, so I need to go to this tool, or I want to do its search policies, so I need to remember that other URL.
00:05:42.079 --> 00:05:46.879
It should be about building things out for the user to be as simple as possible.
00:05:46.959 --> 00:05:57.279
And I think by bringing that single endpoint of co-pilot and kind of making it easy to do things from there is where organizations will see the most adoption and the most value from that investment as well.
00:05:57.759 --> 00:06:02.959
I'm going to ask you about co-work now, because you've introduced it very, very well in terms of that.
00:06:03.120 --> 00:06:04.560
So what is co-work?
00:06:04.720 --> 00:06:05.839
How does it work?
00:06:06.160 --> 00:06:07.279
What's the benefits?
00:06:07.600 --> 00:06:14.480
So co-work is, sorry, I should say co-pilot co-work, because there is a clawed co-work as well, just to confuse everyone.
00:06:14.639 --> 00:06:20.560
But it's Microsoft's way of taking almost the planning and loop element behind the scene.
00:06:20.800 --> 00:06:25.360
And sorry, that's loop with a small L before I confuse people with lots of Microsoft Lingo.
00:06:25.519 --> 00:06:28.000
So it it kind of goes away and takes action.
00:06:28.319 --> 00:06:41.439
If people have used the researcher agent in M3 Copilot, it kind of asks you a question and then goes away, researches it, looks up a bit more information, pulls that together into a big plan before executing.
00:06:41.519 --> 00:06:43.439
And co-works like that as well.
00:06:43.680 --> 00:06:47.439
The difference is it actually takes action as well.
00:06:47.600 --> 00:06:49.839
So it can help you build a presentation.
00:06:50.079 --> 00:06:53.680
My favorite button is please sort out my calendar.
00:06:53.759 --> 00:07:02.000
I'm I'm sure many of your listeners are working at large enterprises and have kind of double booked, quadruple booked slots in their calendar.
00:07:02.160 --> 00:07:06.399
I have exactly the same, and it's often kind of last minute like, ah, what's the biggest priority?
00:07:06.480 --> 00:07:07.839
Oh, I can't turn up to those.
00:07:08.000 --> 00:07:11.199
And now on a daily basis, cowork goes through and actually takes action.
00:07:11.279 --> 00:07:23.040
I've given it some guidance and said this is a regular meeting, the kind of irregular ones will usually make that usually be more important, or if it's with certain people or certain topics, and it takes action and cleans up my calendar.
00:07:23.120 --> 00:07:27.839
And yes, I have it actually declining meetings without me having to click yes to each one.
00:07:27.920 --> 00:07:29.040
I've trusted it enough.
00:07:29.199 --> 00:07:36.079
Initially I didn't, I now trust it enough and will have it just go out and and do those rejections for me from there.
00:07:36.240 --> 00:07:37.680
And why is it different?
00:07:37.839 --> 00:07:44.480
It's using the latest models most effectively to actually bring those tools in and capabilities.
00:07:44.720 --> 00:07:50.240
I would say some people look and go, isn't this kind of what we were promised with Copilot in the first place?
00:07:50.399 --> 00:07:50.800
Yeah.
00:07:51.199 --> 00:07:52.480
I'm not going to shy away from that.
00:07:52.639 --> 00:07:53.199
Yes, it is.
00:07:53.360 --> 00:08:05.920
It's a lot of those things that we thought it would go and do things for us, build up presentations, but the technology and the way they've built it has caught up with that capability to actually do these things that people are expecting.
00:08:06.079 --> 00:08:08.319
And that's why it's it's really taken off.
00:08:08.399 --> 00:08:09.600
There's excitement about it.
00:08:09.759 --> 00:08:12.399
But as with all good things, there's a bad side to it.
00:08:12.639 --> 00:08:15.040
And I think it well, bad side.
00:08:15.279 --> 00:08:18.959
I'm gonna shy away from that because my friend Dan Laskovitz will shout at me otherwise.
00:08:19.360 --> 00:08:20.879
But it's it's the cost of it.
00:08:21.040 --> 00:08:28.560
It is if you use it as you use co-pilot, um you're you're gonna see a huge increase in costs.
00:08:28.720 --> 00:08:33.519
Three, four, five, even more times the cost of using that in the same way.
00:08:33.759 --> 00:08:39.039
So there's now the complication is when should you use co-work for most value?
00:08:39.200 --> 00:08:41.440
When can you do things in co-pilot?
00:08:41.519 --> 00:08:44.879
Because you still need an M365 co-pilot license for it.
00:08:45.120 --> 00:08:46.399
What is the big value add?
00:08:46.639 --> 00:08:51.200
And I'm gonna be brutally honest, I don't right now have a quick answer to that.
00:08:51.360 --> 00:09:01.600
I'm seeing things where you want to kind of combine a few different sources and take action on those and build things out where you want it to kind of repeatedly look through things.
00:09:01.919 --> 00:09:09.039
I've got a lovely demo of where it goes through my Azure DevOps, looks at things that are out of date and nudges people to go in and update that.
00:09:09.200 --> 00:09:16.960
The example I had of clearing up calendars, clearing up my inbox and kind of marking things in red, or really highlighting what my priorities are.
00:09:17.120 --> 00:09:18.960
It's fantastic for that.
00:09:19.120 --> 00:09:22.720
And then you need to look at will this specifically save me time?
00:09:22.879 --> 00:09:24.639
How much money is that time saved?
00:09:24.799 --> 00:09:25.919
Yes, it is worth it.
00:09:26.000 --> 00:09:30.960
I think we'll see an evolution of people working out the right things to do in there for the cost.
00:09:31.279 --> 00:09:37.519
I think we'll also see an evolution in helping people to understand when it is the right thing to use on there as well.
00:09:37.679 --> 00:09:39.120
So it's a lovely tool.
00:09:39.200 --> 00:09:49.200
I I'm kind of recommending some clients are cautious with it right now because if it wasn't for the cost, I would say go help a leather, do everything in there because it's great.
00:09:49.360 --> 00:09:53.679
With that cost, you've now got to work work a little bit harder to work it out.
00:09:53.919 --> 00:10:04.159
I think we'll see things make, I think we'll see both Microsoft and people around the community make it easier to understand and in even potentially automatically doing that.
00:10:04.240 --> 00:10:05.600
This is no NDA breach.
00:10:05.679 --> 00:10:09.200
This is probably a desire and a view of things within there.
00:10:09.360 --> 00:10:22.480
I think we'll see that kind of integration of items that it will happen and we'll automatically move into the right thing and and almost say, in fact, as I have him here, you can almost imagine Clippy popping up and going, Hey, would you like to use cowork to do this?
00:10:22.639 --> 00:10:26.240
Where it analyses the right things to do there, that it sees the value from that.
00:10:26.399 --> 00:10:29.600
I think we'll see things evolve in that direction.
00:10:29.759 --> 00:10:45.679
It it's a new technology, and I I think I I wrote a blog post years ago about uh Brexit and Microsoft 365, where we saw that the kind of spread of people wanting to move fast and those who didn't like change was causing a real problem um within there.
00:10:45.919 --> 00:10:49.840
I feel that problem's got even even worse as we've gone along.
00:10:49.919 --> 00:10:50.960
It's a lot harder.
00:10:51.120 --> 00:10:56.000
Those who want to take the cutting edge can do so much but take a risk.
00:10:56.159 --> 00:10:59.519
And a lot of people just don't want to keep up as much within there.
00:10:59.840 --> 00:11:03.679
And I think co-work is one of those real situations where we see that.
00:11:04.000 --> 00:11:12.080
If you have co-pilot, because there's multiple co-pilots, do you just have to have the co-pilot license for the Microsoft 365?
00:11:12.240 --> 00:11:15.759
Or can you have another co co-pilot licensed?
00:11:16.399 --> 00:11:21.360
Now you need the Microsoft 365 co-pilot, the one that's uh list price is the$30.
00:11:21.840 --> 00:11:25.600
I think it's about£25 per user per month in the UK.
00:11:25.840 --> 00:11:36.320
And then you have something called copilot credits um that work out, and you can assign you basically assign those and you assign a budget to those, and co-work will use those credits.
00:11:36.480 --> 00:11:49.279
And depending on the type of work, I can't remember exactly off the top of my head, I think light tasks uh are like 100 to 200 credits, medium tasks, so where they kind of going through a bit more data, actually taking more action.
00:11:49.440 --> 00:11:53.679
You're looking, I think it's 200 to 600 roughly.
00:11:53.840 --> 00:11:58.639
Uh, please do, uh, much like AI, do go and check these uh citations from there.
00:11:58.879 --> 00:12:01.919
And then for larger tasks, you're looking at 700 from there.
00:12:02.000 --> 00:12:04.240
And you can buy chunks of these credits.
00:12:04.480 --> 00:12:11.200
What I do like that Microsoft has done, I think FinOps and Tokonomics is a big phrase at the moment, how you manage that.
00:12:11.360 --> 00:12:19.759
They've built a cost management dashboard and you can assign budgets at the organization level, security group level, or soon the user level as well.
00:12:19.919 --> 00:12:26.639
So you can say I've got uh 10,000 credits that I can use and it will cost me X amount per month.
00:12:26.720 --> 00:12:29.840
I can't remember the number, which is why I'm not gonna cite that and get in trouble.
00:12:30.000 --> 00:12:36.159
But you you can kind of say you have this much budget that you've got to use, do with that as you want within there.
00:12:36.240 --> 00:12:41.120
And you could decide to give it to a team, you could just give it to a department, to an individual.
00:12:41.279 --> 00:12:42.799
You have the control of that.
00:12:42.960 --> 00:12:45.519
And that I I really thank Microsoft for.
00:12:45.679 --> 00:12:50.559
Plus, they haven't turned it on by default, you have to actively turn it on, which is really good.
00:12:50.799 --> 00:12:51.360
That's really good.
00:12:51.759 --> 00:12:54.559
So they have put control, which they haven't always done recently.
00:12:54.799 --> 00:12:57.200
So it's a pay as you go service, pretty much, is what you're saying.
00:12:57.440 --> 00:13:03.039
You buy credits and then you go into co-work, you do something in there, and then you draw down on those credits.
00:13:03.200 --> 00:13:05.840
I want to talk about redundant, obsolete, and trivial data.
00:13:06.000 --> 00:13:12.159
So before artificial intelligence, bad archiving just meant a cluttered SharePoint sign.
00:13:12.559 --> 00:13:18.559
Now it means copilot might reference a draft policy from years ago instead of the current version.
00:13:18.799 --> 00:13:25.519
How should information managers tackle data lifecycle to keep copilot accurate accurate?
00:13:25.840 --> 00:13:27.840
I don't I don't think much has changed really.
00:13:27.919 --> 00:13:30.639
I I think you're right that copilot will surface that.
00:13:30.799 --> 00:13:34.559
And as I said before, it will be less obvious that you're getting something out of date.
00:13:34.720 --> 00:13:40.559
The same thing would happen if you had V1, V2, V3, and you end up getting something from V2.
00:13:40.960 --> 00:13:44.320
Just because the name had V2, you wouldn't know that wasn't the latest one.
00:13:44.399 --> 00:13:47.120
And by searching for it, you'd still get the same thing.
00:13:47.519 --> 00:13:58.080
Number of times still, I see horrific folder structures that kind of come up and evolve, and someone adds another one and then they add to different things and you can't find things within there.
00:13:58.240 --> 00:14:04.320
With Copilot, it's more likely you will find the latest thing and the thing that is most active from there.
00:14:04.480 --> 00:14:18.480
So you're in some ways reducing the the risk because the risk that was still there anyway, if you've got a better life cycle strategy, you're increasing the likelihood that you're going to get the right content from that.
00:14:18.639 --> 00:14:26.639
So I think bringing the right thing, bringing the right information architecture is still key, bringing in the right data lifecycle to change the the data there.
00:14:26.799 --> 00:14:32.320
The question I always ask is, we've known for a long time how to do this, but people aren't doing it.
00:14:32.480 --> 00:14:32.720
Why?
00:14:32.879 --> 00:14:33.440
Why not?
00:14:33.600 --> 00:14:39.279
Why are why when all these information architects are coming up with these grand things on there, are people not actually doing it?
00:14:39.440 --> 00:14:40.960
And I think it's it's effort.
00:14:41.120 --> 00:14:48.480
It takes a bit of more time and consideration to do this, and people just want to get what they believe their job to be done from there.
00:14:48.639 --> 00:14:52.159
So you've got to try and make it easy for people to do the right thing.
00:14:52.320 --> 00:14:53.360
I love Scott Guthrie.
00:14:53.519 --> 00:14:56.879
He's always talked about helping people fall into the right buckets.
00:14:57.039 --> 00:15:01.759
If you make it easy for people to do the right thing, they are more likely to do that as well.
00:15:01.919 --> 00:15:07.440
So don't put huge architectures that people have to sit there and work out where on earth different things go.
00:15:07.600 --> 00:15:16.480
Can have that natural thing, listen to feedback and evolve things as well, and you're much more likely to get much better content out from it as well.
00:15:16.960 --> 00:15:18.480
It's a team working together on it.
00:15:18.720 --> 00:15:19.120
Agreed.
00:15:19.200 --> 00:15:28.559
And if you need any help with your information architecture in Microsoft 365, you can head over to Leadership Through Data website where we have an amazing course, which is exactly what Kevin is saying.
00:15:28.639 --> 00:15:31.039
So keep it easy, keep it structured.
00:15:31.279 --> 00:15:40.879
So now I just kind of want to ask so if you've identified your redundant, obsolete, and trivial information and SharePoint, could you then say to co-work, deal with this?
00:15:41.279 --> 00:15:42.080
Go and sort out my data.
00:15:44.399 --> 00:15:45.600
Yes, it's a short answer.
00:15:45.679 --> 00:15:57.200
It's the longer answer that is uh I actually had exactly this conversation with someone on the day after the the announcement of co-work's pricing, so uh a couple of weeks ago.
00:15:57.360 --> 00:16:04.240
And I said, if you can sort out their data in the next two weeks, this will be an absolutely fabulous way of doing it.
00:16:04.399 --> 00:16:05.600
You define your policies.
00:16:05.679 --> 00:16:08.080
You don't whereas before you need to set up rules.
00:16:08.320 --> 00:16:22.639
The lovely thing about generative AI is you can kind of describe, I want this sort of thing, I want this sort of thing, and you could maybe even add tags to your document so people can go and review and and it can recommend what is the the rot content from there and not actually take action.
00:16:22.720 --> 00:16:25.519
Or you could say, I'm comfortable with this now, go and take action.
00:16:25.679 --> 00:16:28.799
The question I have now is how much will that cost you?
00:16:28.960 --> 00:16:30.320
And is it worth it?
00:16:30.559 --> 00:16:32.000
Now, you could still do that.
00:16:32.159 --> 00:16:44.799
You could pick on, maybe pick a few folders, look at the cost of implementing something like that, understand whether you want to scale that across gigabytes, terabytes, petabytes that I've seen in organizations and make that decisions on that.
00:16:44.960 --> 00:16:46.639
So yes, it can do it.
00:16:46.799 --> 00:16:56.559
The question is, is the value of doing that automatically going to be cheaper than doing it with I wasn't gonna say not manually, but with human interaction?
00:16:56.879 --> 00:17:04.079
And I think it's a big risk for organizations to have that trust deal with it, you know, a function of of co-work.
00:17:04.559 --> 00:17:04.960
You could do it.
00:17:05.039 --> 00:17:05.920
It'd be amazing.
00:17:06.240 --> 00:17:07.759
I'd love to know if people have tried it.
00:17:07.839 --> 00:17:14.799
I if I'm brutal, I feel that the migration and tidy up space is ripe to have an AI tool doing that.
00:17:14.960 --> 00:17:20.400
But with all the vendors out there that must be looking at this, it can't be as simple as you think.
00:17:20.480 --> 00:17:25.119
So uh I'd love to see governance risk security that you've got to wrap around it.
00:17:25.200 --> 00:17:30.480
You've obviously got to make sure what you're gonna say yes to is the right information as well.
00:17:30.799 --> 00:17:35.519
There are larger pieces, certainly, with a lot of the listeners that they work towards.
00:17:35.839 --> 00:17:38.799
So let's talk about governance security and permissions.
00:17:38.960 --> 00:17:41.759
So copilot respects users' permissions.
00:17:42.000 --> 00:17:51.119
Many organizations have terrible per permissions hygiene now, such as files accidentally marked, available to everybody.
00:17:51.359 --> 00:17:59.440
What practical steps must an organization take to audit their information architecture before a copilot rollout?
00:17:59.759 --> 00:18:04.160
So, I mean, the same as if you're doing a migration, the same as if you were moving to another system.
00:18:04.319 --> 00:18:05.759
You you need to look at there.
00:18:05.839 --> 00:18:08.640
And there's a certain amount of engaging the owners.
00:18:08.799 --> 00:18:14.400
Hopefully, you have clearly defined owners within there, understanding their level of commitment.
00:18:14.720 --> 00:18:33.599
There's things like the SharePoint, I always forget what Sam's called, SharePoint Advanced Management that gives you oversharing reports so you can start to dig a little bit deeper into yes, we've got something that's given far too much access to people and within there and get some insight into some of the bits there.
00:18:33.839 --> 00:18:39.440
So it's using a lot of that different information to to try and get the right thing.
00:18:39.680 --> 00:18:44.640
In terms of kind of doing with a copilot rollout, it's it's about reducing risk, really.
00:18:44.799 --> 00:18:48.079
So you're reducing the risk of those permissions being wrong.
00:18:48.240 --> 00:18:56.240
There's two simple ways of doing that, having a smaller number of people who have copilot or reducing the amount of information that they have access to.
00:18:56.400 --> 00:19:05.519
And when when we found uh I think it was about two or three orgs that I was working very closely with in the early access preview of copilots, they kind of knew there was some content.
00:19:05.680 --> 00:19:11.839
Often their kind of core intranet content was fairly well governed, and so they were happy for that to be surfaced.
00:19:12.000 --> 00:19:21.519
I remember one that had this set of sites that was uh linked to a kind of case management system that was absolutely awful.
00:19:21.599 --> 00:19:22.799
And they knew it was awful.
00:19:22.880 --> 00:19:27.680
They had they put a lot of complex permissions in, but never refined them.
00:19:27.759 --> 00:19:31.279
And as people have moved roles, they'd left those complex permissions.
00:19:31.440 --> 00:19:40.160
So we just went, great, exclude that from SharePoint, set up the rules and sorry, exclude that from copilot, set up the rules so that content isn't surfaced through.
00:19:40.319 --> 00:19:45.039
Reduce the risk of the wrong content by focusing on the areas that you're comfortable with.
00:19:45.200 --> 00:19:52.720
And yes, there's going to be a bit where you have that arguments, you know, you you adjust that level, and there's going to be a point where the levels, yes, we should have access, no, we shouldn't.
00:19:52.880 --> 00:19:57.200
That's part of the discussions that you need to go through and and work through within that.
00:19:57.440 --> 00:20:03.359
I would encourage people to I've had a few conversations where SharePoint's been turned off.
00:20:03.440 --> 00:20:10.799
So you've got access to Teams, you've got access to email content calendar, but not SharePoint, because they feel that's where the risks are.
00:20:11.119 --> 00:20:13.039
I'm very, very mixed on that.
00:20:13.119 --> 00:20:15.119
I I'd say not the whole of SharePoint.
00:20:15.440 --> 00:20:20.480
Try and pick up some areas where you are comfortable, look at those business areas and roll it out to them.
00:20:20.640 --> 00:20:25.680
But uh it's it's a discussion to go through on on people who understand their data.
00:20:25.839 --> 00:20:27.119
And it's always fun.
00:20:27.359 --> 00:20:30.079
And I'm saying that with inverted commas if anyone's listening.
00:20:30.559 --> 00:20:32.720
Let's talk about labels and classification.
00:20:32.880 --> 00:20:42.240
How heavily should organisation rely on Microsoft purview and sensitivity labels to restrict what co-copilot can ingest?
00:20:42.400 --> 00:20:46.880
And where does the human behaviour usually break that system?
00:20:47.200 --> 00:20:53.440
I I think there's a big value conversation to this if an organization's already put in their sensitive sensitivity labels.
00:20:53.680 --> 00:20:54.880
That's always a hard word to say.
00:20:55.039 --> 00:20:57.359
They're attention labels, they've got a good strategy.
00:20:57.599 --> 00:21:00.400
Copilot's absolutely in a great place to support that.
00:21:00.559 --> 00:21:10.000
If they haven't, it's a fairly big rollout to look at your to make sure those labels are connect, are correct, valid, that you've engaged people with this.
00:21:10.079 --> 00:21:15.599
This is not a technology project that you just turn on a load of labels without impacting people.
00:21:15.839 --> 00:21:20.319
It has to be related to how people are going to use it on there.
00:21:20.480 --> 00:21:25.680
I've seen a lot of organizations just turn on labels without any communications or just a single email.
00:21:25.839 --> 00:21:30.400
Funnily enough, all people do is to put it to the lowest level and tag it without without looking at it.
00:21:30.559 --> 00:21:39.759
If you then have your AI labels because you're, you've got the you've all the appropriate licenses for that, then everyone grumbles because it's always wrong within there.
00:21:40.079 --> 00:21:43.599
So I I I think it's something you need to look at.
00:21:43.680 --> 00:21:44.720
It's something you need to work out.
00:21:44.799 --> 00:21:52.480
If you're going to invest as a compliance strategy and security strategy, absolutely build that into your co pilot story.
00:21:52.640 --> 00:21:59.119
If you're just doing it for co pilot, it's a lot of effort and you've really got to make sure you consider the value of that worthwhile.
00:21:59.279 --> 00:22:00.720
Do I think people We should do it?
00:22:00.880 --> 00:22:01.200
Yes.
00:22:01.359 --> 00:22:04.880
Do I think they should look at the cost of doing it versus what they're going to get out?
00:22:05.279 --> 00:22:06.079
Even more, yes.
00:22:06.400 --> 00:22:18.480
If you need any help with your sensitivity labels, please head over to the Leadership Free Data website where we do some amazing courses on that with Microsoft 365 Records Management and SharePoint as well.
00:22:18.799 --> 00:22:24.240
So one of the key questions that I want to ask you is about mindset when it comes to copilot.
00:22:24.319 --> 00:22:28.240
What is the biggest barrier to adoption you see right now?
00:22:28.480 --> 00:22:36.480
Is it a lack of technical skills or is it a lack of trust in the information that comes out of copilot?
00:22:36.640 --> 00:22:39.119
Or is it too worried about using AI?
00:22:39.359 --> 00:22:40.559
I'd add a fourth in there.
00:22:40.640 --> 00:22:42.960
I think all of those play into it a bit.
00:22:43.200 --> 00:22:50.319
What we've seen, the biggest one, is even if you get people engaged, the habit of using copilot itself is one of the issues.
00:22:50.400 --> 00:22:52.400
You've got used to a way of working.
00:22:52.559 --> 00:22:53.759
You're asking people to change that.
00:22:53.839 --> 00:22:58.640
And I found this myself, despite day in, day out telling people, here's all the cool prompts you could do.
00:22:58.799 --> 00:23:04.640
It took me quite a long time to kind of build that actual habit of doing it with my own content.
00:23:04.880 --> 00:23:08.400
I'd say now I'm very good at searching for stuff with copilot.
00:23:08.559 --> 00:23:14.640
I'm good at rewriting some of my content in there or at least reviewing some of my content with copilot.
00:23:14.799 --> 00:23:17.119
I I think I could do more, especially with cowork.
00:23:17.279 --> 00:23:24.960
Uh I've only got two days to kind of make the most of the licensing with that, but I see people having the whole second brain and kind of digesting their knowledge.
00:23:25.359 --> 00:23:29.359
I've actually now got Microsoft Scout doing that, digesting my client knowledge.
00:23:29.440 --> 00:23:34.880
So I have a client database almost that that gets updated on a periodic basis that I can go to.
00:23:35.039 --> 00:23:40.319
And that's the kind of grounded information for some of my agents for key updates on those.
00:23:40.400 --> 00:23:43.680
So it's not looking through all my team's meshes all the time.
00:23:43.839 --> 00:23:52.240
If I'm looking for what's happening with this project, what's the latest, I've got a single dashboard that's kind of consistently maintained from there as well.
00:23:52.480 --> 00:24:02.240
And so that kind of habit of doing that, as well as I agree with you, the technical understanding to do that, the technical prompting is key from there as well.
00:24:02.480 --> 00:24:10.480
I'd also say that the type of people, I think the kind of prompting world has scared off a lot of people on there, and that there's different types of prompts.
00:24:10.799 --> 00:24:21.920
People see like pages and pages of prompts that get put together, and then they see the examples on copilot that are like one-liners because they've been tuned to work effectively from that.
00:24:22.160 --> 00:24:33.759
So encouraging people, I think, not to think about prompting, but to think about if I was asking Jackie to go and go and have a look through my content, find these things, I would explain a few things in there.
00:24:33.920 --> 00:24:43.519
Treat copilot like that, treat it like a summer intern, that you want to explain where things are, explain what you want to get done, what you want that outcome, you'll get much better results on there.
00:24:43.680 --> 00:24:51.599
Um they they saw Microsoft from the early research that managers were a lot better at using copilot because they were used to delegating their work.
00:24:51.759 --> 00:24:52.720
That's what you're doing.
00:24:52.799 --> 00:24:55.200
You're delegating some of your thinking to something else.
00:24:55.359 --> 00:24:56.880
So write your prompts like that.
00:24:57.359 --> 00:25:07.359
If copilot is handling the retrieval, summarizing and drafting of our information, what does the core skill set of a knowledge manager look like in the next three to five years?
00:25:07.599 --> 00:25:16.480
I'm gonna be honest, this was the question that gave me nightmares uh on getting the right answer because I I think there's so much fear, you know, you touched on it in the last question.
00:25:16.559 --> 00:25:23.599
There's so much fear about AI and changing jobs within there, and we're used to doing things this way, you either got to change.
00:25:23.759 --> 00:25:26.960
I I think there's an awful lot that will remain the same.
00:25:27.200 --> 00:25:39.119
I think it shifts the knowledge that certainly the knowledge management area from being more of a librarian and police officer to you must do this, you must do this, to more of a coach.
00:25:39.279 --> 00:25:43.839
It's kind of nudging people rather than setting explicit rules from within there.
00:25:44.160 --> 00:25:52.480
So kind of defining strategies rather than policies, defining ways of working rather than rule sets from that.
00:25:52.640 --> 00:25:55.440
So I think we'll see people start to move towards there.
00:25:55.599 --> 00:26:07.359
Since the beginning of the year, the the use of AI skills that I think has evolved from Claude and Anthropic is becoming bigger and bigger across the AI state, and Microsoft's heavily heavily building on these as well.
00:26:07.440 --> 00:26:11.279
So thinking about defining what you do as a as a skill.
00:26:11.599 --> 00:26:16.480
If you haven't seen, I think just Google AI Skills Anthropic has got some great websites on it.
00:26:16.640 --> 00:26:19.119
Microsoft is growing into these as well.
00:26:19.279 --> 00:26:24.240
I think skills.sh is another good website with a set of those.
00:26:24.400 --> 00:26:27.200
Have a look and they describe how you do things.
00:26:27.519 --> 00:26:44.559
Thinking about your own work and building up your patterns of things like that, how you review documents, how you create documents, defining that style as a skill means that you can kind of again delegate some of the actual work from it in the same way that people thought they could just get Copilot to write a presentation.
00:26:44.640 --> 00:26:48.720
You could say, give me a presentation on how copilot can help me.
00:26:48.880 --> 00:26:53.839
The reality is you want to go to that level of depth of what sort of thing, how do you define this?
00:26:53.920 --> 00:26:54.960
How do you break this down?
00:26:55.200 --> 00:26:57.680
You can define those ways of working with skills.
00:26:57.839 --> 00:27:00.240
So getting ahead of that will be absolutely key.
00:27:00.480 --> 00:27:04.400
Worker needs to have a look at the way they write and capture knowledge.
00:27:04.640 --> 00:27:10.640
I think more atomic, smaller bits with small bits of metadata, we're seeing markdown files go down well.
00:27:11.039 --> 00:27:15.920
Ways that work for AI more than for users is going to be the effective way.
00:27:16.000 --> 00:27:21.759
Because AI is going to add that layer of that knowledge that you don't need to browse and search in the right way.
00:27:21.920 --> 00:27:23.440
It's going to make it more effective.
00:27:23.599 --> 00:27:26.559
So building out content in the right way will be key as well.
00:27:26.880 --> 00:27:29.839
But what is Claude and what is Anthropic's?
00:27:30.240 --> 00:27:34.480
Athropic is another AI company, much like OpenAI and others.
00:27:34.640 --> 00:27:36.319
Obviously, Google has its Gemini.
00:27:36.799 --> 00:27:42.559
Anthropic has their models and they've bucketed those under the Claude brand within there.
00:27:42.799 --> 00:27:50.720
Uh and I'd say they have led the way in terms of this next generation of large language models that can think and plan and take action.
00:27:50.880 --> 00:27:57.920
So Claude co-work originated that was that's kind of what Microsoft has based their co-pilot co-work.
00:27:58.000 --> 00:28:01.440
I I would say working with Anthropic, they haven't just ripped off the idea.
00:28:01.599 --> 00:28:05.200
And they they are building a lot of really great AI content.
00:28:05.279 --> 00:28:07.680
I think they are leading a lot of the thinking way.
00:28:07.839 --> 00:28:12.079
Open AI, if one kind of thing, oh, they're way behind again um on their.
00:28:12.240 --> 00:28:15.200
I think they are catching up again, keeping up with this race.
00:28:15.359 --> 00:28:17.200
It's a full-time job for many people.
00:28:17.359 --> 00:28:20.720
So let let others do that, let AI do it for you as well.
00:28:21.039 --> 00:28:21.359
Agreed.
00:28:21.680 --> 00:28:30.079
And the last point, I would definitely say in terms of the knowledge managers' skills in the next three to five years, is that actually it's around us having influence.
00:28:30.319 --> 00:28:30.559
Yeah.
00:28:30.799 --> 00:28:41.440
Having our own personal branding, having those conversations with the organization to say actually this is the value, this is what we need to be thinking about, and really having that standpoint of that human value.
00:28:41.759 --> 00:28:45.920
Our human value because a lot of people are so worried about the technology.
00:28:46.160 --> 00:28:49.920
We've still have that human value, and I see that very much so.
00:28:50.079 --> 00:28:54.559
We just need to adapt our humanisms to the to the new world.
00:28:54.960 --> 00:28:57.680
It's been absolutely sensational to have you on the show.
00:28:57.920 --> 00:29:00.720
How can listeners reach out to you if they want to know more?
00:29:00.960 --> 00:29:06.319
Um, you can find me LinkedIn is the place I'm on the most if you look for Kevin McDonnell MVPs.
00:29:06.640 --> 00:29:12.319
I also have a podcast with my friend uh Zoe Wilson and Gary Trinder called The Copilot Connection.
00:29:12.400 --> 00:29:14.640
Uh so we do it around every two weeks.
00:29:15.119 --> 00:29:17.200
Thank you so much for your time today, Kevin.
00:29:17.279 --> 00:29:18.000
It's been brilliant.
00:29:18.240 --> 00:29:18.799
Not a problem.
00:29:18.880 --> 00:29:19.920
Thanks very much.
00:00:04.480 --> 00:00:05.759
Hello, welcome to today's show.
00:00:05.839 --> 00:00:09.199
I'm Jacqueline Stockwell, CEO of Founder Leadership Through Data.
00:00:09.359 --> 00:00:12.800
I inspire and motivate information leaders across the world.
00:00:13.679 --> 00:00:22.000
Today I am with Kevin McDonald, and we're going to be talking about artificial intelligence, especially copiler and co-work.
00:00:22.239 --> 00:00:29.760
Kevin has been an MVP since 2021 and he's an absolute true guide for the modern workplace space.
00:00:30.079 --> 00:00:43.759
Now, Kevin has over a decade in the financial services, shaping digital workplaces before moving into consultancy, where he handles everything from high-level strategy to hands-on adoption.
00:00:44.000 --> 00:00:51.039
In his daily role, Kevin helps support organization strategies across Microsoft's various co-pilot endpoints.
00:00:51.359 --> 00:01:08.239
And he's deeply passionate about helping people unlock their potential using AI to deliver productivity whilst ensuring business understands what to use and when to use it and how to actually get measurable value out of their AI investments.
00:01:08.719 --> 00:01:09.840
Thank you very much for having me on here.
00:01:10.000 --> 00:01:10.480
It's a pleasure.
00:01:10.879 --> 00:01:15.120
We often hear the phrase your AI is only as good as your data.
00:01:15.359 --> 00:01:24.560
From an information management perspective, what is the biggest wake-up call organizations get when they turn on Microsoft Copilot for the first time?
00:01:24.879 --> 00:01:30.400
So I think the the first one when they turn it on, it's actually really easy to turn on, which is fantastic.
00:01:30.879 --> 00:01:38.719
Having been through some products where you have to go round and round in circles enabling things, it's quite refreshing just go dick, it's on.
00:01:38.879 --> 00:01:46.640
It's also one of the scariest things, because then people start using a bit and start investigating what you can get back from there.
00:01:46.879 --> 00:01:48.480
And that's that's the problem.
00:01:48.719 --> 00:02:01.359
It's all about permissions and that the fact that despite for years and years and years talking about governance and information architecture and making sure you're putting the right permissions on, very few organizations actually do it.
00:02:01.439 --> 00:02:08.080
I I remember years ago when I was not on the consulting side looking for a headset, I need to get a replacement one.
00:02:08.319 --> 00:02:12.960
And the top result came back with the head of service desk's one-to-one notes with his team.
00:02:13.120 --> 00:02:15.199
And they included some quite personal information.
00:02:15.360 --> 00:02:21.280
Thankfully I knew him, so uh messed him quite away and said, you really need to change the permissions not to give everyone from there.
00:02:21.439 --> 00:02:23.759
But it's things like that that come up.
00:02:24.000 --> 00:02:25.439
And this isn't a new problem.
00:02:25.520 --> 00:02:28.719
You've been able to do search and bring that for a long time.
00:02:29.039 --> 00:02:40.560
The problem now is those results are coming back into prompts, and you're not I I wouldn't necessarily see quickly if I wasn't checking the citation that it had come from someone's notes from there.
00:02:40.639 --> 00:02:48.800
So you can get information surfaced without it being obvious that you shouldn't see things, and that's that's kind of the risk of it that people are seeing.
00:02:48.960 --> 00:02:50.400
And how severe is it?
00:02:50.560 --> 00:02:54.159
Well, it's a classic consulting answer, but it it depends on that.
00:02:54.319 --> 00:03:00.960
Have organizations actually try to address some of their governance, have they put in place some life cycle for their data on that?
00:03:01.120 --> 00:03:09.599
And then you've got to balance that desire to kind of have the perfect data for AI with actually not being left behind.
00:03:09.840 --> 00:03:15.680
And I think in a very binary world that we live in too often, this is one of those nuanced things.
00:03:15.840 --> 00:03:21.759
You shouldn't either have the perfect data and go with AI or not have the perfect data and not use any AI.
00:03:21.919 --> 00:03:26.639
It's gotta be a scaling bit on there, and it can also be two-tone within that.
00:03:26.719 --> 00:03:30.240
And that's that's the kind of shock of people seeing that information.
00:03:30.400 --> 00:03:35.520
It's really how you handle that that the the top organizations will come out on top, really.
00:03:35.840 --> 00:03:45.599
Microsoft have introduced copilot endpoint acts across their entire stack, from Microsoft 365 to security and power platforms.
00:03:45.759 --> 00:03:50.960
How do you help organisations like map out these endpoints so they work together all seamlessly?
00:03:51.199 --> 00:03:58.800
So I I think the first thing I'd say is working out what challenge an organization is trying to solve within there.
00:03:59.039 --> 00:04:05.439
It's very easy in this kind of AI will solve everything, let's throw copilot everyone, that will that will solve it.
00:04:05.599 --> 00:04:11.919
Actually, when I speak to organizations, trying to get under what they're actually trying to achieve with it is really important.
00:04:12.080 --> 00:04:21.040
If you just go and have that conversation and they want to be more productive and things like that, very good and noble things, but there's usually more to it than that.
00:04:21.199 --> 00:04:24.160
So kind of understanding what they're looking to get.
00:04:24.399 --> 00:04:35.920
So if you're looking to be more effective at handling when you have a security incident, either preventing that or handling that faster, then something like security co-pilot's going to be where you're going with it.
00:04:36.000 --> 00:04:43.839
If you're looking to answer questions on your organization's content in a more effective way, M365 Copilot will be fantastic for that.
00:04:44.000 --> 00:04:51.519
If you're looking to pull information from different sources and actually take intelligent action on those, cowork is even better.
00:04:51.600 --> 00:04:55.839
And I'm sure we'll come back more onto co-work as the top topic at the moment.
00:04:56.160 --> 00:04:58.720
But all of these things have got different ways.
00:04:58.800 --> 00:05:04.720
They've all got different costs, both at running, implementing, making sure people use.
00:05:05.040 --> 00:05:10.160
So looking through the value that they bring from that is key as to whether you go from it.
00:05:10.240 --> 00:05:12.959
And again, a very hot topic with co-work on that.
00:05:13.120 --> 00:05:18.319
So in in terms of the actual endpoints and where people will use those, it's more about what they need.
00:05:18.560 --> 00:05:23.519
The one thing I would say is get this AI into where people live.
00:05:23.680 --> 00:05:30.160
It's why Satch and Adela's comment of co-pilot is the UI for AI love, because you can go into Teams.
00:05:30.240 --> 00:05:31.680
You have that single place.
00:05:31.839 --> 00:05:41.920
You don't have to kind of go, oh, hang on, I want to use AI for incident management, so I need to go to this tool, or I want to do its search policies, so I need to remember that other URL.
00:05:42.079 --> 00:05:46.879
It should be about building things out for the user to be as simple as possible.
00:05:46.959 --> 00:05:57.279
And I think by bringing that single endpoint of co-pilot and kind of making it easy to do things from there is where organizations will see the most adoption and the most value from that investment as well.
00:05:57.759 --> 00:06:02.959
I'm going to ask you about co-work now, because you've introduced it very, very well in terms of that.
00:06:03.120 --> 00:06:04.560
So what is co-work?
00:06:04.720 --> 00:06:05.839
How does it work?
00:06:06.160 --> 00:06:07.279
What's the benefits?
00:06:07.600 --> 00:06:14.480
So co-work is, sorry, I should say co-pilot co-work, because there is a clawed co-work as well, just to confuse everyone.
00:06:14.639 --> 00:06:20.560
But it's Microsoft's way of taking almost the planning and loop element behind the scene.
00:06:20.800 --> 00:06:25.360
And sorry, that's loop with a small L before I confuse people with lots of Microsoft Lingo.
00:06:25.519 --> 00:06:28.000
So it it kind of goes away and takes action.
00:06:28.319 --> 00:06:41.439
If people have used the researcher agent in M3 Copilot, it kind of asks you a question and then goes away, researches it, looks up a bit more information, pulls that together into a big plan before executing.
00:06:41.519 --> 00:06:43.439
And co-works like that as well.
00:06:43.680 --> 00:06:47.439
The difference is it actually takes action as well.
00:06:47.600 --> 00:06:49.839
So it can help you build a presentation.
00:06:50.079 --> 00:06:53.680
My favorite button is please sort out my calendar.
00:06:53.759 --> 00:07:02.000
I'm I'm sure many of your listeners are working at large enterprises and have kind of double booked, quadruple booked slots in their calendar.
00:07:02.160 --> 00:07:06.399
I have exactly the same, and it's often kind of last minute like, ah, what's the biggest priority?
00:07:06.480 --> 00:07:07.839
Oh, I can't turn up to those.
00:07:08.000 --> 00:07:11.199
And now on a daily basis, cowork goes through and actually takes action.
00:07:11.279 --> 00:07:23.040
I've given it some guidance and said this is a regular meeting, the kind of irregular ones will usually make that usually be more important, or if it's with certain people or certain topics, and it takes action and cleans up my calendar.
00:07:23.120 --> 00:07:27.839
And yes, I have it actually declining meetings without me having to click yes to each one.
00:07:27.920 --> 00:07:29.040
I've trusted it enough.
00:07:29.199 --> 00:07:36.079
Initially I didn't, I now trust it enough and will have it just go out and and do those rejections for me from there.
00:07:36.240 --> 00:07:37.680
And why is it different?
00:07:37.839 --> 00:07:44.480
It's using the latest models most effectively to actually bring those tools in and capabilities.
00:07:44.720 --> 00:07:50.240
I would say some people look and go, isn't this kind of what we were promised with Copilot in the first place?
00:07:50.399 --> 00:07:50.800
Yeah.
00:07:51.199 --> 00:07:52.480
I'm not going to shy away from that.
00:07:52.639 --> 00:07:53.199
Yes, it is.
00:07:53.360 --> 00:08:05.920
It's a lot of those things that we thought it would go and do things for us, build up presentations, but the technology and the way they've built it has caught up with that capability to actually do these things that people are expecting.
00:08:06.079 --> 00:08:08.319
And that's why it's it's really taken off.
00:08:08.399 --> 00:08:09.600
There's excitement about it.
00:08:09.759 --> 00:08:12.399
But as with all good things, there's a bad side to it.
00:08:12.639 --> 00:08:15.040
And I think it well, bad side.
00:08:15.279 --> 00:08:18.959
I'm gonna shy away from that because my friend Dan Laskovitz will shout at me otherwise.
00:08:19.360 --> 00:08:20.879
But it's it's the cost of it.
00:08:21.040 --> 00:08:28.560
It is if you use it as you use co-pilot, um you're you're gonna see a huge increase in costs.
00:08:28.720 --> 00:08:33.519
Three, four, five, even more times the cost of using that in the same way.
00:08:33.759 --> 00:08:39.039
So there's now the complication is when should you use co-work for most value?
00:08:39.200 --> 00:08:41.440
When can you do things in co-pilot?
00:08:41.519 --> 00:08:44.879
Because you still need an M365 co-pilot license for it.
00:08:45.120 --> 00:08:46.399
What is the big value add?
00:08:46.639 --> 00:08:51.200
And I'm gonna be brutally honest, I don't right now have a quick answer to that.
00:08:51.360 --> 00:09:01.600
I'm seeing things where you want to kind of combine a few different sources and take action on those and build things out where you want it to kind of repeatedly look through things.
00:09:01.919 --> 00:09:09.039
I've got a lovely demo of where it goes through my Azure DevOps, looks at things that are out of date and nudges people to go in and update that.
00:09:09.200 --> 00:09:16.960
The example I had of clearing up calendars, clearing up my inbox and kind of marking things in red, or really highlighting what my priorities are.
00:09:17.120 --> 00:09:18.960
It's fantastic for that.
00:09:19.120 --> 00:09:22.720
And then you need to look at will this specifically save me time?
00:09:22.879 --> 00:09:24.639
How much money is that time saved?
00:09:24.799 --> 00:09:25.919
Yes, it is worth it.
00:09:26.000 --> 00:09:30.960
I think we'll see an evolution of people working out the right things to do in there for the cost.
00:09:31.279 --> 00:09:37.519
I think we'll also see an evolution in helping people to understand when it is the right thing to use on there as well.
00:09:37.679 --> 00:09:39.120
So it's a lovely tool.
00:09:39.200 --> 00:09:49.200
I I'm kind of recommending some clients are cautious with it right now because if it wasn't for the cost, I would say go help a leather, do everything in there because it's great.
00:09:49.360 --> 00:09:53.679
With that cost, you've now got to work work a little bit harder to work it out.
00:09:53.919 --> 00:10:04.159
I think we'll see things make, I think we'll see both Microsoft and people around the community make it easier to understand and in even potentially automatically doing that.
00:10:04.240 --> 00:10:05.600
This is no NDA breach.
00:10:05.679 --> 00:10:09.200
This is probably a desire and a view of things within there.
00:10:09.360 --> 00:10:22.480
I think we'll see that kind of integration of items that it will happen and we'll automatically move into the right thing and and almost say, in fact, as I have him here, you can almost imagine Clippy popping up and going, Hey, would you like to use cowork to do this?
00:10:22.639 --> 00:10:26.240
Where it analyses the right things to do there, that it sees the value from that.
00:10:26.399 --> 00:10:29.600
I think we'll see things evolve in that direction.
00:10:29.759 --> 00:10:45.679
It it's a new technology, and I I think I I wrote a blog post years ago about uh Brexit and Microsoft 365, where we saw that the kind of spread of people wanting to move fast and those who didn't like change was causing a real problem um within there.
00:10:45.919 --> 00:10:49.840
I feel that problem's got even even worse as we've gone along.
00:10:49.919 --> 00:10:50.960
It's a lot harder.
00:10:51.120 --> 00:10:56.000
Those who want to take the cutting edge can do so much but take a risk.
00:10:56.159 --> 00:10:59.519
And a lot of people just don't want to keep up as much within there.
00:10:59.840 --> 00:11:03.679
And I think co-work is one of those real situations where we see that.
00:11:04.000 --> 00:11:12.080
If you have co-pilot, because there's multiple co-pilots, do you just have to have the co-pilot license for the Microsoft 365?
00:11:12.240 --> 00:11:15.759
Or can you have another co co-pilot licensed?
00:11:16.399 --> 00:11:21.360
Now you need the Microsoft 365 co-pilot, the one that's uh list price is the$30.
00:11:21.840 --> 00:11:25.600
I think it's about£25 per user per month in the UK.
00:11:25.840 --> 00:11:36.320
And then you have something called copilot credits um that work out, and you can assign you basically assign those and you assign a budget to those, and co-work will use those credits.
00:11:36.480 --> 00:11:49.279
And depending on the type of work, I can't remember exactly off the top of my head, I think light tasks uh are like 100 to 200 credits, medium tasks, so where they kind of going through a bit more data, actually taking more action.
00:11:49.440 --> 00:11:53.679
You're looking, I think it's 200 to 600 roughly.
00:11:53.840 --> 00:11:58.639
Uh, please do, uh, much like AI, do go and check these uh citations from there.
00:11:58.879 --> 00:12:01.919
And then for larger tasks, you're looking at 700 from there.
00:12:02.000 --> 00:12:04.240
And you can buy chunks of these credits.
00:12:04.480 --> 00:12:11.200
What I do like that Microsoft has done, I think FinOps and Tokonomics is a big phrase at the moment, how you manage that.
00:12:11.360 --> 00:12:19.759
They've built a cost management dashboard and you can assign budgets at the organization level, security group level, or soon the user level as well.
00:12:19.919 --> 00:12:26.639
So you can say I've got uh 10,000 credits that I can use and it will cost me X amount per month.
00:12:26.720 --> 00:12:29.840
I can't remember the number, which is why I'm not gonna cite that and get in trouble.
00:12:30.000 --> 00:12:36.159
But you you can kind of say you have this much budget that you've got to use, do with that as you want within there.
00:12:36.240 --> 00:12:41.120
And you could decide to give it to a team, you could just give it to a department, to an individual.
00:12:41.279 --> 00:12:42.799
You have the control of that.
00:12:42.960 --> 00:12:45.519
And that I I really thank Microsoft for.
00:12:45.679 --> 00:12:50.559
Plus, they haven't turned it on by default, you have to actively turn it on, which is really good.
00:12:50.799 --> 00:12:51.360
That's really good.
00:12:51.759 --> 00:12:54.559
So they have put control, which they haven't always done recently.
00:12:54.799 --> 00:12:57.200
So it's a pay as you go service, pretty much, is what you're saying.
00:12:57.440 --> 00:13:03.039
You buy credits and then you go into co-work, you do something in there, and then you draw down on those credits.
00:13:03.200 --> 00:13:05.840
I want to talk about redundant, obsolete, and trivial data.
00:13:06.000 --> 00:13:12.159
So before artificial intelligence, bad archiving just meant a cluttered SharePoint sign.
00:13:12.559 --> 00:13:18.559
Now it means copilot might reference a draft policy from years ago instead of the current version.
00:13:18.799 --> 00:13:25.519
How should information managers tackle data lifecycle to keep copilot accurate accurate?
00:13:25.840 --> 00:13:27.840
I don't I don't think much has changed really.
00:13:27.919 --> 00:13:30.639
I I think you're right that copilot will surface that.
00:13:30.799 --> 00:13:34.559
And as I said before, it will be less obvious that you're getting something out of date.
00:13:34.720 --> 00:13:40.559
The same thing would happen if you had V1, V2, V3, and you end up getting something from V2.
00:13:40.960 --> 00:13:44.320
Just because the name had V2, you wouldn't know that wasn't the latest one.
00:13:44.399 --> 00:13:47.120
And by searching for it, you'd still get the same thing.
00:13:47.519 --> 00:13:58.080
Number of times still, I see horrific folder structures that kind of come up and evolve, and someone adds another one and then they add to different things and you can't find things within there.
00:13:58.240 --> 00:14:04.320
With Copilot, it's more likely you will find the latest thing and the thing that is most active from there.
00:14:04.480 --> 00:14:18.480
So you're in some ways reducing the the risk because the risk that was still there anyway, if you've got a better life cycle strategy, you're increasing the likelihood that you're going to get the right content from that.
00:14:18.639 --> 00:14:26.639
So I think bringing the right thing, bringing the right information architecture is still key, bringing in the right data lifecycle to change the the data there.
00:14:26.799 --> 00:14:32.320
The question I always ask is, we've known for a long time how to do this, but people aren't doing it.
00:14:32.480 --> 00:14:32.720
Why?
00:14:32.879 --> 00:14:33.440
Why not?
00:14:33.600 --> 00:14:39.279
Why are why when all these information architects are coming up with these grand things on there, are people not actually doing it?
00:14:39.440 --> 00:14:40.960
And I think it's it's effort.
00:14:41.120 --> 00:14:48.480
It takes a bit of more time and consideration to do this, and people just want to get what they believe their job to be done from there.
00:14:48.639 --> 00:14:52.159
So you've got to try and make it easy for people to do the right thing.
00:14:52.320 --> 00:14:53.360
I love Scott Guthrie.
00:14:53.519 --> 00:14:56.879
He's always talked about helping people fall into the right buckets.
00:14:57.039 --> 00:15:01.759
If you make it easy for people to do the right thing, they are more likely to do that as well.
00:15:01.919 --> 00:15:07.440
So don't put huge architectures that people have to sit there and work out where on earth different things go.
00:15:07.600 --> 00:15:16.480
Can have that natural thing, listen to feedback and evolve things as well, and you're much more likely to get much better content out from it as well.
00:15:16.960 --> 00:15:18.480
It's a team working together on it.
00:15:18.720 --> 00:15:19.120
Agreed.
00:15:19.200 --> 00:15:28.559
And if you need any help with your information architecture in Microsoft 365, you can head over to Leadership Through Data website where we have an amazing course, which is exactly what Kevin is saying.
00:15:28.639 --> 00:15:31.039
So keep it easy, keep it structured.
00:15:31.279 --> 00:15:40.879
So now I just kind of want to ask so if you've identified your redundant, obsolete, and trivial information and SharePoint, could you then say to co-work, deal with this?
00:15:41.279 --> 00:15:42.080
Go and sort out my data.
00:15:44.399 --> 00:15:45.600
Yes, it's a short answer.
00:15:45.679 --> 00:15:57.200
It's the longer answer that is uh I actually had exactly this conversation with someone on the day after the the announcement of co-work's pricing, so uh a couple of weeks ago.
00:15:57.360 --> 00:16:04.240
And I said, if you can sort out their data in the next two weeks, this will be an absolutely fabulous way of doing it.
00:16:04.399 --> 00:16:05.600
You define your policies.
00:16:05.679 --> 00:16:08.080
You don't whereas before you need to set up rules.
00:16:08.320 --> 00:16:22.639
The lovely thing about generative AI is you can kind of describe, I want this sort of thing, I want this sort of thing, and you could maybe even add tags to your document so people can go and review and and it can recommend what is the the rot content from there and not actually take action.
00:16:22.720 --> 00:16:25.519
Or you could say, I'm comfortable with this now, go and take action.
00:16:25.679 --> 00:16:28.799
The question I have now is how much will that cost you?
00:16:28.960 --> 00:16:30.320
And is it worth it?
00:16:30.559 --> 00:16:32.000
Now, you could still do that.
00:16:32.159 --> 00:16:44.799
You could pick on, maybe pick a few folders, look at the cost of implementing something like that, understand whether you want to scale that across gigabytes, terabytes, petabytes that I've seen in organizations and make that decisions on that.
00:16:44.960 --> 00:16:46.639
So yes, it can do it.
00:16:46.799 --> 00:16:56.559
The question is, is the value of doing that automatically going to be cheaper than doing it with I wasn't gonna say not manually, but with human interaction?
00:16:56.879 --> 00:17:04.079
And I think it's a big risk for organizations to have that trust deal with it, you know, a function of of co-work.
00:17:04.559 --> 00:17:04.960
You could do it.
00:17:05.039 --> 00:17:05.920
It'd be amazing.
00:17:06.240 --> 00:17:07.759
I'd love to know if people have tried it.
00:17:07.839 --> 00:17:14.799
I if I'm brutal, I feel that the migration and tidy up space is ripe to have an AI tool doing that.
00:17:14.960 --> 00:17:20.400
But with all the vendors out there that must be looking at this, it can't be as simple as you think.
00:17:20.480 --> 00:17:25.119
So uh I'd love to see governance risk security that you've got to wrap around it.
00:17:25.200 --> 00:17:30.480
You've obviously got to make sure what you're gonna say yes to is the right information as well.
00:17:30.799 --> 00:17:35.519
There are larger pieces, certainly, with a lot of the listeners that they work towards.
00:17:35.839 --> 00:17:38.799
So let's talk about governance security and permissions.
00:17:38.960 --> 00:17:41.759
So copilot respects users' permissions.
00:17:42.000 --> 00:17:51.119
Many organizations have terrible per permissions hygiene now, such as files accidentally marked, available to everybody.
00:17:51.359 --> 00:17:59.440
What practical steps must an organization take to audit their information architecture before a copilot rollout?
00:17:59.759 --> 00:18:04.160
So, I mean, the same as if you're doing a migration, the same as if you were moving to another system.
00:18:04.319 --> 00:18:05.759
You you need to look at there.
00:18:05.839 --> 00:18:08.640
And there's a certain amount of engaging the owners.
00:18:08.799 --> 00:18:14.400
Hopefully, you have clearly defined owners within there, understanding their level of commitment.
00:18:14.720 --> 00:18:33.599
There's things like the SharePoint, I always forget what Sam's called, SharePoint Advanced Management that gives you oversharing reports so you can start to dig a little bit deeper into yes, we've got something that's given far too much access to people and within there and get some insight into some of the bits there.
00:18:33.839 --> 00:18:39.440
So it's using a lot of that different information to to try and get the right thing.
00:18:39.680 --> 00:18:44.640
In terms of kind of doing with a copilot rollout, it's it's about reducing risk, really.
00:18:44.799 --> 00:18:48.079
So you're reducing the risk of those permissions being wrong.
00:18:48.240 --> 00:18:56.240
There's two simple ways of doing that, having a smaller number of people who have copilot or reducing the amount of information that they have access to.
00:18:56.400 --> 00:19:05.519
And when when we found uh I think it was about two or three orgs that I was working very closely with in the early access preview of copilots, they kind of knew there was some content.
00:19:05.680 --> 00:19:11.839
Often their kind of core intranet content was fairly well governed, and so they were happy for that to be surfaced.
00:19:12.000 --> 00:19:21.519
I remember one that had this set of sites that was uh linked to a kind of case management system that was absolutely awful.
00:19:21.599 --> 00:19:22.799
And they knew it was awful.
00:19:22.880 --> 00:19:27.680
They had they put a lot of complex permissions in, but never refined them.
00:19:27.759 --> 00:19:31.279
And as people have moved roles, they'd left those complex permissions.
00:19:31.440 --> 00:19:40.160
So we just went, great, exclude that from SharePoint, set up the rules and sorry, exclude that from copilot, set up the rules so that content isn't surfaced through.
00:19:40.319 --> 00:19:45.039
Reduce the risk of the wrong content by focusing on the areas that you're comfortable with.
00:19:45.200 --> 00:19:52.720
And yes, there's going to be a bit where you have that arguments, you know, you you adjust that level, and there's going to be a point where the levels, yes, we should have access, no, we shouldn't.
00:19:52.880 --> 00:19:57.200
That's part of the discussions that you need to go through and and work through within that.
00:19:57.440 --> 00:20:03.359
I would encourage people to I've had a few conversations where SharePoint's been turned off.
00:20:03.440 --> 00:20:10.799
So you've got access to Teams, you've got access to email content calendar, but not SharePoint, because they feel that's where the risks are.
00:20:11.119 --> 00:20:13.039
I'm very, very mixed on that.
00:20:13.119 --> 00:20:15.119
I I'd say not the whole of SharePoint.
00:20:15.440 --> 00:20:20.480
Try and pick up some areas where you are comfortable, look at those business areas and roll it out to them.
00:20:20.640 --> 00:20:25.680
But uh it's it's a discussion to go through on on people who understand their data.
00:20:25.839 --> 00:20:27.119
And it's always fun.
00:20:27.359 --> 00:20:30.079
And I'm saying that with inverted commas if anyone's listening.
00:20:30.559 --> 00:20:32.720
Let's talk about labels and classification.
00:20:32.880 --> 00:20:42.240
How heavily should organisation rely on Microsoft purview and sensitivity labels to restrict what co-copilot can ingest?
00:20:42.400 --> 00:20:46.880
And where does the human behaviour usually break that system?
00:20:47.200 --> 00:20:53.440
I I think there's a big value conversation to this if an organization's already put in their sensitive sensitivity labels.
00:20:53.680 --> 00:20:54.880
That's always a hard word to say.
00:20:55.039 --> 00:20:57.359
They're attention labels, they've got a good strategy.
00:20:57.599 --> 00:21:00.400
Copilot's absolutely in a great place to support that.
00:21:00.559 --> 00:21:10.000
If they haven't, it's a fairly big rollout to look at your to make sure those labels are connect, are correct, valid, that you've engaged people with this.
00:21:10.079 --> 00:21:15.599
This is not a technology project that you just turn on a load of labels without impacting people.
00:21:15.839 --> 00:21:20.319
It has to be related to how people are going to use it on there.
00:21:20.480 --> 00:21:25.680
I've seen a lot of organizations just turn on labels without any communications or just a single email.
00:21:25.839 --> 00:21:30.400
Funnily enough, all people do is to put it to the lowest level and tag it without without looking at it.
00:21:30.559 --> 00:21:39.759
If you then have your AI labels because you're, you've got the you've all the appropriate licenses for that, then everyone grumbles because it's always wrong within there.
00:21:40.079 --> 00:21:43.599
So I I I think it's something you need to look at.
00:21:43.680 --> 00:21:44.720
It's something you need to work out.
00:21:44.799 --> 00:21:52.480
If you're going to invest as a compliance strategy and security strategy, absolutely build that into your co pilot story.
00:21:52.640 --> 00:21:59.119
If you're just doing it for co pilot, it's a lot of effort and you've really got to make sure you consider the value of that worthwhile.
00:21:59.279 --> 00:22:00.720
Do I think people We should do it?
00:22:00.880 --> 00:22:01.200
Yes.
00:22:01.359 --> 00:22:04.880
Do I think they should look at the cost of doing it versus what they're going to get out?
00:22:05.279 --> 00:22:06.079
Even more, yes.
00:22:06.400 --> 00:22:18.480
If you need any help with your sensitivity labels, please head over to the Leadership Free Data website where we do some amazing courses on that with Microsoft 365 Records Management and SharePoint as well.
00:22:18.799 --> 00:22:24.240
So one of the key questions that I want to ask you is about mindset when it comes to copilot.
00:22:24.319 --> 00:22:28.240
What is the biggest barrier to adoption you see right now?
00:22:28.480 --> 00:22:36.480
Is it a lack of technical skills or is it a lack of trust in the information that comes out of copilot?
00:22:36.640 --> 00:22:39.119
Or is it too worried about using AI?
00:22:39.359 --> 00:22:40.559
I'd add a fourth in there.
00:22:40.640 --> 00:22:42.960
I think all of those play into it a bit.
00:22:43.200 --> 00:22:50.319
What we've seen, the biggest one, is even if you get people engaged, the habit of using copilot itself is one of the issues.
00:22:50.400 --> 00:22:52.400
You've got used to a way of working.
00:22:52.559 --> 00:22:53.759
You're asking people to change that.
00:22:53.839 --> 00:22:58.640
And I found this myself, despite day in, day out telling people, here's all the cool prompts you could do.
00:22:58.799 --> 00:23:04.640
It took me quite a long time to kind of build that actual habit of doing it with my own content.
00:23:04.880 --> 00:23:08.400
I'd say now I'm very good at searching for stuff with copilot.
00:23:08.559 --> 00:23:14.640
I'm good at rewriting some of my content in there or at least reviewing some of my content with copilot.
00:23:14.799 --> 00:23:17.119
I I think I could do more, especially with cowork.
00:23:17.279 --> 00:23:24.960
Uh I've only got two days to kind of make the most of the licensing with that, but I see people having the whole second brain and kind of digesting their knowledge.
00:23:25.359 --> 00:23:29.359
I've actually now got Microsoft Scout doing that, digesting my client knowledge.
00:23:29.440 --> 00:23:34.880
So I have a client database almost that that gets updated on a periodic basis that I can go to.
00:23:35.039 --> 00:23:40.319
And that's the kind of grounded information for some of my agents for key updates on those.
00:23:40.400 --> 00:23:43.680
So it's not looking through all my team's meshes all the time.
00:23:43.839 --> 00:23:52.240
If I'm looking for what's happening with this project, what's the latest, I've got a single dashboard that's kind of consistently maintained from there as well.
00:23:52.480 --> 00:24:02.240
And so that kind of habit of doing that, as well as I agree with you, the technical understanding to do that, the technical prompting is key from there as well.
00:24:02.480 --> 00:24:10.480
I'd also say that the type of people, I think the kind of prompting world has scared off a lot of people on there, and that there's different types of prompts.
00:24:10.799 --> 00:24:21.920
People see like pages and pages of prompts that get put together, and then they see the examples on copilot that are like one-liners because they've been tuned to work effectively from that.
00:24:22.160 --> 00:24:33.759
So encouraging people, I think, not to think about prompting, but to think about if I was asking Jackie to go and go and have a look through my content, find these things, I would explain a few things in there.
00:24:33.920 --> 00:24:43.519
Treat copilot like that, treat it like a summer intern, that you want to explain where things are, explain what you want to get done, what you want that outcome, you'll get much better results on there.
00:24:43.680 --> 00:24:51.599
Um they they saw Microsoft from the early research that managers were a lot better at using copilot because they were used to delegating their work.
00:24:51.759 --> 00:24:52.720
That's what you're doing.
00:24:52.799 --> 00:24:55.200
You're delegating some of your thinking to something else.
00:24:55.359 --> 00:24:56.880
So write your prompts like that.
00:24:57.359 --> 00:25:07.359
If copilot is handling the retrieval, summarizing and drafting of our information, what does the core skill set of a knowledge manager look like in the next three to five years?
00:25:07.599 --> 00:25:16.480
I'm gonna be honest, this was the question that gave me nightmares uh on getting the right answer because I I think there's so much fear, you know, you touched on it in the last question.
00:25:16.559 --> 00:25:23.599
There's so much fear about AI and changing jobs within there, and we're used to doing things this way, you either got to change.
00:25:23.759 --> 00:25:26.960
I I think there's an awful lot that will remain the same.
00:25:27.200 --> 00:25:39.119
I think it shifts the knowledge that certainly the knowledge management area from being more of a librarian and police officer to you must do this, you must do this, to more of a coach.
00:25:39.279 --> 00:25:43.839
It's kind of nudging people rather than setting explicit rules from within there.
00:25:44.160 --> 00:25:52.480
So kind of defining strategies rather than policies, defining ways of working rather than rule sets from that.
00:25:52.640 --> 00:25:55.440
So I think we'll see people start to move towards there.
00:25:55.599 --> 00:26:07.359
Since the beginning of the year, the the use of AI skills that I think has evolved from Claude and Anthropic is becoming bigger and bigger across the AI state, and Microsoft's heavily heavily building on these as well.
00:26:07.440 --> 00:26:11.279
So thinking about defining what you do as a as a skill.
00:26:11.599 --> 00:26:16.480
If you haven't seen, I think just Google AI Skills Anthropic has got some great websites on it.
00:26:16.640 --> 00:26:19.119
Microsoft is growing into these as well.
00:26:19.279 --> 00:26:24.240
I think skills.sh is another good website with a set of those.
00:26:24.400 --> 00:26:27.200
Have a look and they describe how you do things.
00:26:27.519 --> 00:26:44.559
Thinking about your own work and building up your patterns of things like that, how you review documents, how you create documents, defining that style as a skill means that you can kind of again delegate some of the actual work from it in the same way that people thought they could just get Copilot to write a presentation.
00:26:44.640 --> 00:26:48.720
You could say, give me a presentation on how copilot can help me.
00:26:48.880 --> 00:26:53.839
The reality is you want to go to that level of depth of what sort of thing, how do you define this?
00:26:53.920 --> 00:26:54.960
How do you break this down?
00:26:55.200 --> 00:26:57.680
You can define those ways of working with skills.
00:26:57.839 --> 00:27:00.240
So getting ahead of that will be absolutely key.
00:27:00.480 --> 00:27:04.400
Worker needs to have a look at the way they write and capture knowledge.
00:27:04.640 --> 00:27:10.640
I think more atomic, smaller bits with small bits of metadata, we're seeing markdown files go down well.
00:27:11.039 --> 00:27:15.920
Ways that work for AI more than for users is going to be the effective way.
00:27:16.000 --> 00:27:21.759
Because AI is going to add that layer of that knowledge that you don't need to browse and search in the right way.
00:27:21.920 --> 00:27:23.440
It's going to make it more effective.
00:27:23.599 --> 00:27:26.559
So building out content in the right way will be key as well.
00:27:26.880 --> 00:27:29.839
But what is Claude and what is Anthropic's?
00:27:30.240 --> 00:27:34.480
Athropic is another AI company, much like OpenAI and others.
00:27:34.640 --> 00:27:36.319
Obviously, Google has its Gemini.
00:27:36.799 --> 00:27:42.559
Anthropic has their models and they've bucketed those under the Claude brand within there.
00:27:42.799 --> 00:27:50.720
Uh and I'd say they have led the way in terms of this next generation of large language models that can think and plan and take action.
00:27:50.880 --> 00:27:57.920
So Claude co-work originated that was that's kind of what Microsoft has based their co-pilot co-work.
00:27:58.000 --> 00:28:01.440
I I would say working with Anthropic, they haven't just ripped off the idea.
00:28:01.599 --> 00:28:05.200
And they they are building a lot of really great AI content.
00:28:05.279 --> 00:28:07.680
I think they are leading a lot of the thinking way.
00:28:07.839 --> 00:28:12.079
Open AI, if one kind of thing, oh, they're way behind again um on their.
00:28:12.240 --> 00:28:15.200
I think they are catching up again, keeping up with this race.
00:28:15.359 --> 00:28:17.200
It's a full-time job for many people.
00:28:17.359 --> 00:28:20.720
So let let others do that, let AI do it for you as well.
00:28:21.039 --> 00:28:21.359
Agreed.
00:28:21.680 --> 00:28:30.079
And the last point, I would definitely say in terms of the knowledge managers' skills in the next three to five years, is that actually it's around us having influence.
00:28:30.319 --> 00:28:30.559
Yeah.
00:28:30.799 --> 00:28:41.440
Having our own personal branding, having those conversations with the organization to say actually this is the value, this is what we need to be thinking about, and really having that standpoint of that human value.
00:28:41.759 --> 00:28:45.920
Our human value because a lot of people are so worried about the technology.
00:28:46.160 --> 00:28:49.920
We've still have that human value, and I see that very much so.
00:28:50.079 --> 00:28:54.559
We just need to adapt our humanisms to the to the new world.
00:28:54.960 --> 00:28:57.680
It's been absolutely sensational to have you on the show.
00:28:57.920 --> 00:29:00.720
How can listeners reach out to you if they want to know more?
00:29:00.960 --> 00:29:06.319
Um, you can find me LinkedIn is the place I'm on the most if you look for Kevin McDonnell MVPs.
00:29:06.640 --> 00:29:12.319
I also have a podcast with my friend uh Zoe Wilson and Gary Trinder called The Copilot Connection.
00:29:12.400 --> 00:29:14.640
Uh so we do it around every two weeks.
00:29:15.119 --> 00:29:17.200
Thank you so much for your time today, Kevin.
00:29:17.279 --> 00:29:18.000
It's been brilliant.
00:29:18.240 --> 00:29:18.799
Not a problem.
00:29:18.880 --> 00:29:19.920
Thanks very much.
