Up next in 10
The Cloud Show with Magnus Mårtensson ft. Robert Folkesson - Ep. 20
985 views
Feb 23, 2024
Join this live session with Magnus Mårtensson ft. Robert Folkesson for the next episode of The Cloud Show with Magnus Mårtensson on January 31, at 01:05 PM (EST). The show is about cloud leadership and all the important questions relating to cloud projects. Certainly, many matters when a company is going to and wants to be successful in the cloud, are about technology. However, there are many additional matters, adjacent to technology, that we also need to tend to regarding business strategy, human resources, organizational change, planning for a technical cloud approach, and many more questions. These conversations are critical for a healthy cloud and for a swift and accurate cloud approach. GUEST SPEAKER Robert is CTO of Active Solution - a consulting services firm based in Stockholm, Sweden, focusing on helping customers build cloud-based solutions. As the CTO of Active Solution, he is responsible for the overall technical and architectural strategy for the company. He is also responsible for constantly keeping our skills and knowledge up to date through cutting-edge projects, training with the world's best experts, and knowledge sharing between colleagues and projects. he is also a Microsoft Regional Director. That doesn't mean that I work for Microsoft, but that they have recognized my knowledge and community leadership by awarding me the title.
View Video Transcript
0:00
Hello, hello everyone and welcome back once again to the Cloud Show
0:05
And today, as always, we have a wonderful superstar as our guest
0:10
We're going to talk to Robert Falkison, and he is very much involved in this AI space right now
0:17
All the companies, all the customers of his company are calling for AI services
0:23
How do we do this? How do we get it into production? And what should we do with it
0:26
So we're going to talk about how to roll AI into the enterprise. What do you do? How do you do it
0:32
And with that comes a huge challenge in terms of data sovereignty
0:36
How do we manage data regulations and moving data? And are we allowed to use the cloud services and so forth
0:43
And of course, while that is taking so long, how do you avoid in an enterprise to catch that horrible disease called Shadow IT
0:53
where people are just like, I'm not going to wait for AI. I'm just going to do it my own way and you get a lot of shadow IT
0:59
How do you avoid this? So all up on AI and Enterprise with Robert Falkison
1:04
Welcome to the Cloud Show. Hello, Robert
1:18
Hello, Magnus. How are you? I'm fine, thank you. How are you
1:24
I'm really good. I'm really good. This is such an exciting topic that you brought for us today
1:28
But first, let's check up on you. Who are you? And what do you do
1:33
Yes, so I'm the CTO of a company called Active Solutions. And we specialize in cloud development, heavily focusing on the Azure PASS services
1:44
And also, we're doing quite a lot of AI development. And actually, we have been for quite many years
1:52
I think we did our first cloud-based AI. AI solution some 12 years ago or something like that
1:59
Right. So yeah, cloud, pause services, AI, all coming together. And multiple NDPs work in your company and have a really solid base of, of, of
2:11
hardcore deep dive tech people. Yes, exactly. Absolutely. And, and right now you have, what is it, is it only, only one AI MVP or is it two
2:25
It should be two now, I guess. Two. With Alan also switching to Erienne, yeah, yeah
2:33
Yeah, that's really good. Peter Erna Holm as well. Yeah, yeah, exactly
2:38
And so you're based out of Stockholm, Sweden. Yes, exactly. Right. So very happy to be here at Cloud Show
2:45
Thank you very much. Thank you very much for being here. Excellent to have you
2:49
And I know that when I reached out to you to talk about these things, you know, get on
2:55
showed as the star of the show. I knew that we're going to talk about AI. I was like, he's going to say
3:01
AI. And we did. You said AI. So yeah. Yeah, easy guess. Yeah. Now, but it's, it's interesting because
3:09
as I said, we've been working with it for quite a couple of years. And now that the generative AI
3:16
and cognitive services have exploded and there is a lot of interest, we've seen the last, the last year
3:25
seen a lot of experimenting, a lot of proof of concepts, but the last couple of months
3:33
we're now seeing companies, organization, enterprises, trying to figure out how to get this
3:40
into production, how to really start using it. And also, there's kind of this huge demand from
3:47
the business to be able to use these kind of services, to be innovative, to come up with new ideas
3:55
to optimize their processes. And there's a huge pressure now on the existing organizations to come up with ways of
4:04
allowing this in a good way. Yeah, definitely. I get it. I mean, basically it's everything we did before, all of our great products and services we
4:13
did before, except now with AI. Yeah, exactly. And the services are out there
4:20
Yeah. People can sign up for subscriptions on a lot. lot of different services now and do stuff that you couldn't dream about just a couple of years ago
4:32
And of course, the organizations and different departments, they need it. They want it
4:39
They need it for the competition. They needed to innovate. They needed to go faster and have better time to market
4:47
So what we have been seeing is that either you try to structure it and allow it in a structured goods
4:55
safe way where you can make sure that you follow the regulations GDPR for example if you in EU or you will have shadow IT that will pop up and you will have kind of in the beginning of the cloud
5:12
services exclusion. I think there's a parallel to that. Yeah, and that's a really
5:18
that's a really tough challenge to have, I would say, that if it's taking too long to do it the right way
5:25
we're going to do it anyway. Yeah. In any way we can
5:30
Yeah. The marketing manager will flesh out his card and know her card and kind of sign up anyway
5:39
Yeah. So what is the sort of the, I know it's a big, big thing to answer in a short time
5:46
but what is like the structured approach? What does it like entail at a high level
5:51
How would you go about it? If you're a large company, you're going to use AI now
5:55
and how? Yeah. So now I'm heavily focused on the Microsoft Azure platform
6:05
So this is going to be colored by that, of course. The other cloud platforms have other ways of doing it
6:12
But on the Azure platform, you have this concept of a landing zone
6:18
And the landing zone is a way, it's a concept that makes
6:25
it possible to deploy resources in a way that are secure by default? So you set up this zone where you
6:35
deploy your resources, where it already has cost control, it has the monitoring, it has the
6:42
security, security boundaries, it has the secure access it might need to internal resources
6:51
Everything set up the way it should be. And there are excellent reference landing zones
7:00
that you can use from Microsoft that will show you how to do this
7:04
So that is one part of the puzzle that you need to be aware of
7:09
And then there is also the possibility to use these services from in Azure from Microsoft in a way that makes it possible
7:20
to decide what region in the world will your data be stored in
7:25
and guarantee that there will be no training on your data, your customer's data
7:33
It's all going to be in your hands and you will have total control of it
7:40
And they will make sure that if you store it in the US
7:46
it will be stored in the US and not go anywhere else. So that is a big
7:51
That's a big thing that that sort of the data sovereignty and the data privacy, right
7:56
So that a company's data is their oil, it's their bloodstream, it's their life
8:03
Yeah. Of course, if that went to someone else, it would be catastrophe and it wouldn't work at all
8:10
And it would be, well, it would destroy the company that offered the AI service if they leaked
8:16
the data of the customer. Yeah. And so that is kind of making it possible to have
8:25
a technical IT platform ready to start experimenting in and start move fast in, but under
8:33
very controlled and limited forms. Yeah, you get a box, but the limitations aren't really in the form of limited services
8:47
You can still use the open AI services but inside of the cloud platform. Yeah
8:53
Yeah. So we'll still be able to use, for example, not chat, GPT, but the GPT models
9:02
Yeah. And the Dali models, but in a very controlled environment. Yeah
9:09
So you were saying in the start of this conversation that your company are focusing active solution
9:17
focusing a lot on platform services. And that's been the case since the dawn of
9:23
of the cloud that this has been a focus like you should be using platform services you shouldn't be
9:27
using infrastructure services and i think that is the the key thing that has happened here i hope
9:33
that you agree that we have turned a corner and come to a place where all of these
9:39
AI services are exactly that they are platform services you just connect your data and you start
9:45
using it yeah exactly so and that that is also true about these services so what
9:53
Microsoft is doing is they making the same services that you have available or the same models available as services inside of the Azure platform And they also providing the guidance with the cloud adoption framework
10:13
that you can use to plan and plan your organization for it and plan your strategy around it
10:22
And they're also providing guidance in form of the well-architected framework that will help you live up to all the regulations that you need and all the security audits and all these kind of things that you need to live up to as an enterprise organization while still being able to provide the business with their ability to be really far
10:52
and try out these services and build proof of concepts and even go to production with them under safe um
11:01
under safe uh in a safe way yeah yeah yeah get it um so so let me let me ask me this so this seems like a lot
11:11
right there's there's there's cloud adoption framework there's with the with the landing zone concept
11:16
there's well architected framework and then there's data regulations and all the things this actually does
11:22
like a complicated thing. And I suppose this is an open goal. How do you start doing it? Well
11:26
you hire your company. Sure. That's an open goal. But I mean, how does a company approach something
11:32
like this in a way that makes financial sense and that they feel that they are in control of security
11:40
and so forth? It's huge. So, so yes, but I would say that, for example, the cloud adoption framework
11:47
that will help you guide your business around your cloud journey, including the AI part
11:56
It's not really complicated. It gives you a lot of guidance. So Microsoft has kind of package this up as a guidance package for your organization
12:08
that you can take part of, it's available online, of course. So it's really easy
12:17
to get assistance and get guidance from the cloud adoption framework. It's not hard
12:25
It's kind of made to be able to guide you in these questions
12:30
And it's the same thing with the well architected framework is basically best practices
12:37
and checklists and assessments that will guide you through the whole process of doing it
12:45
And then of course the implementation details, you will probably need help from people who have done it before
12:53
Right. You need to find some experienced people. Either, of course, hire them if you can find them, right
12:59
Yeah. Or just get them in some way to guide, help the process along more quickly and more to the point, exactly
13:09
It's a good point. So about this problem, then, the flip side of the course
13:15
point, if it's taking too long or if the company isn't really on the ball to set up this secure
13:21
and safe space called a landing zone or what not, right? The shadow IT thing, what do we do to avoid it or how do we manage that if it happens
13:32
Have you seen some bad scenarios out there? Yeah, yeah. I mean, it is hard. If you don't start moving pretty fast
13:45
and try to tackle this problem, it's going to pop up, especially in larger organizations
13:52
it's going to pop up. And so the advice is to start looking into this now, right away
14:05
How can I make the cloud journey, including the AI parts, as good as possible in a structured way
14:15
the help of all the guidance that is out there. And I mean, there's, it's, it's very hard to pin down shadow IT services being bought from
14:29
people with, like, with, with access to a card. Yeah, with a credit card. Yeah
14:36
Yeah. It's very hard to get a grip on. So what you have to do is kind of turn it around and offer
14:45
the business a way to be innovative, to be faster to the market to let them use these services
14:53
but in a controlled in good way. That a good point yeah So in terms of that in terms of like guardrails or setting up protections it be things like for data sovereignty we set a policy that says you can only deploy
15:11
databases in the following regions. Exactly. Exactly. So you will give your business a way
15:18
to use these services, but under the controlled form that you decide upon. And that makes it
15:26
possible to live out through regulations and also have cost control. So you won't have any bad experience in that area, which
15:38
Yeah, that's a big thing as well. If you turn the knob on a service, on one of these AI services
15:45
and you're not really sure what you are doing. I'm just going to try. This thing goes to 11
15:50
I'm going to turn it to 11, but you don't know what you're doing
15:55
that actually make you really hard. Yes, or something gets really popular and you start getting a lot of requests
16:02
and you start using the services which are pay on demand in some cases a lot
16:08
And you don't have the cap on the cost set up properly
16:13
That's also, of course, an issue, a risk, a big risk. It really is
16:20
So what is the key controls to? to set a cap on consumption
16:28
How do you do it? Yeah. So if you use the services in Asher, you can set an actual cap on the actual consumption
16:37
But what you can also have is a good way of being notified when it starts reaches these levels
16:43
So you have good cost control built into the platform. And you can also set caps on individual usage of models, for example
16:53
So you can cap the usage for a specific use case inside of your organization and make sure that you won't have any bad surprises in the end of the month
17:06
Right, because if you hit the cap, you have used it a lot and then you have to start asking the question, are we using it properly
17:14
What are we using it for? Is there a business value enough from this cost
17:18
But even more importantly, you can get an alarm even before that so you know that, okay, now we've got a spike
17:27
It might be a good thing. Our business is going good. It's getting popular. We are getting hyped somewhere, which might be good
17:37
Then I can turn it up, but in a controlled way. In a controlled way
17:42
That's, I think, is good advice. And I think that is what people need to hear now, that they are a little bit scared of the
17:48
thing they don't know how to manage yet. And there are controls in place for you to slow things down
17:57
Exactly. Yeah. Well, what I really like you saying here today is the part where you said that
18:03
the only way to really combat or keep away Shadow IT is to ensure that you can offer a really good platform for them to work in
18:14
So they want to use that one. And the way to do that is to use the existing guidance to set things up in a proper way so that people will have quickly enough, will feel that they get the ability to start using these things
18:30
Yeah, exactly. And you can still set it up in a way that makes it easy for the business to develop their new services, to deploy their new services and to get it into not only into Provo Concepts and Experiments, but also into production
18:46
but all under secure and control forms and with cost controls and with policies that will control where the data is stored
18:55
And also making sure that your data is your data. It isn't going anyway
19:00
You can even have your own encryption of the data with your own key if you like to
19:07
Yep. So. Like a whole bring your own key thing. And at that point, not even Microsoft can get into your data. No
19:14
If they wanted to, which they don't, let's let me underline that they don't
19:19
But in fact, they couldn't. Yeah. It's going to be impossible. Exactly
19:25
All right, good, good. All right, brilliant. Well, thank you for coming and sharing all of this
19:29
This was a big lot for a short conversation. And if we have the time, we could go for several hours on this topic
19:38
Most definitely. Most definitely. But I want to thank you very much for being with us today on the Cloud Show
19:43
and thanks for everyone who was listening or who is watching this right now
19:48
And everyone have a good evening or day or wherever time is on you are in
19:53
All right. Thank you, Magnus. Thanks. Bye
#Distributed & Cloud Computing
#Enterprise Technology
#Intelligent Personal Assistants
#Machine Learning & Artificial Intelligence
#Management
#Networking
#Other