so welcome to this live broadcast from Copenhagen on a I would say very exciting day for a Microsoft partner we uh built a lot of excitement early this uh year when Microsoft announced everything about the new Microsoft co-pilots coming out later this year and uh I'm happy to say that we as a partner in fellow mind have been working with the co-pilot these recent uh weeks and it uh it certainly is very exciting um so let's have a conversation about that and and in Ai and in general and I'm happy to invite my colleague uh flying in from Finland on today's talk yeah hi everybody I am mikoni I'm uh from Finland I am our fellowman finland's uh competence lead for our uh apps and Innovation business unit and actually I'm super excited to be here as well uh so glad to finally get this thing out that has been uh talked about a bunch over the last six months or so yeah indeed yeah and uh me I'm Henrik Kim Christenson uh CCO here in fellowman Denmark also very much working on new technologies from Microsoft I've been doing that for many years and of course the Hot Topic of this year is the whole AI story from Microsoft and and specifically these uh co-pilots and um the idea about the today's session uh is basically to kick off a whole uh range of broadcast in these coming weeks uh we thought it would be make sense to do the the basics uh today and then you will see uh next week uh there we'll have a a session deep diving more on the Microsoft 365 co- pilot U now that it it is there then the following week uh we go maybe a bit more technical doing some open AI use cases um and we'll maybe also talk a bit more about that today what that is is then uh come the week of the Microsoft ignite of course I guess we all excited to to go there either in person or digital uh we have this U I'm I'm one of the the people that are fortunate enough to to go there in person and we are planning for a uh how should we say live from U ignite uh session where we will try to highlight what we we learned from the day one so to speak there going to be so much information and uh it's basically our job to you know pack it down and and provide the most valuable information out there yeah yeah I I I I think most people that have gone to these uh conferences in the past know the almost frustration of being there and just knowing all the the things you're not hearing but a good thing is that it is H today with with how Microsoft sets up these conference are very much digital so so much content is recorded and and shared actually almost as it as it happens so it makes so much more sense to to then also have a conversation about it Wireless happening and and maybe uh direct people in in in a certain direction and then to be announced we'll have a what we call the episode five uh the aftermath of Microsoft ignite I I guess and our current planning shows that it will probably be the the week after ignite the ended but today's uh session uh sometimes and even for us I guess it's a it's a it makes sense to to maybe step back a bit and and and just think about what is the context of this what is the the foundation what is the platform we're building on uh it's also a time for us to maybe say okay we've been working actually with quite some customers uh while we were waiting for for the for the co-pilots working with some of the tools and and features that that are at hand and and that is uh in in summary what this session is about uh we we would like to ensure that we on a common ground so to speak have a common understanding about what is AI and especially the AI in in in the Microsoft space um then we'll also have some open questions maybe uh we we'll have our take on it will AI take my job and and so on uh but hopefully also really talk to what is the business value uh what are the opportunities out there and I guess in the end our key message here is get started oh yeah oh yeah definitely now it's now it's actually the perfect time to do that and we'll I guess later on also uh give a give a message in in how simple it is to you know move in this landscape right now and U and uh but I guess yeah let's get started and and you know dive into the first topics and what we are talking about today exactly yeah so talking about AI we uh Microsoft uses this slide and we basically just copy pasted it because we actually think it it it talks to a very very important message first of all being that AI That's that's not anything new in the terms of artificial intelligence I I think you could you could say that the ideas about it was formulated back in the former Century in the 1950s where they saw the potential of computers at that time uh they also then formulated what we could do with computers I guess a lot of the the Sci-Fi actually derives also from from that period in terms of where where could It Go um then we saw the the maturing of machine learning and I think the the limiting factor here for many years has actually been the compute power available and that's what this slide maybe also talks to that you can say stepwise from artificial intelligence over machine learning to deep learning and now generative AI it has accelerated right uh it is really kicked off in these re very recent years and um you can say working um in any productivity tool within the last 10 years even your your mobile phones and so on youve seen glimpses of this actually working in in some areas but but I guess it's also fair to say that we were all maybe taken by storm and surprise almost when uh we we saw this very first publicly available and working generative AI in the form of of the um chat gbt yeah and I guess really to summarize the slide I'd say that if you've had like previous previous experiences with a you've maybe had a bad stigma or or maybe some bad bad interactions with AI where you've had to spend a bunch of time Gathering data and uh and then mushing it up together and and learning based on that and getting some analysis on your data um that those have been big projects and they have been quite you know unstable whether they will provide any benefits or not now we are not there anymore we have jumped into a new scenario where the machines understand our natural language and that's I guess the the biggest thing that they are they can Now understand us and generate new things for us quite simply yes yeah so so you you could maybe very simply say it's not such a geeky project big project to get productive with the AI anymore exactly I uh in in in recent uh weeks I've actually had a a conversation with um person that is now retired but working one of the major companies here in in Denmark and and I talked with excitement about the AI now and he was like yeah I I think we've been there before haven't we and he mentioned a competing product through Microsoft Watson and uh how they've spent way too much money on that project and never really saw how they could harvest the benefit of of that really yielding the the results so or delivering on the Promises so I guess we're here to say that we we strongly believe that investments in in this area use of this now is something that will uh yield results um but again we will get back to also some examples on that but looking at it in this way then um uh we we basically went to a whiteboard a couple of weeks ago here in in in fellow mind and and and brainstormed a bit about okay is that is there a way to maybe visualize uh this development and and also in that way try to to relay how things are building on top of each other also because there are a lot of confusion here we even confuse ourselves I think sometimes internally in in in all these classic abbreviations and um and and and things that for even from a Microsoft perspective is almost branded the same named the same thing but it's very different things and at least in terms of what the the user U experiences but reading it from left to right on this slide you could say okay so we we had the notion of the the AI we have the the notion of machine learning a very important point is here when we're not going into that discussion today that this is not the artificial general intelligence we're talking about here because I guess more smart people can talk more to to that than I can but but but basic that's not what we at here that's why the ADI is at at the bottom there but then you could say AI came the large language models the lnms that kind of it's in the shadows it's uh in in behind everything we're talking about here yeah it's of a top level term for these uh models that we are using uh there could be models from uh open AI such as the GPT model there could be La Lama from uh meta and all of these it's sort of a top level term for most of these generative AI tools Yeah Yeah so basically what you're saying is that there's not a single tool out there it's actually a huge exactly and that's probably something that you've heard before you've heard people talk about these GPT 3.0 GPT 3.5 F gp3 three 4.0 and you've heard about Lama and all of these uh different different terms uh so these are basically the um model behind the U chat interfaces that we use to type our questions to that's the model to do the generative AI ability to and then we have the tools like chat GPT which is the the actual interface the where you type your questions too so that's typically where people uh you've heard these terms so how do how do they differentiate between them themselves uh typically people use the term open AI right now which is obviously a company uh and and company that has built the GPT services and and chat GPT it's sort of a synonym for the chat GPD to use the word opena I mean um people use it so much that you're probably not wrong using it but uh technically no I even sometimes use the examp it's kind of like U we we saw Google maybe more than 20 years ago 25 years ago and and uh we we then in in the years to to come basically changed the whole notion of you're not searching the internet you you're Googling right yeah and maybe we're at a point now where almost chat gbt from open AI is now taking that terms of of that conversation with the computer is chat gbt EXA even though it may not be chat gbt exactly yeah but again to the to to the to the slide here uh so as you also pointed out open AI is basically a company uh you could say open AI is then the company also Microsoft invested in a couple of in in recent times um huge investment but they in that sense bought access to their services or their models everything and that's why Microsoft is now able to say so we now have Asher open AI so so all the the things you see from open AI is now also available in Microsoft and why is that important why why is that interesting well first of all it's because they can then start to build their own services and end user facing services and we we'll show that in a minute uh but it all it's also important for for businesses out there for for us as a partner that we are now able to buy access to the compute power on the ai ai Services we're able to speak to uh where did this run where is my data stored is it worked on my data because it's with within the boundary and and the well-known setup of Asher uh simply um and and you'll later today today give some examples of of this your own chap chat gbt so so I'll just leave that there and I feel like the uh the Microsoft take the co-pilots that have been introduced I think they are like the biggest technological productivity boost that we have been introduced in in in our well lifetime I guess could be the right or wrong thing to say but it's really the biggest one of the biggest technological boosters we have but basically we're standing on the shoulder of saan Adela announcing that hey first come the internet then came the iPhone and now we have a AI so yeah I think it's a fair comparison to make um just just briefly because we're now naming a lot of things we're not showing jet chat gbt we're kind of expecting that you may have even tried it out there but I think one of the confusions and frustrations with chat gbt sometimes is that it it it you you ask it a question and then it generates something and and as we tend to say it's a 100% conf confident but 70% correct because it's it's it's just generate generating thing but one thing you could do is um instead using um Microsoft's uh offering now called the the Bing chat for Enterprise and basically what that means is that if you have any Microsoft license for Office 365 you will also in the Microsoft Edge have this new icon in the top right corner some of my colleagues have showcased this before but it's just merely to to show this is that now I did a fun thing here saying make me amazed here and then it just went to the internet and then it started to amaze me and I was just about to stop it because you just rambled on with all kinds of things that are nonsense but I the most important Point actually being here at the button that it is now not just generating content but it also shows me where did did this uh this come from another version of this and that's when we we turn maybe even more to uh being productive in the context of your own content is then that I have another window open here um where I now showcase the Microsoft 365 chat um so it's somewhat the same as Bing chat but the the the very important change or CH of functionality here is in the back end uh it now looks as my own data so this in this example I have here uh I have no a meeting coming up with my colleague in um in in the Netherlands Rick so I asked that question and it basically comes down to there by the way is the next broadcast 16 I'm doing that live from Seattle with Rick but it also States some other information here and even again link I can link click directed to that and be taken to it I think that's a a productivity G that that's exactly one of those so if you if we think about uh what we have done is basically we have the first chpd and you use it use it to you know ask your basic questions it doesn't have access to live internet okay the next step we use chat uh bing Enterprise uh chat uh which is connected to internet you could ask it questions about any live event that has happened uh right now and is found from the internet and it's also secure because it is uh for your Enterprise and uh then we have the M365 chat now it has the tools to connect to your specific data as we could see whether that be a topic about what meeting do I have or what kind of conversation I had in the emails or uh could you find me a document from my SharePoint structure or something similar to that helping you get on with your day and helping you create and find your content and maybe it's it's obvious but but the the funny thing is you don't even know what you're looking for right because in your in your chat you're just searching for something and it could be a document it could be a conversation it could be an email it will know everything and it will understand your question and direct you in in the right context we we all know the frustration of going back and forth between all our different office tools to find we know it's there yeah but but where final short example here is also in teams then we had a meeting internally here in felline where we discussed okay what is our experience with the Microsoft co-pilot up until today I asked it to to to to listen in on the meeting so doing a transcript of the meeting and then I ask okay what were the main topics it actually Ables to these were the main topics I can actually agree to that is even able to again link back to who said that and where is located and then also said please also summarize all the topics and for each topic name a responsible person then it came out with all these different topics it even named who who said that and in in in what context again it's able to link back to the transcript and show it but it also shows me some topics with the no owner I was a bit puzzled with this first one but it was a St statement from one of my colleagues more as a a shout out but U yeah a bit funny okay so um going back very briefly here to I'm confusing myself here a bit I just wanted to show you um um the co-pilots um it it is maybe again a bit obvious but but but talking to this how the co-pilots work and as a generative AI it can sit and fit into the tool that you using now I showed the Bing Chad Enterprise and the co-pilot in in Edge I see it in teams but of course it's also word it's a PowerPoint word here in Excel and and and so forth so in summary and as a takeaway here also from from from this first session is let's get started because it is there there now it's it's not something that is coming this I think the slide is actually summarizing it quite nicely even just with a computer almost and a Microsoft personal account you're able to get to Bing chat then just understand that it's more in the category of uh other free services that they may use your data um whereas if you then move to the Bing chat Enterprise your license then you have this uh commercial data protection enabled it's your data it stays with within the boundaries of of of your setup and then come the Microsoft 365 co-pilot we'll have a lot of other sessions and and talks about what that is in in more detail uh even today just for the Danish AUD audience there is something at at 3:00 otherwise next week and then uh also from our internal conversation one uh things to be cautious about uh when when doing these talks is uh Microsoft decided to to go with co-pilot for a lot of things so so they are they are branding and naming a lot of features here um to to just be aware when having this conversation also in your own organization what co-art are we actually talking about and then uh as a technical person and also this is also part of the excitement right that the toolbox is actually there one thing is the things you can license and buy as as as software as a service from Microsoft but it's also possible to really deep dive in and I think that's one of the things you've been doing a lot in Finland yeah we've been doing a bunch and and we are super excited that we just have the tools uh right now to push these things and what it means is that actually companies are able to do quite complex scenarios with AI generative AI right away right now and that's the biggest news news relating that we have the tools here yes okay but yeah uh I guess uh generally like uh the like six months ago or what nine months ago whatever it ago it was when we first heard Microsoft announce these tools and and and really arm us with the Azure capabilities uh where we can do our own co-pilots that are safe and secure uh where we have these co-pilots we started already at fellow mine to you know Branch out to see like what can we do with these tools what is the status the state of AI today and uh and going going uh to that uh AI is quite what we what we found out is that AI is like quite near uh us in our everyday work basically all of these tools that we are now getting they are quite like Plug and Play Almost where you can just have your end users uh have your end users just add these tools to their uh everyday work life as in having have them in teams have them in uh off your office tools and so forth and uh they are mainly about pushing productivity like whenever you right now if if you are at the SC you think of a scenario where you are wanting to find information from somewhere what do you do you basically start to search it and uh ask around from people and water cooler conation maybe yeah exactly but that's by chance right that you pick up on on on something yes exactly and uh and what these tools do they basically allow us to be so much more productive by finding the information for us helping us gather information and giving us understanding into whatever we are doing and and this is basically something uh that uh there has already been studies uh relating to this technology there was a study made by Harvard Harvard MIT Boston Consulting Group that really drilled down on the uh work of knowledge workers and uh they they found out that using this technology the knowledge worker were able to gain a 20 to 40% uh productivity boost in their everyday work and I can totally believe it if we like if we relate this number to the tools that we have now uh seen so we can get help starting off uh a presentation about uh let's say our like U our deliveries for the past six months we could have a presentation made by AI from that it might not be 100% correct ride of the bat but it's maybe 60% or 80% ready and then you just move around a bit again you're productive or you find the document you're looking for uh you remember what you had in your last meeting what kind of uh tasks you had there all of these things they all Point towards having better productivity and as you can see in the image it's it's really really about just chatting in one user interface reaching out to your other tools uh from the same view and moving forward with in your daily process what I also find a bit mind buckling sometimes is that now you you you see this is a a demo video and Microsoft has has of course set up things to to make to make that available but that that's a creativity we should be now showing both as integrators but but also you you as in companies out there that that think about where are your content today and all the context switches you're making today maybe to gather that information and then think about how can we bring this together in the old days so to speak we index and we did searches but it really didn't work that well but with with the um the generative AI available again we can ask the question it understands the context it understands all the sources and then you can really start to bring content together exactly and and the the biggest part is whenever we see these they like feel so futuristic but in reality they are already here so this is something that you can already uh work in and and one thing that we quite often hear about are some typical concerns around the AI like uh if I use these tools will will the data be used to you know be train the next models and will other people have access to my data uh and and these are these are something that we quite often hear and uh and my my answer is always that once we use the co-pilots or we use the Microsoft tools uh available they are all all already built with uh the security that they won't send the data somewhere it's your private chat GPT that you can use for yourself so even security is also compliance wise that you can say that the practice you have around the Microsoft cloud today also applies to your uses of the AIS AI tools in the Microsoft yeah it honors the rules uh that what you can reach manually is what the robot or the AI can also reach so you couldn't I couldn't ask uh for for my boss's salary uh from from that and uh and have something else is wrong then yeah sure um then we have a bunch of Humane concerns can I trust its answers will these tools take my job well we'll see uh but uh but to first answer and can I trust his answers this is something that the readymade tools typically it's somewhat some somewhere in in the middle you can trust it but then you have to be cautious at at some time uh sometimes as well but when we create our own custom models which I'm getting to uh in a second um we can also measure or change the amount of um H hallucinations it can do to a degree we can say that you have to be 99% sure or otherwise you say that I don't know the answer to this one okay in that sense yeah yeah so it's not being too creative and just answering something yeah and also will these tools take my job this is something that we hear really a lot uh and I think it does have an effect that when we are more productive uh we don't have to spend time on these typical scenarios but what we see is that then instead we'll be able to have and and spend more time on things that actually matter rather than searching for the information makes sense and ethical concerns finally uh like is there a potential for misses Microsoft is really pushing a lot and they have also released statements that they are pushing that the base model behind everything you don't have to do anything but the base model is built so that it won't have like racist comments or sexual comments and so forth so this is something that really like drives in that this is usable but one thing that I really have to point out that we have these co-pilots and they are mainly for productivity gains uh but we also have the opportunity to build our own custom models and and how I differentiate these two is that we have on on on one side we have these uh productivity tools the microscope that everybody can access and then the other side we have these custom tools that we can basically build these AI features on top of your core processes your core uh services or products and really differentiate yourself in in the market these are tools that are customade made for you and custom made to really bring you the competitive advantage and this is based on your own data once again safely and securely and it could be if we think for example here we'll we have another episode where we go over all of these use case scenarios but for this specific if you think about a customer service agent we could have a self-service portal based on your own data that you can just ask questions about uh your own service when my about some details about your contract or or or so and not have a service agent there and this is not your typical chat user interface this is something that can act as a super something that people would kind of recognize as what they've for instance experienced with chat gbt or being a chat Enterprise that is able to understand the question you have a conversation about it and the answers EXA then Falls within that the topic uh that that you you're in here for instance something about some Machinery or whatnot exactly and um and now that we are um approaching our time um also we've we've done a bunch of work with many companies and we've built a maturity model around uh where companies are right now in this AI space uh and most mostly we see companies they are at the middle segment where they've heard about chat gbd theyve May been using it a little bit and only a few of these companies have already moved towards piloting these CH gbt use cases and and and co-pilots and uh and custom open AI uh SE ments in their own company with their own data and then basically also this this also relates the message of it's it's time to get started because the tools are there we have a practice always already established around a lot of these things and we see a lot of our customers now really moving ahead and and back to the uh will it take my job well the edge here is maybe that you should be concerned about the the colleague that figures this out and is much more productive about that you can even raise that to on a company level you should be concerned about the competitor adopting this and being more productive being able to deliver faster than you can exactly and I think like one key Point uh is exactly how to get started one of the easiest things to get started is if you're worried worried about your own data and the use cases one of the easiest ways to get started is use your own documents that you have gathered for five 10 or 15 years uh already we can just push those to the AI and you already have a readymade AI bot that can answer your HR questions your product questions your contract uh questions what whatever you have documents on we can push these to the bot and it's as easy and uh we will share these links afterwards but that is basically a call up call out for Action also here that uh We've try been been doing this for for some time now we have some good practices some good uh packages so to speak to to get you started because it's also important to maybe step down from the very generic explanations and get into your business and your business process and see where we can apply some some Ai and I think that is maybe the the closing remark also thank you for listening in on today's session uh please reach out to us and to to fellow mind in general again what we showed just before is on our website and there are some contact forms in there and then um hopefully also with the today's kickoff of this series I I hope that we will see some of you again uh coming up next week but and the weeks after but um yeah that is that's all from us and I uh will Deep dive much more on these topics we cover today uh on the later that's right yeah much more thank you perfect thanks for much