 Good morning. Good afternoon. Good evening. Good. Good. Whatever. I just say hello to everybody I'm very happy that you're all here for our first UPU DMA be innovation talk on artificial intelligence in marketing sales and service. I Really would like to say welcome to everybody I can't do it in all your languages because there are so many languages so many countries participating But I really would say I'm very happy and very proud to be the moderator of this session Just to give you some idea who you are and who I am. I'm Martin Itcher. I'm from Germany I'm a well 25 years veteran in marketing and CRM wrote several books about that and working for a long time in the German direct marketing association and as well I'm Member of the board of the European Direct Marketing Association and chairman of the board of the worldwide marketing association So I'm very happy and I'm very proud to be part of this Session we have today and I would like to give over to Olivier Boussard from the DMA be from the UPU Olivier it's your turn Thank you Martin. I hope everybody can hear me Okay, and welcome to you all to this To this session and on behalf of the Universal Postal Union's direct marketing adversary board It's really a great pleasure for me to welcome you all to this innovation talk As you may know the UPU's direct marketing adversary board is a group of postal operators and private sector stakeholders Who aims is to promote direct marketing? through postal channels As you know in view of E-commerce and the data revolution and the use of digital media channels Direct marketing is currently undergoing a profound transformation yet remaining a key activity for post In June this year We also the series of five online sessions Which address the main drivers of this new ecosystem in which direct marketing operates as well as how postal operators can adapt and Respond by building on new channel strategies We also address during those sessions the impacts of the current COVID-19 and they make Situation and the direct impact on marketing and direct marketing Those webinars were very very successful. We we had more than a hundred participants Today we even broke that record And the full report of the session for your information is available on our website or on request You can contact myself or at the boss are and and get the full report with very very interesting insights From those five online sessions. We had in June Given the success of those webinars. We had in June we decided to continue the conversation And to organize the support of Martin Our moderator a new series of workshops in the form of innovation innovation talks with global experts During our discussions in June a number of topics emerged as being of great interest to direct marketers and postal operators And one of those topics was the impact and value of Artificial intelligence on the way we do business as our marketers or as managers responsible for post operators marketing It's all this reason that we have decided to stop today our new series of webinars with this specific topic and to this hand We're very very happy to welcome today a professor Peter Genst Professor again. She's a global expert in digital transformation business intelligence and innovation and you will share with us today experience views and perspective on In marketing sell and services So on behalf of the UPU and the direct marketing at this report I would like to thank professor Gange for taking the time to be with us today It's a pleasure and I wish for a very enjoyable hour of discovery and discussion and back to you marking. Thank you very much Thank you, Olivier and thank you for this nice opening I'm I'm really proud that we talk top the results from June So in June we were very very successful and we were very proud that we had more than 100 participants Today for our two sessions today. We have 237 registrations and this is About an hour old so it's probably even more by now It went up over the last 24 hours from 200 less than 200 to more than 250 And all these registrations are coming from more than 68 countries And I just see that most of you already found the The comment and the sec the chat section of this presentation So I'm very happy that all of you are saying good morning. Good afternoon. Good day. Hello And I I really urge you to keep on so it's very good that we get some interaction and that we get some questions We are all about dialogue. We are about Not about monologue So I would really be happy that we get into a dialogue with Peter And that we get a lot of questions after his presentation just to let you know The whole presentation Will be in in english But the there is some possibility for the for the The desktop you can put that one in other Languages if you want to do that, but there is no direct translation here So the the presentation will be in english. So if you go up to the small world map, you can switch to another Language for the for the surface of your desktop Um, again, there's the chat pot on the lower right. I see that most of you already found that So whenever you have a question you can put it down there and just write it And then we will see that you wrote this question or this comment And I wouldn't be really happy that you send a lot of comments or a lot of questions What we will do is I will switch over to peter in some seconds and he will do about 30 minutes presentation Um, so that we are have enough time of 15 to 20 minutes afterwards to have a really really lively discussion Um, and I would be happy if all of you can can go into this discussion The the topic of artificial intelligence Is so interesting so that i'm really sure that we will have a lot of questions from you So that was it from my point of view. I would like to switch over now to professor. Dr. Peter Gange He is one of the most renowned experts on artificial intelligence He just wrote a book about artificial intelligence in marketing sales and services He has his own wikipedia page So I think that this does one of the things you should know that he is a real expert And I'm very very happy and we are really how not that he said, okay, I will do this with you So peter thanks a lot for being with us and I would like to give over to you It's your floor and we are happy that you are here So, yeah, thank you so much for the introduction invitation Great honor to be with you at least virtually And I'm really happy to share my thoughts on AI and marketing sales service And now hopefully you can see the screen. I think it takes some seconds. So I have to Share that so This looks perfect Looks perfect, okay No, that was n sharing So No, we can't see your screen anymore. Yes a second. So Okay, once again, I have to share that So now you should see the presentation. Hopefully we see the presentation and now we see it perfectly. You're set great The problem is always on the user side. So I'm talking about artificial intelligence. So I think I have to improve my Human intelligence skills. Okay AI and marketing sales and service I think there's no other technology out there that is overhyped and underhyped at the same time You know on the one side we have a lack of fantasy But we just think of marketing automation or self-driving cars And on the other end we have the science fiction stuff You know when honor schwarzenegger comes into play and talks about terminator and that stuff And I think we have to discuss the use cases in between And I think it's really important to fix expectations because there's so much buzz around the AI And sometimes it's not AI in marketing. Sometimes they use AI For marketing as a marketing slogan as a marketing claim, you know a lot of software vendors claims They have AI inside in order to better to sell their products So I don't want to to spoil the excitement. I'm a huge fan of artificial intelligence is a powerful tool I think it's important to fix expectation And hopefully I can give you some insights how we can use AI and marketing sales and service Okay, um, let's have a look at the agenda. I would like to touch upon the following things I want to start with lead customer prediction and profiling to know the next 1000 customers Based on AI Then there's a huge hype around conversation and AI So how can we use chatbot and all that stuff for marketing and service? Then I would have a look at the media planning and finally I would come up with some use cases how you can use AI for content marketing. Hopefully They are something for everybody in case there's something missing. I think we can jump on that having our Q&A session, okay, let's get started and Yeah, sorry for that You we have to talk about the excitement of AI and now I want to bore you and have to bore you with data Maybe you know the slogan data is in your oil. I don't like that. It's not that delicious I think data is in your bacon because there's no AI without data as simple as it sounds. It's all about data And I have a short contest for you guys It should be very easy. Maybe you have an idea. Who's behind this profile? It's a person you all know It's born 1948 raised in England Married for sick time two children successful business wealthy Like to spend holidays in the Alps and like dogs So it's not that easy to get this feedback in this digital environment So maybe I'm not sure if you have an idea who is behind If not It's our Osios born, huh? Okay, but guys have another profile for you. It's Seems to be the same profile. No, it's actually it's it's completely the same profile. It's not only similar So it's the same person But it's not Osios born and we all know Osios born do not have a twin friend It does not have a twin prince Brother, so it's another person and you all know this person And this is Prince Charles so maybe you know the the definition of customer relationship management treating different persons differently by how the hell Should I treat these persons differently because they seem to have the same profile? okay, but Facebook for example makes a difference So there's not only Facebook. They're a huge amount of data out there explaining customer insight What's the difference between osios born and Prince Charles? It's not only facebook It's instagram our websites. You name it. There are tons of data out there you can use To get better customer insights Even though that we all know that they do not write their own facebook fan pages, of course But they're very much telling and if you use AI you can try to track what's going on on the internet You can try to find Intense triggers and you can really derive insights about the audience And it's not only the demographics that are important It's also what's going on on social media. For example This example is related to the b2c area But it also works for to be to be area and I think you all know with where with the so-called Demographics or the thermographics the thermographics is The companies are related to a certain industry has a certain revenue and number of employees and all that stuff So this is company a Lafayette and the other is kiko milano a completely different cast Company so but if you look on the website if you look what's going on on social media You will find the difference between these two Companies so what i'm trying to say we have big data And I can help us to unleash the value of big data Today we have 44 Set up bytes with a lot of data Um And what I would like to show is that for example for companies you can capture so many data out there As I mentioned, so what's going on on social media? What is it web traffic on the website? What are the people looking for how they are connected with each other? There are so many data points you can take into account In order to derive customer insights in order to run campaigns that are fit with your Audience so having this said I think now we have to explain what is AI about but please keep in mind It makes and breaks with data Okay, uh another short Contest quiz for you guys. Well, I think it's pretty easy to answer What's the difference you might say? Oh, that's really easy So there's a muffin and there's a nice dog. So something to cuddling with something to eat. Hopefully that is a difference Um, but you won't believe there's no AI out there that could make the difference Provided we only have these amount of data but because of Big data we have tons of pics on insta on facebook you name it And what you can see 10 years ago human being where better when it comes to immigrant recognition But then we have a tipping point the AI tries to catch up And now AI is doing a better job when it comes to image recognition, but it depends on the data You need to have trainings data Um to use AI and that brings me to my definition AI as a study of how to make computers do things at which at the moment people are better There are a lot of tasks in marketing which we human being do a great job But AI is catching up. So that's a definition It's not that fixed definition. It's a rather dynamic definition and maybe you remember 1996 there was Gary Kasparov and it was a world best champion player and it he lost against an AI which was created by IBM. It's called deep deep blue He never lost the game, but then he are lost again He'll against deep blue nobody could imagine and the funny thing Once AI reaches a level of maturity and could win against A chess player. It's not called artificial intelligence anymore and then 2016 people thought It's AI would be great when it would be possible to win against lisa doll I'm not sure if you know lisa doll lisa doll is a korean Guy, which is the world champion in playing go go with the oldest board game in the world And it's really hard to play it requires strategic thinking So you have to be creative all the privilege we human being have So the machine cannot be creative you would say And I can strongly recommend a documentary at netflix Because they show this battle between man and machine And the software is cool deep mind alpha go And lisa doll Went to the press and was pretty optimistic because he never lost the game and he will win again He lost the first game. He lost the second game. He lost the third game And then again he stepped in front of the press And I think this is really amazing because he said this move was really creative and beautiful So what i'm trying to say AI is not only about to automate processes But also to come up with new things to be creative and beautiful And the first the first game you won't believe lisa doll won The first game everybody felt relief We can win against the machine and this move is called the god move because no AI could predict that move And no human being have ever played that move Just think of that what lisa doll have ever played this move if you have wouldn't be challenged by an AI So and this is my definition of AI Augmented intelligence. How can we use AI to come up with better decisions? How can AI help us to do a better job? The question is some bad tongue say oh, that's not new AI So we have regression maybe even an excel sheet is some kind of artificial intelligence But what's new is that AI becomes more and more autonomous AI learns to learn If you compare that to the example of gary kasparov, that's completely different In the case of kasparov the AI was trained the people explain the AI how chasers works In the case of go the system was completely left unsupervised So the system learned to learn how to play go and I think this is a tipping point This is a game changer when it comes to AI the system are getting more and more proactive They are more and more autonomous and they learn to learn and that's completely new and that's good news Okay, but we don't want to talk about AI for gaming. We want to talk about AI for businesses And in particular we want to have this deep dive how can we use AI for marketing service and sales And there's a research conducted by McKinley two years ago and they try to figure out Which is the biggest lever for AI in the business environment because you know, you can use AI for predictive maintenance You can use AI for HR for example And the interesting thing that they found out that the most important part to use AI is the customer facing area So I think it's worth looking at the customer facing area because we can really unleash the value of AI Um, and sometimes you need that begging that McKinley Do some research underlying the importance of AI in the customer facing area Okay Now it's a rather complex slide But it's really important because I would like to outline the solution space of AI in marketing sales and service So on the left side You see all the use cases that are related to automation So it's mainly about how can we automate processes? How can we gain efficiency? How can we reduce cost and all that stuff? And you see the bubble size reflects the level of maturity and the color reflects the level of usage So I give you an example for example real-time bidding This is a technology based on AI, which is pretty advanced. It's highly automated But the business impact is not that high In contrast to that we have pricing based on AI for example amazon is doing a perfect job Using AI for personalized pricing. Okay, this has a high business impact It particularly are in europe It's not that mature because companies are afraid of using that because you know that kind of Customer democracy. They don't want to display different prices So they're struggling with that even though the dynamic pricing based on AI has a huge business impact. Okay On the other hand, we have the more creative part of marketing when it comes to Brand management, the strategic perspective of marketing media planning storytelling Influential marketing and all that stuff And you see the bubble are not that big and they're not the deep colored, but it's changing So AI is more and more tapping into the more creative part Okay, so this is more or less the solution matrix of AI and marketing sales and service And now I wouldn't like to touch upon some of these use cases And give you some illustrations showing some best practices. Okay Um, and I would like to start with elite prediction. So how can we use AI to predict the next customer? So um, and this is example Of stroller. I'm not sure if you're aware of stroller. It's a pretty popular company At least and in Germany, they do a lot of digital things. So they run a lot of campaigns They had a lot of digital companies that belong to that company. So, uh, they are campaigning all the day. So and the question is What is the right audience to engage with? So what is the next customer and this is usually done by human beings by sales reps And what we did is we used AI in order to predict the next 1000 customers. How does it work? Um, you have customers Hopefully, maybe you have a customers And then the a customers are enriched by different data As I mentioned before this could be social media data digital data websites. You name it And then you have a very detailed view on the company. You have more than 10,000 data points representing a company Not only the thermographics the industry number of employees all this stuff, but also the digital and social data And then AI automatically detects so-called look-alike audience Companies you don't know but they seem to be similar Okay, and the system also provides you okay, this is a company you should approach while linked in This is a company you should approach via the traditional way And then you can engage with the audience and then the system have to learn So the system helps you to detect leads it helps you to profile the leads Um, and this works really really well And the discussion came up. Okay AI is going to replace a sales rep and I said no No sales rep should be afraid of an AI a sales rep should be afraid of an employee that is armed with AI because He can focus on the real valuable tasks like empathy Delivering a perfect customer experience. So I wouldn't say that AI is replacing sales rep AI helps you to to keep track what's going on Helps us to find intense help us to find triggers we can jump on And so this is really a powerful tool if you you want to come up with new customers You can also use that Technology to enrich existing customers Because maybe you want to have more insights about the existing customers Maybe you want to run a campaign and that it's important to know Is there anything I have to know maybe they launch a new product? Maybe they are struggling with a competitor. Maybe there's a new CEO or whatever So there's so many insights you can use for campaigning for marketing in order to approach your audience So it automates the process to a certain degree so that you have as a marketer more time For the really creative part of your job So the next one it's about service and engagement, you know the slogan services in marketing So I want to talk how we can use AI to deliver And not only to react The correction So let's say it's a good base of customer engagement on the right side You need to feel that you weigh when AI come into play And you know, I think you heard of this bot chatbots based on AI Um, what does it mean? I give you a short overview for you when we talk about chatbots Um, of course, we can use these chatbots within our messenger system. We can use it Are at facebook messenger we chat whatsapp you name it We also have chatbots on our mobile for the mobile Websites or we can use it on the web to answer questions And we have also this kind of smart speaker so you can talk to alexa amazon google home and all that stuff So they all try to deliver customer experience based on an automated process To be honest, I have to say the current status is not that good. There's a lot room for improvement If you try to check out how the bots are working right now, um, they're not really Delivering good customer experience and my maybe that the reason for that is that a lot of companies Under estimate the effort that is required to come up with chatbots. It's not a plug-and-play solution And I would strongly recommend Not to install a bot if you could not ensure a certain quality Because you can kill customer experience That doesn't mean that you have a perfect ai that doesn't mean that you have the perfect bot But you have to ensure that you answer the question of your customers Interestingly enough facebook now requires some so-called escalation parts So you have to ensure once the ai is not able to give an answer you have to define the workflow how it's rooted to the Human agent giving an answer. So the yet you really can assure 100 percent reliability So smart bots work in a way that once they get To a certain limit that they have a smart hand over to the human being. I think that's a perfect mix And I have to say there are so Really awesome achievements in the field of conversational ai and I will provide you with some examples and therefore I have a short video Hopefully you can hear that because I think it's it's really amazing. So I just start this video that gives you an idea What ai is able to deliver when it comes to conversational business 60 percent of small businesses Don't have an online booking system set up We think ai can help with this problem So let's go back to this example. Let's say you want to ask google To make your haircut appointment on Tuesday between 10 and noon So what you're going to hear is the google assistant actually calling a real salon To schedule the appointment for you Let's listen Hi, I'm calling to book a women's haircut for our client. Um, I'm looking for something on May 3rd Sure, can you warm For what time are you looking at 12 p.m? We do not have a 12 p.m available. The closest we have to that is a 115 Do you have anything between 10 a.m and uh 12 p.m? Depending on what service she was like, what service is she looking for? Just a woman's haircut for now Okay, we have a 10 o'clock 10 a.m is fine. Okay. What's your birthday? The first thing is lisa Okay, perfect. So I will see lisa at 10 o'clock on May 3rd Okay, great. Thanks. Great. Have a great day. Bye So hopefully this could give you some some flavor how it could look like to have this customer experience based on ai And it's not a rule-based system. Most of the bots we have right now are if then so it sucks so much time actually To create all these rules and here we use deep learning It's a kind of a neural network that automatically find the pattern and come up with the answers Maybe i'm not sure maybe on the right side. There's amazon alexa That would be crazy then we have two bots talking with each other in a pseudo empathic way And I thought the machines understand zeros and one in a better way But anyway, this is a narrow intelligence The system wouldn't be able to play go of course, but it's able to to answer question of your customers And it's not based on on on on rules There are a lot of other examples Also for daily life, and I hope you all are in strong committed relationships And you're not going to break up with that but in case you want hopefully not Please do it with honesty and fairness not just text it or use what apps What apps because in the time of commutational business you can do it in a much smarter way And I will just show that to you Hello Hi, can I talk to diane, please speaking Hi diane, I'm calling on behalf of john to schedule an appointment for what the appointment for you to come pick up Your belongings from john's apartment Excuse me John would like you to remove your belongings from his apartment. What are you talking about? I'm very sorry, but john has decided to end your relationship. Is this a joke put john on the phone I'm sorry. John's not currently available. What where is john? I cannot believe this. This is bullshit How does 10 a.m. Sound are you fucking kidding me? Oh, if 10 a.m. Is not available. What time is suitable for you? Fuck you Okay, let's get serious again, but you can see conversion conversation AI will be part of our Daily life. There's so much use cases you can apply Not only the customer service area When it comes to conversation AI, it's important to understand that there are two different layers You can see it in the picture. We have an audience You have company and you want to engage with the audience you can do that via different channels devices touchpoints, whatever The first layer is to understand the request the intent of your customers Maybe you heard of natural language processing natural language understanding In this field, we are pretty much advanced using AI Uh particular where it comes to the english language. So it's it really works well to to detect intense The second layer is we have to give an answer accordingly And usually here we have more the approach the rather approach of if then rules But you can also use artificial intelligence not only to understand the intent of the customers But also to give the right answer And I think in the last two years they have really big great achievements what you can see here It's called meaner. It's a framework from google. They launched in January, february of this year And and you can check it out in the internet and you see some dialogue on on the left side There's no human behind it's just artificial intelligence. It's not rule based They train billions of data to come up this kind of dialogue And if you can see what about the quality you see and please be aware of that we human beings are not perfect So if you look at the benchmark, it's 86. So this is a measurement of communication quality And efficiency and you can see meaner is pretty close to human beings There's another really exciting technology out there. It's called gdp3 It's part of the open ai framework of ellen musk And this is a really amazing tool automatically Generating text messages And ellen musk decided not to launch this tool and not to put it public Because the tool is that smart that you could create fake news easily And ellen musk do not want to run the risk to create fake news So what i'm trying to say is now we live in a world of if then it's a rule based world when it comes to chatbots But this will change in the upcoming years because we have great achievements We can use artificial intelligence not only for Detecting intents Maybe voice intents or text intents. We can even come up with perfect customer appearance in terms of giving the right answer, okay Um, but conversational ai means more than just talking to your customer having chatbots, whatever I think that goes beyond Maybe you think okay, I've alexa at home and i'm ask alexa About the weather and all that stuff but The devices are getting smarter and smarter. So you have to think beyond that echo alexa or google home I give you an example last year in las vegas as a cs They introduced the first toothbrush Equipped with amazon alexa with ai inside and and and you would say oh does it make sense? You must be really desperate to talk to your toothbrush, right? And it's not that easy to talk to your toothbrush while you Use it. So of course the idea is maybe you do it not in the right way. Maybe it's used then the system can automatically Order a new one. Okay. So devices get smarter and and within the pit of sex tech They also launched the first vibrator Equip this ai with alexa. You see there are no limitations So what i'm trying to say all the device are getting smarter There's ai inside and they will be connected with each other You know that internet of things smart home Internet of humans and at the end of the day internet of everything Okay, so it's not only this conversational interface talking to your customer. It becomes The brain of a new digital ecosystem Which is evolving over the time and that not only works For the smart speaker you can also apply that in the messenger world For example, if you look at we chat we chat cover the whole customer journey based on conversation ai Everything is done by ai in this ecosystem. So it goes beyond How can i use we chat? How can i use amazon alexa? I know that sounds rather visionary, but I'm pretty sure that we have this conversation ai at the brain for the upcoming ecosystems. We are living in Another example, it's a more creative part. It's uh the job of media planning You know, the question is how to spend my media budget So spend it for for google for newspaper for for asm television So the question is how can i derive the optimal media mix? How can i get the the perfect budget allocation? and as you know, that's not That transparency process because there's sometimes some black box and and you know, they're kickbacks and all that stuff It's so ai have do not it does not have an agenda So you can use ai for media planning to come up with a rather objective and empirical based media plan and ai is not only taking into account the static and linear relations but also with dynamic and non-linear relations So if you use ai for media planning, you get a rather objective data driven recommendation how to use your money And there's one h&d doing that job providing this ai as a service. It's called blackboard seven And I think I like the slogan the future media agency is not an h&d. It's an ai So I think sooner or later ai will out glass human being when it comes to media planning Okay, now let's come to content marketing content creation and also the more creative part of storytelling So as you can see, we are pretty much advanced in the field of how can we organize and classify content based on ai tagging classification all this stuff another interesting thing that you can use ai to increase your likelihood that The customer is going to read your newsletter that it's convert in some way Okay, you can can use ai to predict which other words that are most likely to resonate with the audience Another example is we can use ai to personalize content Then of course we talked about the chatbots, but the finishing touches. How can we use ai to create new content? Does it work? Yeah, it works Um, maybe you know that a lot of press releases weather reports sport news are already written by an ai But even the more creative part is written by an ai. This is an example. It's uh used by by a um ibn watson the ai ai framework and What they did they call all the data out there on the internet that are related to marketing then they use ai Uh to create the next edition of the drum you can read it on the internet It's not that bad You can also see a video where all the employees and editors are bored because they have nothing to do It's not the driverless card the driverless um magazine Another example uh lexus uses ai to create a commercial sport That doesn't mean that you press the button the ai button and the ai automatically Creates a new commercial, okay, but it helps you to come up with the right storyline It helps you to come up with the right visuals. So it's a kind of co-creation. It supports the process of being creative another example In this field of content creation Is netflix you all know netflix and you all know the trailer you can watch at netflix And these trailers are not created manually. They're based on ai and big data So maybe some of you guys would like to see Fast driving car some of you might see some actors So the ai automatically predict What are the trailers that are most likely that you will watch the video or the film whatever So it's a kind of hyper personalization. We personalize Not only the message we personalize the content So we have one to one production of trailers at netflix And if you think that through the end, I think we will also have advertisement that is one to one personalized to your audience and actually Tencent is working on this these days and I will provide you with some insight How that could look like if you use ai to automatically uh This play adds in a very personal and dynamic way Impressive and this brings me to my last slide so that we have enough time To have this discussion and hopefully I can ask you quick answer your questions I just want to leave you with one final Slide and this is a short outlook We talked very much about Marketing use cases and new technology that that avoid over the time But behind all this are human beings And I think we are all aware of the fact that ai will have an impact on our jobs, of course I know this slides in the first glance looks rather complex, but it's pretty easy So on the left side, you'll see jobs that are safe On the right side the job that are vulnerable that are at risk The upper jobs are good paid the lower jobs are not good paid. That's the whole story And so I do the reality check you see the teacher So the good news it seemed to be safe or that's good news or bad news You do not get a lot of money. So I'm used to that Um, you see a lawyer for example, I'm not sure if that is really the right position We call there's a software called laverton. There's an ai based software that go through tons of contracts Finding the the issues and also generate new contracts You see the sales rep you just think of the historic sample I wouldn't say that the sales have got to be replaced But the job profile will change over the time in a sense of augmented intelligence And then I find some really interesting Job profiles, maybe you're looking for a new job. Maybe this is some foot for thought We have a machine people ethic manager. We have a memory augmented a theorist We have an inbuilt marketing architect and we had a bot creative You might say I'm kidding now. There's a real these are real Job description and I think it's very easy to to think of job that will be replaced by ai But we usually have a lack of fantasy Which jobs will evolve over the time in marketing sales and service because of ai, okay Hopefully I'm not gonna replace by ai because I love this kind of presentation and in particular I love this kind of exciting discussion after this So thank you very much for your attention And hopefully we have an exciting discussion now And hopefully we have a discussion among real people and you're not represented by your digital agent. Thank you very much Thanks a lot peter in behalf of everybody. First of all, I do just some clapping So just imagine that this is about more than a hundred people now clapping That's one of the things we haven't really got Artificial intelligence or or even support for to do the right clapping in in valunars I think that's something the tech should really change Thanks a lot for the presentation. I personally found it very very interesting I'm quite sure that there will be some questions. So everybody now you have that Comment box the chat box in the lower right of your screen Just try to use that one and if you have a question for peter and I'm sure there are some questions Just write it down there and we will be happy To to use that question. Oh, the first ones are already coming Um, give me a second. I will do that that way So here's the first questions coming from Olivier What advice would you give to a direct marketing company that wants to start using ai to answer marketing efforts? Or for its clients, what could be the first step? Yeah, the very good question. I think you should be very clear about the use case Because I know it's sometimes confusing. There's so much opportunity is the window of opportunity of that big using ai So where to start I think is the most important thing And usually I would say the customer journey is a perfect basis in order to figure out which is a perfect use case so what is the best phase of the customer journey where you can start and then you should Figure out what is the use case that has a high business impact And it's also feasible in a sense that you have the data. Once again, it makes and breaks with data So you have to be very clear understanding. Do we have the data in order to run ai? So sometimes I start projects and we unfortunately We have to come to the conclusion that we have to wait for the ai project because before you have to have a data strategy You have to be very clear which data are available Which data are worth fostering which data are missing and in case data are missing how can we bridge that gap? So my advice would be not to start with a full blow and and you know start with 100 power point slides Explain what I can do in the marketing area Try to find a showcase Try to figure out the potential for a certain use case and that was exactly the idea why I showed so many use cases And and then try to showcase this use case to get the feeling how ai works and then step by step you can extend this approach But really my advice is you have to be aware that you need to have the data in place Otherwise you cannot apply ai and by the way because I think that's a really important point a lot of people Are telling me we do not have the data to run ai There are so many opportunities To get the data For example, there's open data. There's so much data free of charge or you can synthesize data You can artificially generate data to use or maybe you share Data with your competitors. Maybe so it's to set up Alliances to to share data for example, so sometimes I think it's an excuse We do not have the data you need to have a data strategy in place to come up with your first showcase using ai and marketing Peter there's another question regarding the post offers. How how can we use Give me a second. There it is. How can we use ai artificial intelligence in post offices? Or let's let's not say just the offers, but the whole post operator from the From the sending the post collecting it and using it. So how how would you say? So it's also the So it's it's not a physical one also the digital post office or I would say let's stay with the digital physical one first I think we we all talk about multi channel only channel management So I think in the physical world, it's not that easy. We can can use QR codes. We can use, you know Digital signage and all that stuff. I think the strengths of ai to be honest is how can we use digital data? And how can we deliver a perfect customer experience in the digital world? But even though you need to have customer insights and if you want to deliver perfect Customer experience in the post office, you know what your customer is interested in and and you can even use Digital data using ai and you can transfer this insight also in the physical world I think there is not that clear cutting dividing line. This is digital world and this is a physical world You can use the insights that come out of the digital world based on ai also for the physical world And delivering a customer experience. So I think the biggest potential of ai is if you can apply it in a digital environment That's for sure, but you can use the learnings also in the physical world And what I see is that at least some of the postal operators are already using ai in a very interesting way. So Starting with for example sorting the post They use ai in Letter recognition and to to recognize where the letter should go I know that a lot of the postal the parcel services are using chatbots in customer service To to give answers to various my package at the moment So I think there are quite a lot of of application areas where postal operators are already using ai or ai technology So there's the other side of ai is often data protection. So Is it some kind of Things going against the other each other or is it something where where bait data protection might even help or how do you see that? So in particular Europe, we are faced with the GDPR and Sometimes not that easy to do ai because you need to use data But I think sometimes we use data protection as an excuse not to use ai There are several ways to use ai that are completely compliant with data protection You don't need to know the name of your customer. You can work with anonymous data Work with anonymous data. You can use artificial generated data So to be honest not that easy because the data projection is quite challenging And you have to have some ideas how to use ai but I can say It's not a promoter actually for ai projects, but I would say it's not a showstopper There are several ways to run ai projects that are compliant with data protection Um, and I think I really appreciate data protection, but first of all, I think we should Focus on the value of data. How can we unleash? The value of data and not to talk in the first Step, how can we protect data? So I think sometimes the Discussion goes in the in the wrong direction, but the good news you can use ai And you're not in conflict with this with data protection all this stuff Speaking of data, what are the first steps for a data strategy? Yeah, very good question. First of all, you have to know which data are available Okay, so you asked to have a very clear understanding of the current situation Then you should try to figure out which data are missing and Data strategy means you should come up with ideas how to bridge that gap. So maybe how can we capture data? So the most important point is to to have a big picture of the current status when it comes to data To figure out which data are missing and to figure out how can we can we fill that gap? And as I mentioned, there are a lot of ways to fill that gap For data alliances open data synthesized data and all that stuff So but this should be really the first step before you think of using ai to have this clear understanding of your data And data strategy also means that you have an idea how to work on the data How to get the missing data and once again, I think this should be the first step when you're trying to to enter the ai project theme Are there any ai innovations, you know of in the logistics section So that's not very much related to to the marketing area But I think when it comes to internet of things, so as I told you every device gets smarter So I think if you if you track device if you track things You can can use ai. So there's no I think no boundaries, right? I would say I'm not no expert in logistics to be honest But I think when it comes to physical goods, they also equip by QR codes They all get more and more digital and so they produce data you can use to optimize logistics Internet of things and then a rapid thing that also belongs to logistics, but I have no Exciting use case I could talk about Well, what's coming to my mind is uber for example They don't see themselves as a as a taxi corporation They see them as a data corporation and they use a huge amount of of ai to to make sure That the right uber cars at the right position at the right time They use ai to to detect how much money they can Get at a certain point of time for for a certain right So that would be one of the examples that is coming to my mind From where should we start is there are some kind of minimum requirement for ai Which is a very interesting question is we are not a big post operator, but one of the smaller ones So what is the investment in a where where does it start? Okay, I think the good news is we have some kind of democratization of ai so 20 years ago You have to spend million of dollars to run an AI project now The the most advanced ai software out there for example google tender floor just to give an example is open source is free of charge And I talked about big data But you can also start with small data and the new the good news is there's also a new method It's called transfer learning you can even if you have to have small amount of data You can try to transfer them in an even bigger environment So the the good news is there's no minimum requirement when it comes to budget Of course, you need some data scientists or people that are eager to use that You can start with a small amount of data to get the first insights the software is available So I think the really good news is that there's no Our minimum requirement I think the the most important thing is you have to be Eager to do that you have to be willing to do that you have to be curious about that I think that's the most important thing and I wouldn't say you need at least 100 000 euro you need 1000 data sets That's that's that's not the point I think you you should try it and the good news is it's not that minimum requirement for my point of view There's a question about why can we not have a universal language for ai like it programming language available For anyone to use I think that's you just answered that a little bit a bit that most of the well not most but a lot of ai systems are now open source And you can use them free of charge and there are a lot of data sets out there which you can use free of charge So do we have some kind of universal language for ai or is it very very different from the different systems it depends on the system and If you look at the different systems, I explained we have amazon alexa google home So they all came up with their own ai language because they want to battle about the ai Who has the best ai system in place? So we're working for quite a long time to get this universal language of ai to get the standard Everybody can apply but don't forget their companies behind they want to push certain Ideas and even though if you talk about open source That doesn't mean that we have this unit language universal language So we have to be very much aware of the fact that we have to manage different ai languages I would really appreciate to have this universal ai language, but I think we won't have this in the near future Peter thanks a lot for your very interesting presentation. Thanks a lot for answering all these questions I see that we are one minute before 10 at least german time. So one minute before the end of our thing. So Once again, thank you very much and a clapping from everybody all around um for all of you Who found this interesting We will have the next innovation talk the number two talk on December the 8th And we have a totally different topic which has but there are some similarities We will have a topic which is social selling with linked in So you all know linked in probably but you found that over corona of it covet covet 19 social selling is getting more and more important and We found a book about that which is linked in versus locked out and we will get an innovation talk about the linked in strategy You will get tips and tricks how to use linked in and we will even do live profile checks with all of you So keep on and I would be happy To see you again on December the 8th for our next innovation talk in our innovation talk series Thanks a lot. Peter once again. Thanks a lot to the upu dma b And thanks a lot to everybody being in this innovation talk and i'm happy to see you again Early december and that's the end for today Thank you everybody and take care and please stay healthy most of all things. Thank you. Bye. Bye