so let's start with introductions uh my name is Sian uh I'm the founder and CEO of smart guide uh and uh I have background in AI studied AI long time ago before it was cool and also business and then I spent 10 years with McKenzie helping uh companies uh to digitize and of course I am a very passionate traveler visited about 70 countries uh hello everybody uh my name is Roman alova uh at smart guide I the chief marketing officer uh I also study marketing uh and uh lately I've been devoting my time and effort a lot into Ai and implementing AI also in smart guide uh and uh I'm very passionate about the topic as well so that's why uh we're very glad that we can share uh our insights and things that we already learned also with with you and mystify the topic a bit uh for you from the the tourism sector and from destinations and before we get uh into AI itself uh let us introduce a little bit how we go there what we are doing so U um if you can go to the next slide please um very good so we noticed that uh uh tourist destinations uh very often develop um their own apps to guide their visitors but they struggled with the development it's expensive takes lots of time uh and then very often uh the the end users The Travelers did not use it because nobody wants to use a different app in every place they travel and that's why we started smart guide as a platform can think of smart guide kind of as a YouTube uh for travel guides so that anyone who has interesting stories to share can easily publish them in the content management system without any it development and uh you know with a few clicks uh can have guide uh both in an application and also on a website then if the guide uh turns your phone into a personal tour guide uh now covering more than 1,200 destinations so it's basically one app for all these places and uh it will show you what are the places to visit recommend you uh tours to take how to it triner how to explore destinations and attractions and Outdoors it works based on GPS when you arrive at some place tells you interesting stories from local guides about the place that you are visiting also in audio so you don't have to read you can just listen to that you can watch videos uh but also works uh indoors uh by scanning QR codes and uh we can spice up the experience uh by by uh showing for example uh augmented reality experiences and the guide is available now in 30 uh different languages so it's a great tool for international audiences and as we guide uh the visitors uh we get to know them very well we know exactly where they came from from which channels uh how what are their interests and how they use uh the content what they consume uh what where they go where they spend their time and because we use GPS of course all of this is gdpr compliance so aggregate not on a personal level but we can see how tourists from different uh demographics behave in the destination very accurately and uh that's actually you know we first had this uh crazy amount of data that enables smart destination management and that's also inspired us to uh put some AI behind it and that's why we spent uh the last few months uh really deeply involved in Ai and that's something we would like to share so uh you all know uh that AI is already transforming travel for quite a few years uh just when we look at the experience when you book a flight ticket uh the price of the ticket it is set dynamically now for most of the airlines using AI you then if you need any support while buying the ticket you already use chatbots uh to communicate um about that then when you check in you know you go at the airport you have facial recognition AI uh and then even if you you know are not happy about the flight and you call uh the airline uh they will usually use uh sentiment analysis AI to see how angry you are and if they should connect you to a live person or if it's okay to let you talk to to a machine so uh there is a already in many aspects of traveling but what we found out is that uh it's very difficult to uh to grasp a and what would we do for Destination managers because it's not of course the fault of the destination so unlike booking.com an average destination doesn't have an uh AI team available so uh we wanted to show you a few practical examples of how to use AI but the last thing before we uh go there I don't want to bore you with the different terminology of AI and and the theory but what is good to know when you come out of this Workshop is that uh even though now everyone is talking about jgpt uh not about the PO and also the pitfalls of jgpt not all of AI is uh the same as jgpt there are many different types of AI so C GPT is a generative AI large language model it's built on deep learning neural networks but they are completely different types of AI so like you know if you have self-driving car it works on a very different principle than C gbt so we now talk about some principles and some problems of AI don't extrapolate it uh on all the other uh applications but let's dive uh right into the interesting how you can use AI practically uh yes so uh we looked into several different uh applications that could be relevant for Destination management and first of them uh pertaining exactly to to generative AI as Yan mentioned is uh content creation content creation about the destination obviously uh nowadays uh we can see that the destination obviously have a lot of content created about themselves but very often uh it's uh quite like a lot of facts it's little as if a historian would uh create the content and not a guide it's uh not really as engaging as maybe people and visitors uh might expect and uh that's why if you send uh visitors or if you if you direct them into the content on or or the standard Viki style content available on your destination website they might not find it uh as engaging uh this is exactly where jgpt and similar models and generative AI steps in because unlike um some people might uh think that AI is um or should be predominantly used for for uh like calculations or or non-creative task per se but is actually quite the opposite and generative AI is quite uh useful for Creative tasks such as creative creating engaging content uh it's very efficient because unlike uh standard human copywriters who literally have to write everything from scratch uh when using generative language models for creating destination content uh you can uh create vast amounts of content pretty fast with models not just like jgpt but also for example Google gini which is uh basically a direct competitor of jgpt uh and there is a lot of benefits uh obviously also pitfalls but we will cover both uh in terms of benefits uh beside the speed uh very important thing is that you if you create a Content using jgpt uh so far legally it's still uh your content in terms of copyright so so you can use this content uh for example on your website in your communication on social media Etc uh another very important thing is that uh you can uh introduce many different styles of the narration and uh what's one thing that we found out when we were uh testing and and uh playing with the abilities of these models is that jgpt and similar models can be pretty funny and they can create very engaging content often times a lot more engaging than uh the the guide content that is uh currently available one Pitfall and that is like a very uh important and uh critical thing to keep in mind when working with chpt is that it uh needs to be fact checked every output out of chpt needs to be fact Che fact checked because it has a tendency to so So-Cal hallucinations and it uh literally uh because the way it works uh the way these models Works basically they they predict every uh other word the the the the the most probable word based on the vast amount of data these language models were trained on uh the most probable word in terms of context is what it gives uh so that's why it can create sometimes very uh creative and nicely sounding content uh which is absolutely not true so if you work with uh these tools such as jgpt uh it's necessary to talk to them or or communicate uh with these tools in certain uh language in certain Manner and certain structure in order to get uh really quality uh outputs that are actually usable and uh that make more that don't make more harm than than good eventually so in terms of uh tools for Content generation there are uh we different types of tools obviously but we can we can uh generalize uh and and divide them into three main groups uh based on the output so uh first type is the uh is are the tools that provide text outputs uh that is for example CH GPT or Google Gemini uh basically you input text a command or a prompt uh as it's called Uh in the AI um lingo and it provides you an output in form of text another area uh of tools another group of tools is where you uh instruct it with a text and the output is visual so so an image for example and the third is uh the text output which generates video output this is the newest uh newest uh and still not as uh spread area of of content generation but uh there are already very uh nice examples of these toools such as for example uh open AI uh Sora which uh created very very lifik videos maybe uh short preview all these videos are generated by AI they are extremely lifelike uh they are on the first glance they're are very hard to distinguish from reality uh they use different styles they use animation they use they use uh different perspectives and as you can see down here they are all generated based on certain prompts that were given uh in a in a text form so uh how to how to how to communicate with with these models such as for example Sora or ggpt in order to generate uh the content that you that you wish for uh in case of uh in case of tools for example such as mid Journey which is a text to image generator uh it's uh fairly not to say simple but it's more simple than communicating with your GPT because mid Journey as an image generator uh has a quite narrow usage so uh you can Define type of visual you want you can Define the perspective you can Define the aspect ratio you can Define even the for example if you want the photography you can Define the type of device uh which it should imitate or emulate and uh you will get uh an output which uh can be uh scaled and and the prompt is not as uh hard to craft and you can get very very lifelike results this is an example of a picture generated by by mid Journey uh on the left side you can see uh they're both uh Old Time Square in Prague on the left side you can see AI generated picture of Prague and on the right side it's the real photo of prag and you can notice that while the AI generated picture is beautiful it's really just like still size it's it's it has nice colors it has nice detail nice perspective if you look closely yeah you will notice that it's uh not actually the Oldtown square or actually the astronomical clock looks quite different uh in reality as the AI generated it it's pretty close but it's not the same uh this shows uh very very clearly why it's important to fact check because uh this is one of the like biggest pitfalls of of uh content Creation in AI that it's on a first glance it can be very effective it can be really stunning but when we look into details we can find a lot of false loads that's why for example if you generate content for our destination we advise not to uh create content of places that are well known because uh while the pictures can be can be uh very pretty they might not be might not be factual uh what we uh really really uh do suggest is using uh these generator generation tools for for example brainstorming uh logos if you want to create like a logo for a destination you can use uh these tools to to brainstorm ideas if you want to create a poster for example for an event uh it can create you the base visual uh you can also change uh styles for example if you have a picture by the way the first picture is also generated uh by by AI uh and we uh iterated the style into cubism into into into art deco so you can really really use the the content creation tool for for playing and like very flexible use of the content but we really advise against uh using it on Real uh places real destinations uh now when it comes to the structure and how to communicate and craft the actual prompt that will get you to the good results as I mentioned with the tools like mid Journey it's easier because uh you only only need to to Define how the visual will look in the end but C GPT is a lot more wide in terms of usage so uh you need to follow certain structure and certain rules and luckily uh we can already based on experience and observation and based on several research uh made Define several principles that are Universal and that you can use uh uh to op optimize your output and to to find on your output so uh the syntax and the structure of the prompt itself self uh is very decisive in terms of uh what output you get and uh five basic principles which I would like to highlight uh which I would really help you uh if you adopt them uh in content creation uh is first ggpt uh is like a really very well-trained student who just came out of of of studies but is not experien experienced enough and needs instruction uh that means the better instructions you define and give to the model to to jgpt the better results it will give you uh really think of it as as if if it was an intern and uh they they're full of information but they uh know no processes so so this is how to treat the Tool uh also uh within chpt uh there is something like something called meta instructions so you can Define uh General instructions for for the entire uh set of outputs that you use which will be applied to all con conversations with jgpt that you have for example if you want to maintain certain narrative style or you want to Omit certain words or you want to on the contrary uh always mention certain phrases uh you can Define this uh across across uh all conversations and then it uh facilitates and and speeds up up the content creation very much because you don't need to repeat the instructions all the time one very important thing that we also found out uh empirically is that jgpt doesn't like negatives it's always better to uh structure the The Prompt in a so so let's say positive way uh so instead of uh don't use negatives uh uh if I wanted to instruct jgpt not to use negatives it's better to say avoid using Nega negatives it's similar as in military language uh chpt tend to miss the don't or the negative not and only focus on the verb and do the exact opposite of what you what you ask uh one other thing that is that is also empirically learned that it's very important to repeat key instructions so for example if uh you want to uh generate a text that is strictly five 500 characters long and this is like a key prerequisite for the output it's recommended to repeat this this again something that you wouldn't do with a human uh it sounds not really natural but with jgpt it's always good to repeat key instructions and then repeat key instructions and if the prompt is uh too long it's always or too complex it's always recommended to break it down either to uh steps within one prom or sometimes even better to to break it down into a series of proms now uh it's recommended that chpt um prompt has has certain attributes and certain elements uh which will help you better Define the the output uh you don't need to use all of them and sometimes based on the context and based on the use case you don't even need to but it's recommended to really think of these seven areas or these seven attributes when you structure your prompt so first of all it's uh always necessary to define the task um so that is the key thing you want uh the the the model to do in this case I used an as as an example that a destination manager would like to create a Facebook post promoting an event wine festival or or or basically an event in in their City uh that should have certain length that should be in certain language uh is also interesting that you can uh use instructions in one language and output in the other that is something very useful especially for for for uh information centers uh and uh when you want to when you want to uh create a task it's always good to define a Persona of the of the U or basically Define an expert which will help CH GPT personify uh the task or personify uh um the the the Persona that you want to emulate and uh give you a better better output it's always good to to provide Conta so uh in this case I want uh tgpt to design an engaging post uh to invite people to the annual Wine Festival I want your GPT to become an experienced marketer and social media copywriter so that it knows uh knows uh based on its already learned knowledge base uh how to approach the task I want to give it a context that uh the Persona is focused on local events and promotion of of local events which further Narrows down uh how uh tgpd understands how I want the output to look then it's always good to define the style or the form so for example in this case uh we want the text to be original targeted towards families and children uh and it's always good to provide an example so for example if you already wrote uh similar Facebook post before or if you saw a Facebook post that uh you think uh sounds good you want to get inspired by it you can use that as an example and jgpt will imitate that particular style uh and it's always good to use uh defining instruction so for example if you want to avoid certain expressions or if you want to use for example emojis or whatever uh detail instructions you wanted to give uh it's always good to mention it in the prompt so if you apply all these uh attributes uh to to to your prompt uh you will get a structure that looks something like like like this example prompt here and uh then when you when you put it to jgpt you will get uh you will get an output uh that is very close to to your original to your original instructions and uh this is very very time saving and and very efficient especially if you do repetitive tasks because uh it's uh obviously possible to scale generation of such content and you can uh reuse these instructions uh in in future content generation and uh in the end you can keep a consistent style of your writing you you can even talk in foreign languages so so for example if you want to create the same variant of the Facebook post in uh different languages this is this is uh possible in GPT and uh there are also uh other uh ways to to use ggpt not just to generate content but also to uh rewrite content so for example if you have an existing content uh which you want to uh rewrite for example for families or for children uh this is something that is possible with jgpt and uh um it's very easy for example to take a factual uh text that is meant for adults and uh talk and and uh task to GPT to rewrite it as a fairy tale for children so you will reuse the the existing content but you will be able to uh Target a lot of different audiences and a lot of different segments uh in different style and in different languages which will very much uh change the way how you how you can address different different groups of of visitors and different segments and in smart guide we already played with it ex extensively and we created uh a copywriter AI copywriter based on JG pt4 uh if you uh scan these these QR codes you will be able to see it yourself in the app one is uh even like a funny uh guide and we we were quite surprised how how humorous ggpt actually can uh can be so Hello to friends from bled who are on the call and also to salsburg land that I also see on the call so another example uh can be to expand uh some uh descriptions when you have maybe just one sentence or sentences description of a place like in this example you can use the model to uh write it up into um you know a longer description that is not nice to read so the second topic is uh using AI for translations which is a very obvious use case so I don't know now if you were able to follow my mandering if not it's okay you cannot speak all the languages nobody can actually AI can and that's exactly uh the advantage so uh if we go further so we can use uh machine translations uh of text that got if you tried you know Google translate five years ago and you use D now uh it improved so much that uh you know now we are reaching levels depending on uh the source language and target language but the accuracy is actually very very high uh and uh even if you use the uh AI translation and then just uh you know sanity check it with a human you can save about 70% of the costs that you know would normally cost to do a human translation so that of course allows you to save costs or prepare your content for your visitors uh in many more languages um now uh tourist destinations that we talk to very often they are uh they came from the uh world of physical printed materials printed guides brochur sleep plets and uh nobody wants to have uh you know 95% correct translation in a print you know in a book that you print uh that's not okay and uh you know your visitors will not be happy about that but uh online world is different so you can see examples in for example booking.com that uses uh AI translations uh in the reviews or Google Places and travelers are used to that uh we also ask them about that and the important thing is you know when they have to prioritize do they get the content 95% correctly translated in their language or do they have to view it in a language where they are not so proficient almost all of them say yes I'm okay with a 95% correct translation but I want to know that it's an AI translation so but it's you can use AI translations in digital forms but inform your customers that or your visitors that it's an AI translation and that's exactly what we do uh as well in smart guide where we now offer translations in certain languages it's also built into the content management system so if you have your content on Smart guide you can just you know you have content like in check language here for example you can just click on at translation and uh you choose the language and it will automatically use DL to translate uh the content uh into English so it's really easy really fast and then yes you can of course fine tune the translation to make it even better but saes tons of time you can also now uh automatically translate uh your guide into C Lang es so with basically a click of a button which can be very effective of course to expand the audiences so that you can reach uh you know you have usually four five Target main target markets but like this you can reach also the you know other 25 that uh cumulatively can be quite big quite important uh what is good is that uh you will uh see the analytics of the actual usage so uh you can translate your destination guide into certain languages and then you will see how many people actually use it on polish or Korean or other languages and based on the actual usage you can then decide where you invest in some uh you know making the translation fact check it by human basically so uh you can use your transtions budgets very effectively and of course it has benefits for SEO if you publish more content on the website in different languages to RIS from speaking other languages will be able to find you finally small trck that I noticed that some tourist information centers use but definitely not everyone uh there are uh tools ready for uh translation in the real time uh when somebody actually comes into the tourist information center doesn't speak your language uh then uh you can just have uh you know all the information center staff uh have Google Translate a free application on their phone where uh they just press the button let the T in different language speak to this app it will understand what they say translate into your language and you can just press the button listen to that or you read it and then you reply in your language and it will translate it back to that language so a very easy way how you can it should be basically uh I think installed on most of the forms of the stop the tourist information centers uh now when it comes to tourist information centers and Visitor Center staff uh AI chatbots uh actually have also another usage that can be used uh not obviously to to replace uh but to uh work with the the the human staff at the tourist information centers mainly because uh it's not possible to keep all the time the tourist information center open all the time but it is possible to have uh an AI checkboard available 247 that will be able to answer uh relevant in a relevant way questions that tourists might have because uh on one hand you can send tourist to to websites but um they rarely uh really read all the information there they are not as engaged uh and they uh tend to uh finally not get the information that you actually want them to get uh but if they have an interactive way of of getting the information such as and chatbot you can uh provideing with information even when the when the INF center is closed and uh the AI check BS uh there are basically a lot of tools on the market that are readily a available uh that allow creation of uh efficient root planning chatbots uh that allow for example 24/7 VIs assistance because it's trained on on important information uh that you want to give to visitors and it saves a lot of time uh by answering the repetitive queries because tourists do tend to uh ask very similar questions all the time uh and one big advantage of these tools is that they are really easily uh trainable using your existing content so existing uh knowledge based allbe it in version of websites or uh if you have it in documents such as PDFs everything that you have in a written form can be used to train such chatbot and to demonstrate just how easy it is uh I would uh like to show you how to build literally in a few minutes uh uh um from scratch a chatbot in a tool that's called D AI we will lat share share a link with you where you can uh play with it yourself uh this is how the interface looks uh right after you log in login is free you can use your your Gmail account to to do it very quickly and uh when you create a chatbot the only thing you need to do uh is to name it I will name my chatboard like very uh rudimentary my chatbot and here I can choose whether I want to use existing files or whether I want to use URLs uh to save time I will I will use URL so for example if I want to build a chatboard for carsbad uh town in Czech Republic and use their website I can just go uh I prepared a list of of uh like the most uh popular sources from from from Google yeah so I use the the website of the city I use Wikipedia sources note that I use uh different languages of sources it doesn't matter you can mix German English uh Czech polish uh Spanish it will understand all the languages because it can translate it so so uh you will then be able to not only ask uh the the chatbot in your language but also um task it to to answer you in a certain language uh so as you can see I can just very easily uh uh copy and paste uh URLs and create a chatport it will learn uh and train on the files I I provided it you can see that it's literally seconds it's it's very fast it's uh extracting all the information uh from all the sources that I provided and within minutes uh you are basically ready to go and uh implement this chatbot into uh your website now when I ask it uh for example recommends an itinerary in Carlo viari for a family with children there it goes yeah so so uh and this is really something that you can do yourself uh you can use the link uh that's provided here we'll obviously share share the the presentation so uh we will share also the link and you can use it to create your own chatbot and uh then when you have your chatboard ready which is literally as easy as I just demonstrated you can just copy uh the script uh of of the chatbot and paste it U either onto your website or onto any uh online location where you want the headp to uh operate uh with the paid versions of D you can obviously customize the functionality customize the design you can add your own logos your own uh colors you can uh use also more um because as you saw this this result was on GPT 3.5 Which is less smart uh with a paid version you can use even the smarter jgpt 4 uh but already with this version this is very very useful uh tool to to implement to destination to to tourist information centers and destinations another great uh use case of AI in destination management is for personalized recommendations why personalized recommendations matter I mean we are luckily out of covid which is great news for tourism the bad news is that uh in many places the overcrowding and uh uh comes back in you know when you look at Venice it's sometimes look like this you cannot even pass through uh the streets yet at the same time there are still many places even in menis you know like this authentic Gondola repair Workshop that is just literally a few steps away uh from the main channels and it's a very interesting place to visit to get a real sense of when is uh but not so many tourists go there they all stick to the same uh places to the top 10 you know everyone uh will Google the the top 10 places to see in Venice and that's where uh the personalized recommendations uh step in so uh if you can tailor the recommendations similar to what Netflix or Spotify does you know Netflix also shows different recommendations uh for movies to watch and for different music to listen to based on what's interesting you because everyone has a little bit different taste why don't we take exactly the same to travel guides because then we can spread the crowds and uh make also the experience for everyone more interesting because it's more relevant now here we can also speak from our experience because we built uh an AI uh ourselves to do exactly that and this is when I mentioned there are different types of AI so this has thing to do with chbt uh we used uh the more traditional uh machine learning algorithms to train uh our own AI uh to recognize B on 20 million places of interest when it reads content so you will write any content it will go through that and find out like this traffic light semaphor in London is about sightseeing art and sculpture then we see exactly how traveler interact with the content we know whether they visited the place we know whether they read the content they finished listening to the audio or they even liked it bookmark F and based on that we built their personal interest profile and this has different weights of different interests for each person and now when we know what the person is interested in and we did go through all the content you have available for travelers we do exactly the same that Spotify do with music and we uh give personalized recommendation not to like a segment but every single traveler based that you know what maximizes their profile so then in each destination we would still show uh the top places to visit but besides the top places because everyone wants to know the top places uh we also uh show a carousel of recommended places which are never in the top always focus on places out of the beaten path and this really can help to uh distribute uh the crowds more evenly uh across the city and avoid the congestion I tried it myself on on my profile just to show a specific example so you know I very often travel with kids and we uh prefer nature over cities so uh I got my interest profile you know it corresponds I have some like memorials and battlefields there because that's what uh you know we are also quite interested in like World War I topic now with the with with the kids uh I can see I for example in Pilsen I visited the zoo it will know that I was there so it's actually Tik Tok and now when I go to some uh other destinations uh it will you see that yes I get a top places but then in my personal recommendations I very often if there is a zoo or a park uh you know with playground for kids I get these places recommended as well so what this can do is that you know I get inspired to visit the destination because normally uh maybe uh you know berau would not promote the zoo it's not one of the top attractions but for me it's relevant and I did not know that there was an interesting Zoo there so uh this you know I will go there with my kids and spend an extra half day at least in the destination so this is uh can help to um extend the length of stay of The Travelers uh and spread them to to promote the L visited spots and the final example kind of the Holy Grail of AI in destination management is uh itinerary planning so uh itary planner is basically a comp compan that can prepare you know tourist will share their preferences where they go what what they like what they want to see how much time they have available it will also take into account the conditions you know is it summer winter what's the weather forecast and so on and prepare uh a plan for visiting the destination special for them uh now uh we talked about C GPT and last year I mean it was the first year of of C GPT so 6% of Travelers already used CH GPT last year to plan their uh travels and of course now with the growing popularity of of Chad Bots this percentage is increasing and it's very easy right you probably trade it yourself if not I hope you will do after this workshop and so you ask GP what you know you can do in in your destination you get a result and the result is well Britten the top places make sense but there are a few problems with this uh one problem is that very often the tgbt is making up the content so that's the problem of hallucination because it's uh you know write compl nonsense to the places in your second problem is that uh even if you write the same prompt three times you will always get a little bit different answer it's written in a different way it essentially recommend the same places over and over so it looks different but it's the same stuff and so again you don't really uh help the overcrowding problem and the third problem is like you get a text output and the bigger problem for the traveler but now how do you use it right like now you okay do I get this output on my phone do I copy paste the text save it somewhere in some notepad and then you know look up the places on Google and then you know it's quite hard to use so uh what we are working actually on right now we don't have it yet but if some of you are interested in the topic we would be very happy to involve you in in some Pilots we are now integrating chat GPT directly inside smart guide so that we get the very natural interface interact think with traveler that the can do but then we only link the results what chb is recommending they link it to your official content so it will not be making up stuff about the places uh it's your content and furthermore we know the tourist very well as we showed in the personalized recommendation so when we send a prompt to the ch that the user is writing the prompt we will attach to the prompt uh FYI the you know this traveler is interested in you know natural sites and traveling with children for example or something else so uh we provide already the background information on the traveler to C gbt so that we can get different results for every traveler using gbt and finally we make it more actionable so that we turn uh the recommendations uh into uh smart guide so that we add the existing photographs of the places and the GPS coordinate and you know websites where you can buy the ticket and so on because we already have all of this information so we can include it and it's very easy to use for the traveler so if this sounds interesting we believe that this is the future so now we are actually working on this future and very happy to to get in touch uh yes so as you can see based on these uh five examples and five use cases uh there is already a lot of uh ways how to implement Ai and there is uh a lot of existing tools which can be just taken and implemented so there is no need to uh reinvent the wheel and uh it can help tourism boards with operation efficiency improving customer satisfaction and uh making tourism more sustainable by recommending uh less visited places and spreading tourists from the overcrowed hotspots and uh many of these uh approaches are already uh mature in other Industries uh not only uh in tourism and tourism only starts adopting them so luckily there is a lot of readymade solutions that can be just plugged in and in connection with this uh we would like to uh ask you and uh run a poll uh about your experience uh with AI tools and your destination so uh I will now uh run a quick poll uh here within within uh the uh Google meet and now you should be able to see the poll and uh you can you can uh you can vote so I will give another 10 seconds maybe uh so that you can you can express your experience five 4 3 2 1 Okay so I will now end the ball so yes so it looks like uh some of you already have experience and now I would like to run another poll if you do plan to use AI Tools in your destination U also based on the new information that you maybe got okay so how wait five more five 4 3 2 1 okay thank you I'm end in the poll okay so the majority uh all the participants do plan to use Al that is nice to hear what is interesting is there is actually no one who uh is who said no that they would not be using that so that's that show the direction okay so now yes I'm sorry go ahead so we just wanted to thank you a lot for uh taking the time uh and we were glad that we could share some of the insights that we gathered uh you know playing uh with um with AI and actually using it uh over the last uh couple of months and years and uh we would uh definitely be happy to to stay in touch uh we will this is you know field that is evolving really quickly if we did the presentation a year ago uh it would have looked very different than when we did it now and uh I believe we will uh move uh even in the next year uh long way so if this is a topic that interests you if you want to ADV ideas if you want to uh Implement uh AI in your destination management uh please get connected and uh we'll be happy to chat e