YouTube Video IZes7utRxzw  ## Metadata - Author: [[YouTube]] - Full Title: YouTube Video IZes7utRxzw - Category: #articles - Document Tags: [[case]] [[demo]] [[documentation]] [[drupal]] [[funding]] - Summary: YouTube Video IZes7utRxzw - URL: https://www.youtube.com/watch?v=IZes7utRxzw Okay. Hello everybody. Um yeah, in the next 15 minutes it's going to be a short session. I want to talk a bit about triple AI and specifically things that some of you built um already out in the wild. So first about myself, my name is Michael or people know me as Schnitzel in the community. Um I've been quite for a long time in Drupal, so more than 15 years. And this is probably my third year Drupal con. I stopped counting at one point because it just becomes crazy. Um, and I describe myself as an open open source maximalist, meaning um, I believe everything should be open source. So, and obviously Drupal is. So, that's why I'm here. And what I really want to talk about today is how Drupal loves AI. And if we talk about that, I also have to talk about all these companies. So, all these companies joined the Drupal strategic AI initiative. So all of us together, we are donating money and time to make sure that Drupal AI is as good as it gets. And yes, we're also looking for more people. So if your company is not on there yet, I really encourage you because um together we can really build what Drupal AI is and how it works. So we obviously love Drupal and we love AI. So um yes, we're one of the investors. I have teams that um work on all of this every single day. So we want to make Drupal better. We want to make MDI better. So what I've done is I've talked to different of some of these companies um but also not some of the outside. I went around and asked different people of like what have you built with AI and I want to show you some of the things what is already possible today because when we talk about triple AI people think like oh it's not ready yet and can I really use it and I can show you yes you can I'm going to show you a couple of examples so for example 1x internet has built they have one customers which is the world cancer day and one of the things is that they allow people to upload stories about cancer so their ways of how they um successfully fought cancer. Now, in anything in the world or in the internet, right, if you have a public form, you're going to get a lot of spam. You're getting a lot of things. And they were overwhelmed by the amount of submissions. They were clearly wrong. And what they did is with AI, Drupal AI's capabilities, they can now automatically moderate that content to check is it spam, is it probably not true, and things like that. Now, there's still a human in the loop, right? We hear this a lot that the you can still make sure that in the end somebody verifies all of this but what they were allowed is like they have a much better quality and it's also much faster. So and they show you can use super in production. Another company in the UK they have a large French telecom company and they wouldn't allow me to tell you the company name but Google for large French telecom you figure out who it is. Um anyway they um one of the problems is that they have a lot of assets. So pictures, audio and videos and internally it's a problem that the team always use. It's like hey I I have this picture am I allowed to use it? Like is it one of ours? And what they implemented with AI is a reverse image search that they can figure out if you have a picture or an asset to actually figure out where is the original one of it and do we have the license to use it. And in the past the marketing teams they just kind of opened that this is possible or that they can use it and now they actually have a tool. So they have an AI pattern matching system and they have now 99% confidence where before they could not do it again build with Drupal in AI already in production. Another one is the Southwork Council. Now it's a government and specifically old governments but they have a lot of PDFs. So they have more than 2,000 PDFs that were originally meant to be printed. The problem is PDFs that are meant to be printed are not very accessible. So what they did, they used together with Drupal, they used an AI to automatically convert the PDFs into actual websites. So not just OCR on the PDFs, but they actually create full-on websites, HTML, CSS, JavaScript fully automatically from a PDF. Again, Drupal obviously is extremely good in this. And what they now realize that they they do this 240 times faster than before when somebody manually need to do that like they hire people to convert this with the power of AI. Now again there's still a human in the loop and make sure that everything works but it's it's a possibility how we can do it much faster. Another one is called DB Shanker um for a company or a company called Cudos and what they do they do business applications in Drupal. DB Shanker is a company that ships around containers. So they basically do ship management for containers in Drupal. Has nothing to do with your website or anything at all. So they just use this for business. And one of the issue is that they get a lot of email. So there's a support team, right? Let's say you you have a container um from DB Shanker that get shipped for you and you have a question about the container like what's the current status, where is it? So you send an email to their team. In the past um the email had to be read by a support team and analyzed and decided is this spam for who it is who do I forward it to and what they do now with Drupal AI they use an automated email categorization system so the email comes in gets analyzed by the AI and automatically decides oh that's probably for this team and this team and this team now the support team is still there they still do the final decision but it's just much easier for them they don't actually have to read through the whole email to understand who do I need to forward it to because they already have a summary generated by AI. And the other thing they do is they go to a lot of events as we also all do like Drupal events and they now have an automative report generated that compares events that tells them you should probably have gone to that event or things like that. So what they brought down is really like exception handling over like a container we don't know where it is that used to take weeks they now have it in minutes because the capability of all of this. So again use Drupal AI has nothing to do with the website but they can make it much easier. Another one is in Switzerland where I'm from and there's a company Leap that's a triple agency and they have um a customer the Kton Baselshot um that is basically a state or a canton and they have a government website right and these government websites we all know they're full of information completely overloaded it's sometimes very hard to find data and what they did is they had a chatbot that you can just ask in natural language like hey I just moved here how do I register my child for school or how do I get a new passport and things like that. They just built this p this chat bot and they had 50,000 questions in the first year um answered and they asked like the people like did you like this answer more and there was definitely information or the data shows that people really like this. So super simple chatbot connected to Drupal. You can build this with Drupal AI within a couple of minutes today and they're already using it. Another one is very interesting. Um so there's a c a customer um in it's a global company but they're in Germany specifically Buring Engleheim. What they did they had 45 different websites for different countries and regions and they said we want to bring this all together into one Drupal site. The problem was right they have diff every of every of these different 45 pages they have a contact form. Each of them has a site map. Each of them has an about us. Each of them has like this different pages. And they needed to find a way to connect all of them together uh for like redirects and SEO reasons. And instead of doing it manually like to just visit every site, they gave an AI access to all the 45 websites and said, "Find me similar pages." And it ran through. It took a couple of minutes and hours to go through, but the AI came back with the sites that are similar. And that saved them weeks of manual work of somebody trying to figure out similar pictures on different websites. And um in the end um they had they did not had to migrate manually. So the idea was that they thought originally that 80 to 95% has to be migrated manually. And in the end this was all fully automated. So again, this is not an AI using it in daily users or visitors, but this is an AI that helps a team that needs to do a migration that stands in front of a problem. How can we do this much faster? Another one is freely give one of actually the company behind or that originally started Drupal AI. Um they have company the ESP group that's a train company um in Europe. If a train is delayed, you are eligible for a compensation. So if your train is delayed for two months, you get money back. Now imagine if a train is very heavily delayed or multiple trains, the website for the claims, they have up to 20,000 claims per hour that everybody in these train system wants to have a refund. And they were able to automate 90% with this with Drupal AI because of these claims they could automatically verify if this is a valid claim or not. Um so they reduced 80% of the processing time but also interestingly it increased the customer satisfaction meaning the customers they're already pissed their train is already delayed and now they were able because it was a super simple claim process they were able to show that it's much easier now or the customers are more happier afterwards. That's obviously big for this company, right? They didn't want to have the delay of the train, but we can go to them and say, "Hey, we can help you increase the customer sat um satisfaction." One of the last one is a company called Drop Solid. They worked with a customer Pitpa um and they do uh customer service. So there is a water supplier company and um customers just have random questions, right? What is in the water? How often is it tested? things like that. And even though they have all this information on the website, they still get a lot of customer service questions um in emails and things like that of information is already on the website. But we know people are just lazy or they don't know how to use or what for whatever reason, doesn't matter. They send an email with customer requests. And what they decide is like, hey, could we build a super simple bot that looks at the question in the email, goes to the website, tries to find a possible answer, and suggests that answer to the customer service team. So, it's not a bot that automatically responds to the person that asks it, but it's a supporting tool for the customer service agent. So, the customer service agent receives an email and there's already an AI that tried to answer it. And what they could show is that um they have a 15% reduction in the time that it needs on service requests. So it is faster and easier to do all of this. Again, this is not an AI that is used by the customers by the by the visitors. It's used by the internal team. And now the last one um is what we've built um within within Amazing. So we we all use CHT on obviously and the problem with CHT the privacy is extremely bad. They store your data forever. They share with other companies and all these things. We said we want to launch a CHTP competitor but with private so I know where the data goes and um we needed a way for having an MCP server right it's all built in MCP and Aentic whatever this means but we needed to use a tool that can do this and we also needed something that can do user management and we thought and thought and thought like what could we use could we build it ourselves and then we realized well Drupal already speaks MCP and Drupal has user management so we just use Drupal in this chatb clone. It looks like chatbt. You never realize that there's Drupal in the background, but Drupal has all these tools, these APIs already um available. And so we have it fully open source, right? It can be private and um we can now also connect to other servers. So another case where we where you can use Drupal not just to build websites, but a lot of other things you can build. And again, we didn't have to write all this code and Drupal AI already allows us. So, I really hope that this short overview shows you that Drupal A is already useful. And even more one thing, if you build something with any AI related in Drupal, please please tell us about it. We as the team behind the strategic initiative we agreed with the Drupal association that usually a case study is a I don't know raise a hand anybody ever submitted the Drupal case study on drupal.org work. Yes, it's quite a long process and you need to do a lot of information and we agreed with them that we're going to have a fast track. So that fast track is called Paul Johnson. So if you have anything you're it doesn't even have to be in production. If you do anything with Drupal AI or AI related in any way with Drupal, please talk to Paul. You find him on drupal.org. He's in Slack. He's an employee of Forex Internet. Please reach out to him because we want these cases that I just showed you. We want to show the world how Drupal AI works, what is useful, and what you can do. And he is going to help you get your case study out. And if you're not there yet, but you're like, "Hey, I would really like to use Drupal AI, we just launched on Drupal.org a demo system." So on this QR code or on drupal.org/ai/demo, /ai/demo. You can get a 7-day for free Drupal AI demo. You can try it out. There's AI keys behind. You don't have to pay anything at all. Um, and you can try Drupal AI. You can see all kind of different things what you can build with it because we obviously don't want to show the world what can be done. We also want all of you to use it, to try it out, things like that. Again, you don't need to pay anything. It's just a Drupal site that is spin up for you already configured and you can play with it right away. And then when you're ready and you actually want to use it on your own Drupal sites, Amaz trial there again. So you're seeing we really want to incur or help all of you use Drupal AI. You don't have to add a credit card and anything at all. So use if you have a Drupal site and you want to try Drupal AI in it, download that provider. You can use it for 30 days and um play with it. And if you have any questions or problems, reach out to us. There is a AI Slack channel. There's an AI channel in Drupal Slack that has a lot of people in it that work already in AI every single day. And that's it. Uh we don't have time for questions unfortunately because I use all my time. But I'm here the whole day. uh come to our booth, ask us if you have any questions and yes if you want to do a trial that is for the provider and this one's for the demo. Thank you., YouTube Video IZes7utRxzw, YouTube, 01kabwra3rwya342zwm88mdaah.