Event Tech Pod | MergeLabs

0:00 Hi, I'm Faith Keiser of Event Nation and Event Tech Pod and we have a very special guest. I'm really excited to share what he's been doing over at Merge Labs. This is Ali Taghavi. So, Ali, why don't you let us know what problem your product is solving and what the basic philosophy is behind it? >> Data fundamentally has been a problem we've been suffering from as event organizers. I've been myself organizing events for 20 years and in the day-to-day work as a planner, you're needing information 10 12 months before, but most people you talk to get stressed when you talk about 2 weeks ahead, especially when it comes to the coordination work. We're using generic fragmented tools like a monday.com, Asana for tasks, spreadsheets for building our schedules and then of course really good solutions around ticketing and other areas in events that has been developed. So, what we're trying to do is to create integrated solution where you can do all the coordination work. If our infrastructure that is dedicated for events, then we will also be able to really set us up for having that context rich data.

1:06 >> I see. So, you're trying to solve the problem of siloed data and one person being the source of knowledge for an entire event. >> Exactly. >> So, let's get into it. >> Hey, I wanted to talk about context graphs or knowledge graphs. This is how you want your data to look like if you want to start using Lems. And the way you can think about it, each of these things is connected to schedules, maybe to venues, maybe to run sheets. These are the elements that we tackle inside of the software. You use it as you would use a monday.com. For if you want to see how was the 25th edition of the same event, you have those contexts connected to it. And so, as we look at the actual software, summary board where you can see all the operational activities. So, for example, setting up the registration and attendee services, the venue teardown, speaker content management, and inside of each of those there's the act individual activities. And then you have them on a Gantt chart, and we have the tasks. And monday.com, the core element, the atom of that sits in tasks. The type thing that you use for everything. For us, that atom, that core element, is the activity. The activity is the key link that allows us to create a knowledge graph. And that's very important for any coordination work because we need to know the who, what, how, where, when information in a chain from multiple different perspectives. Like it's four-dimensional chess that we're playing within our heads as organizers, right?

2:38 >> Yes, every event planner feels that way. >> Activity is the core, and tasks are assigned to the activities. When I move it to complete, okay, it has a meaning. For example, it can start a chain reaction of another thing that can get started. So, somebody gets a push notification like this group has finished, you can come in and start setting up the audio visual, for example. Let's say this is the one where you have the conversations around that activity, discussing that problem, having that conversation, and nested in the context of that specific activity.

3:08 So, each conversation you create, you can link them to the activities saying this is conversations for this activity. Food, transport, cable, company, that's how you lose context, right? This is how you isolate context out from all of this. >> So, if I'm a planner on a team and I'm having a chat, can I tag a conversation to end up in an activity? >> Yes, exactly exactly. There's more things for the activities. You can define the the relationship between activities. All of this is data, it's added to the knowledge graph, and then LLMs can tap into in a structured way to be able to use it. So, again, the relationship between activities are provided there.

3:49 When you create run sheets, for example, or run a show, you don't just create them for each perspective randomly. You take the activities and you link them together. Like, for example, you want to add a map, but another person gets another run sheet, but they're looking at the same activity. So, these will move together. You build a schedule in different views. Here I have some activities that are linked together, right? These are for two different roles, but they need to see the same thing. You save time by you building the schedule once. You see it in all these different views.

4:19 >> So, is the AI creating these schedules for each individual? >> This is phase two. We haven't gotten that far yet with our solution, but this is also why we are building everything so hyper modular. We're basically creating micro handles for So, phase one is to help you just create this context search data. Phase two is going to be that we train our own LLM to do the work for you. The important part there is that AI You don't want to use tokens, for example, figure out conflicts. You have space conflicts, you have a user conflicts, and that's going to be way too expensive.

4:52 >> Yeah. >> You want the software to do that. AI is going to take away the clicking, the doing, the finding, the knowing, and just summarize it for you. That's what you need the LLMs for, rather than actually processing math for you. And this is what you need when you process conflicts or issues. So, that's coming, hopefully January, February next year, so in not too many months. >> Okay. >> But first step is to create the piping system for that to work. Then, of course, perhaps the most practical thing is each person can then receive their bundle of information, right? So, they get it whether as a link like that or you they can scan a QR code and get it on a mobile app.

5:30 >> Who exactly is putting this data into the system? >> Usually, you will have a few project leaders who are the main people who would create the infrastructure. We have a whole system user sub role management infrastructure that you can use. So, you first need piping of the whole thing, and then you can get there. Playbook about you can create playbooks out of it. So for example, this is an event with all the different schedules I showed you. If I go here, I can clone this event. Let's say this is going to be the 27th edition. All the tasks anybody has built. You can say, "Hey, I'm going to reuse probably most of them. Let's move them to plan." This event, I'm going to have this one in Berlin time zone instead of the late time zone, and then this is going to be next year. Let's start it in November. And in seconds, you have that already created. All the schedules are transferred, adjusted in time zone. You have so much time saving happening because of this playbook effect. And then And then going back to >> So you're saying a team would have a playbook that they use for most events, a certain type.

6:33 >> Exactly. And the system automatically adjusts and corrects the dates and stuff like that. You don't have to go back and unmerge them again. You All of that tedious work Oh, there's a direct link between an activity from previous clone and the cloned activity. So this is what you need if you want to say, "Hey, how did we do this last year?" Or past 3 years we've done this. Which one was the most efficient one? This and sort of questions you really want LLMs for. You can't answer them without this sort of contextualized data.

7:04 >> So the playbook is really the guide and the template. >> Exactly. >> And the assignments, if somebody leaves, somebody else comes on. >> Exactly. >> This is really exciting that this is available. So this is just so amazing what you built. >> We've been working hard for it. >> Yeah, it's it's ready. And this is the website mergelabs.io. I think if anybody want to see more details, they can come and book a demo. >> Thank you so much. We really appreciate your participation here on the pod.

7:34 >> Thank you, Faith. It was a pleasure. Have a good day. >> You too.