Event Replay: Building an AI-Powered Workforce at Brooklyn Sports & Entertainment

[00:00:00] Chris Nicholson: Welcome, everybody. Thank you for joining today's OpenAI Forum. My name is Chris Nicholson. I'm a member of the Global Affairs Team. We're excited to be joined today by the team from Brooklyn Sports and Entertainment, home to the Brooklyn Nets, New York Liberty, and the Barclays Center. Brooklyn Sports and Entertainment recently launched an AI lab to help employees across the organization explore how AI can support their day-to-day work. So today, we're starting a conversation with David Simbandumwe, VP of AI, and Keia Cole, Chief Digital Officer, to learn about the AI lab, why they started it, and how they're helping employees move from ideas to creating working prototypes that improve their day-to-day work. So we're gonna turn it over, after that, to three of the AI lab participants to share demos after we get through some questions. Now, let's jump into our fireside chat with David and Keia. Hi folks, how are you?

[00:01:12] David Simbandumwe: Good. Hi, Chris.

[00:01:14] Chris Nicholson: Great to see you both here. Can you both introduce yourselves? David and Keia, can you both introduce yourselves in your roles at Brooklyn Sports and Entertainment?

[00:01:24] David Simbandumwe: Sure. I'm David Simbandumwe. I'm the Vice President of Artificial Intelligence at Brooklyn Sports and Entertainment. So my role is really to drive our AI strategy across the organization, and in practice, that really looks like building the infrastructure that supports our AI use cases, managing the accompanying change, and then also ensuring that we enable AI capabilities for our staff across the organization. Then it also has a build component where we build solutions that are larger than what individuals can handle.

[00:01:57] Chris Nicholson: Super cool. Keia, want to introduce yourself?

[00:02:00] Keia Cole: Yeah, I'm Keia Cole. I'm the Chief Digital Officer here at Brooklyn Sports and Entertainment. I lead all of our technology, data, what I would call digital AI across the business and the arena. I'm lucky enough to say that this is my second time working with David. We worked together at a prior organization before he joined me here to lead our Gen AI efforts, so it's fun to be back together.

[00:02:29] Chris Nicholson: Very cool. What was the goal of the AI lab when you launched it?

[00:02:35] Keia Cole: So the goal really was to increase AI fluency across the organization. David had joined us and done a really robust discovery process around all of the AI opportunities across the company. We were talking with our CEO about how do we accelerate some of this work and really get more people across the company involved. We came up with this idea of the AI lab, creating structured ways for people to build AI solutions with the right incentives, and having an opportunity to share that with the rest of the organization. The thesis was pretty simple, which was we've done this really big slightly top-down process but how do we get people who understand the work the best actually building and working on these solutions.

[00:03:30] Chris Nicholson: Totally. Keia, I know you've gone through digital transformations before. Could you tell me what was your vision when you started this lab? What did you think concretely you might get out of this?

[00:03:45] Keia Cole: I think the vision was really just to sort of kickstart the organization and really to create some fun and some buzz around AI and really show people the impact that it could have on their work firsthand. I think the program in a lot of ways exceeded our expectations. We had this science fair style showcase of the 14 teams who participated in the first cohort, and I got a chance to thank everybody there. It was one of my proudest moments being at the company, and I've been here when the New York Liberty have won a championship. It was really just a super special experience.

[00:04:32] Chris Nicholson: That's really cool. Now, I know you probably get asked by other people in the org, maybe by some other folks in the C-suite, how does this connect to the strategy for BSE? What do you tell them? Maybe Keia, then David, what do you tell them when you're explaining why AI matters for your org?

[00:04:49] Keia Cole: In terms of how the lab connects to the strategy, this is really a key component of our skill building in AI. It's not the only thing that we are doing with respect to AI, but it's...

[00:04:58] Keia Cole: With respect to AI, it's about getting people more comfortable with the language around AI, the tools around AI, and trying to solve their own problems. Then I think it creates a pipeline of prototypes that David and his team can work on. And David, what do you think?

[00:05:18] David Simbandumwe: I think one of the things that we've noticed after doing the labs is just the quality of the work that comes in when people are requesting new activities. We get a detailed view of use cases that each team has, well thought out and well organized, and we start so much further ahead when we go to the build phase. So that whole enablement across the organization and just increasing the level of literacy has been extremely beneficial.

[00:05:50] Chris Nicholson: So it sounds like the lab is a place where people are really starting to get familiar with what AI can do. When they walk in with problems they're aware of, things they'd like to solve, frustrations, challenges, time-consuming tasks, it sounds like what comes out are great ideas for where AI can apply to those areas. Do you agree with that?

[00:06:11] David Simbandumwe: Absolutely. I think what the lab does, because it's more of a virtual lab, has really put structure around that. In the pilot phase, we had three weeks of workshops on how to define a problem, design a solution, and then build toward that solution at whatever level. Then four weeks in which people sprinted to build, with weekly coaching sessions and cohort meetings by level, really creating the time for people to address a problem they are working on. Once people solve one problem using AI, it becomes so much easier to identify the next problem, create a solution, and walk through the process again.

[00:07:07] Chris Nicholson: It's a motion they repeat, practice, and get better at. We see that in sports a lot too, I hear. So, as other folks in your organization are learning about AI and its applications, what would you say you have learned from this experience with the lab?

[00:07:26] David Simbandumwe: One thing I learned, which was super surprising, is when we collaborated closely with HR and our comms team about bringing this to life. We needed to have levels so anyone could participate and allow for cheap and cheerful solutions. What was surprising was the level of ambition and the challenging problems people wanted to solve. These were some super ambitious projects, and it made me really proud to be part of this organization.

[00:08:16] Chris Nicholson: That's really cool. It's great to see excitement around the technology and applying it to problems. Every month or so, someone stops into my office to show something interesting they've achieved with AI, saving them a lot of time. You've both been through digital transformations before, previous waves of technology. How is AI different?

[00:08:44] David Simbandumwe: I think it's the speed at which things are changing. I've talked to Keia a lot about this. Every three months, we have to rethink our delivery methods because the tools are so much better. For the next iteration of AI Labs, we have a whole new suite of tools for people to deliver on. The solutions are much more complex.

[00:09:14] Keia Cole: For me, the key difference is accessibility. It's much easier for someone to pick up and build in this environment than in prior transformations I've been a part of. Previously, you had to come as a business stakeholder, ask the tech team to do something, and then wait for it to be done. Now, people can at least get to a prototype more quickly and move to production faster.

[00:09:49] Chris Nicholson: Yes, and I've heard from others involved in transformations that trust is a super important...

[00:09:56] Chris Nicholson: ...of bringing AI in and getting people to accept it. Would you say that? Is that what you see?

[00:10:03] Keia Cole: Yeah, I mean, absolutely. As part of the AI lab, you know, for each level, there was, you know, money essentially that people could win. This notion that if you're generating value for the company, we will return value to you. We had external judges that evaluated the solutions and ultimately everything was so great that I think everybody got like a small thing for participating and kind of getting to the finish line. But I think what that signaled to people was building this fluency as an investment in our organization and in our workforce. And I think that was a pretty powerful message that as people saw what other people built, it was like, okay, yeah, I want to do this next time. And we have more people joining now that we're in a kind of always-on format. So yeah, I think sort of rewarding people for generating value was one way in which I think it felt safe to experiment in like something that the company really valued.

[00:11:06] Chris Nicholson: Yeah, I love that thought of AI as an investment in them. So you brought in employees, like frontline workers who are really close to the work. Why did you start there?

[00:11:19] David Simbandumwe: I think just naturally you get better products if you bring people in who actually work with the activities day to day. They've actually seen the pain points, they've seen the gaps between the systems, and they actually know what use cases fit together and which use cases add value. And it also helps a lot with the change management if they've been a part of the process all along because they've had a hand in the solution and it's a lot easier to see the value and to adopt the technology. If you've seen it grow, you've seen it come together and you understand what trade-offs were made as you're building the solution.

[00:11:53] Chris Nicholson: Yeah, super cool. So did you two see, were there opportunities your frontline team spotted that really surprised you that other people might've missed, that they surfaced for you?

[00:12:07] David Simbandumwe: I think conceptually you recognize that there's always savings associated with automating pieces of a workflow, but just actually identifying what those are, whether it's the reconciliation process for expenses, whether we're gonna see some examples of content creation where you know that there's an opportunity there but just actually putting a finger on it and actually identifying what that specific workflow is. I think that's one of the bigger benefits of having frontline people attached to it.

[00:12:33] Chris Nicholson: For sure. Keia, how about you?

[00:12:40] Keia Cole: I think one area that, I don't know if it's necessarily surprising but just seeing the full benefit. There was one project which is not in the demos today around using AI to predict when we should cook food for events in the arena, how much food, when to fire it. You know, you're trying to feed 17 to 18,000 people at a time and how do you use AI to solve that problem. And you're feeding 17 to 18,000 people across concessions, different club spaces. And so that was one where I was like, oh yeah, this is an operational use case that people have been doing for a very long time but we can make this more effective using AI. And then, yeah, I think the demos not to steal their thunder are really some good like surprising examples too.

[00:13:33] Chris Nicholson: Yeah, yeah. And just to zoom out for or add context for our audience, David, you mentioned reconciliation. What is that process and why is it hard?

[00:13:41] David Simbandumwe: It's not necessarily hard. It's just comparing two different spreadsheets and trying to align expenses together. So it's just a tedious process that somebody has to do on a frequent basis. Whereas you can have AI go through and take the first pass at it and then all you're doing is fixing the items that fall through or the items that you have low confidence in.

[00:14:01] Chris Nicholson: So I'll bet there's a lot of people in the community who have considered creating something like an AI lab, considered trying to help their organization move along. What for you two, what are the tips and tricks that help enable employees, that from the lab enable the employees to get to success more quickly?

[00:14:20] David Simbandumwe: I think the one thing, the two things that really helped out a lot were we started with really two weeks of training and we taught people how to recognize a problem, how to identify, how to break down a problem into its elements and then how to actually turn that into architecture and categorize that in terms of a project. And then after that, we spent a lot of time with each of the teams, working through office hours as well as having one-on-one coaching sessions. So I think that support was helpful, especially with people who weren't quite as confident when they started the process.

[00:14:54] Chris Nicholson: How did you use that lab to make it safer for employees to experiment and learn in public?

[00:15:00] David Simbandumwe: Yeah, I think what was really helpful is we had a lot of top-down support. So the message was constantly reinforced by all of our leadership team, the importance of AI, the importance of this program, and they actually put resources behind it from an infrastructure perspective, as well as from prizes associated with winning each of the levels. And then after that, it's spending a lot of time with the individual teams because naturally when you're doing something new, there's that point in time where you're just unsure about the next step, and just ensuring that you can spend time with the teams, walk them through that, convince them that everything will be demoed fine, and that even if you're not 100% successful, there's a lot of benefit in the learnings associated with going through the process.

[00:15:49] Chris Nicholson: Totally, totally. Now, I would love to hear from you too, you've got an AI strategy, it's got more steps ahead. From this lab, what have you learned that's gonna shape that strategy going forward?

[00:16:00] David Simbandumwe: I think the biggest thing that we've learned is just not to shape the strategy in a vacuum, right? So building this pipeline of prototypes from people across the organization and having people across the organization influence that strategy. It's not just going to be something that comes from us as a technical team if it's going to be successful. And one thing we've been actively doing is looking for more opportunities for people to build on their own, and what does that look like from a tooling perspective? So it's not just conversational AI tools that we're going to give to our employees, but what will give them the ability to build further up the stack than they have previously? What does that look like from an infrastructure perspective?

[00:16:50] Chris Nicholson: Yeah, cool. Well, first of all, thank you both so much. We're going to get back to you with questions from the audience. But first, we're going to go to the demo. So we've got three demos today. The first demo is going to come from Bridget, and she's going to show us Operation Sellout. Bridget, over to you.

[00:17:11] Bridget Schelzi: Hi, everyone. For this year's AI Lab, my teammate Rhys and I created a custom GPT called Operation Sellout, which is a copy engine designed to drive single game ticket sales for the Brooklyn Nets. The problem we're solving is that this is currently a time-intensive workflow for our marketing department. It takes our copy team over 50 hours per season to create individual content across channels for each home game. It can also lead to an inconsistent voice with copy being developed across formats, channels, and timelines. Consistent messaging is key to brand authenticity for our audience. The manual process can also cause a content bottleneck with 500 required copy fields per season.

The custom GPT we created allows us to upload the season's home game schedule, and using the established guidelines we input, the GPT will create custom copy across email, app, paid social, as well as broadcast and in-arena reads. So now I'll take you into the back end of the GPT so we can take a look at the solution architecture. Within the instructions, we've inputted a variety of governance rules, including priority generation, order, what not to say, themes to include where applicable, as well as the content fields required for each channel. We also uploaded a variety of knowledge files and brand guidelines so it can generate structured copy, as well as the home game schedule for the season and a demo metadata sheet for this presentation.

[00:18:43] Bridget Schelzi: Over here, I asked it to send me all of the copy for the first five home games of the season. Within less than a minute, it's generated all the copy fields that are ready for our review so we can approve it before it goes live. Operation Sellout has cut down on a ton of time for our copy team, and we're excited to implement it next season and we hope to apply it more broadly across the organization as well. Thank you.

[00:19:07] Bridget Schelzi: And I'd like to welcome Chris back.

[00:19:10] Chris Nicholson: Thanks, Bridget. OK, we've got a second demo, it's called The Milestone Tracker, and this is from Jacob. Jacob, over to you.

[00:19:18] Jacob Wolf: Thanks, Chris. Brianna Stewart is within striking distance of becoming the fastest Liberty player to reach twenty-five hundred points, a key moment for the franchise. In the old world, nobody on the content team knows until it's already happened. Everyone is scrambling at the buzzer to build the graphic, write the post, and tell the story. Too often, we would just miss it altogether. A milestone is a notable career franchise mark. Think a player's twenty-five hundredth point, a new all-time leader, or the fastest ever to reach a marker. For a team, each one is a ready-made story, a graphic, a post, or a partner moment. Miss it, and the moment's gone. Tracking these...

[00:19:52] Jacob Wolf: And it's gone. Tracking these used to be a manual reactive 12-step process filled with hours of pulling numbers across spreadsheets, lots of room for human error and no shared source of truth. So we built the Liberty Milestone Tracker, a real-time command center that watches every Liberty player against WNBA and franchise records and tells us what's about to happen before it happens. That 12-step workflow collapsed to five approval-ready steps and what was previously one to two hours now takes about five minutes, up to 24 times faster and feeding one source of truth for stakeholders across departments. Milestones are auto-ranked by editorial priority from tier zero franchise-defining stat like a new all-time leader down to a tier three nice-to-have post-game note. And the timing isn't a guess. A pace model blends each player's career rate and recent season form to project games to reach ETAs. The transparent rules-based tiering engine also classifies how big each moment is with the appropriate content response. It also pulls additional context automatically for applicable milestones. For example, Breanna Stewart will be the fastest to 2,500 points in franchise history, passing Cappie Pondexter's 130 games to reach 2.5K. The moment it's reached, the previewer builds a graphic in the Liberty's brand style with different sizing and formatting options and fires an alert to the team in Slack with a copy and creative already ready to publish in different sizes for social. We also built skills on top of the data, pulling from the NBA courtside website to get pre-game, in-game, and post-game alerts and looking into additional angles. That gets us ahead of in-game moments that we could never anticipate before tip, and it layers on context, like when a player is among the youngest or fastest in league or franchise history to get there. And the same engine powers the business metrics tab so stakeholders across the organization share one set of numbers and always have the latest. The Liberty Milestone Tracker is powering new possibilities for storytelling in sports and it's never been easier for us to highlight all of our players' achievements. Thanks everyone and welcome back, Chris.

[00:21:49] Chris Nicholson: Thanks Jacob. I love how y'all are using AI to be so fast and responsive. Just like your athletes are doing on the court, you're doing it with data, engaging your fans. And that leads us actually to demo number three. It's Gideon and it's presented by Mariia. Mariia, over to you.

[00:22:08] Mariia Vasylenko: Thank you, Chris. Yeah, this is Gideon, an AI-powered insight engine that transforms data into strategic narratives. The reason why we built this because the minimum turnaround time from data requests to insight delivery is two weeks. And we have only one team who have an access to raw data and output really varies depending on individual analyst expertise and time they have available. So we really wanted to close the gap and we wanted to streamline the data delivery process. But also we wanted to understand the why behind the data. We wanted to gather deeper insights. So we wanted to benchmark this data against historical patterns, look at behavior and to deliver a narrative our team can act on immediately. So the way it works, you ask a question about different types of data, AI pulls data from different data sources, first-party data, third-party data, VAP, AI generates a plan. Data analysts can review the analysis, preliminary analysis. They can redirect. And once they approve, the report can be delivered to the person who requested it. So let's see it in action. It is gonna be based on synthetic data for demo purposes. So here we can ask any type of question. Let's say, if New Jersey is still a meaningful growth segment for the next. That's a very typical request that we send to our data analysts. Right now, AI pulls the data from different data sources and data analysts can review all the data sources, they can review preliminary analysis and they can redirect the tool. They can provide the feedback. And once they are happy with the output, they approve and generate a report. Report is very user-friendly. You don't need to be a technical expert to understand it. We have late insight, interactive dashboard with key data points and recommended action. Thank you so much. Back to Chris.

[00:24:08] Chris Nicholson: All right, thanks Mariia. Well, we're gonna go to audience Q&A here in a sec. So I'd like to invite Keia and David back on. Hey folks.

[00:24:17] Keia Cole: Hi.

[00:24:19] Chris Nicholson: So actually, before we get to audience Q&A, I remembered a question I had that you shared a little bit with me about before the show. So you mentioned a couple use cases. In your data-rich business, there's a couple use cases that affect the fans. I think you mentioned the ticketing workflow and one other. Can you dig into those with me for a second? How is, you start with data, you apply the AI, how is that affecting your business and the fan experience?

[00:24:49] David Simbandumwe: Yeah, so I think one of the ways...

[00:24:50] David Simbandumwe: Yeah, so, I think one of the ways that that affects the business overall is it makes it easier to pull analysis and it democratizes access to the data. So instead of having to go to a single team to actually pull reports, pull analysis, all of a sudden you can do it on the back end, you can do it as a frontline worker, you can actually access the data, ask it questions, and you don't need to do a lot of technical work to pull the information. And so we are working on solutions that make it simpler to access our rich data stores, building natural language query engines, and we're actually looking at what Gideon would look like as a fully functioning product with a semantic layer underneath it.

[00:25:32] Chris Nicholson: Got it. And Keia, how do you see it?

[00:25:35] Keia Cole: Yeah, I think a lot of what we're doing, we really started focusing on our ticketing business, it's like the core of... A big part of how we generate revenue and how do we better enable our sales reps. And there, I think we're putting them in the position to have a more personalized, more thoughtful interactions with people, easier prep for the meetings that they have. And so just really making the team more effective and more efficient, essentially, again, with bringing together various data sources into their prep.

[00:26:10] Chris Nicholson: Yeah, got it. So personalization, making communications more meaningful for their prospects target audience. And then, David, what you were saying it struck me that you're just really, by enabling the org, you're making every, you're raising kind of the intelligence that everybody has access to with the data and the AI, and you're making your organization more responsive to things, right, as they change. Is that right?

[00:26:31] David Simbandumwe: Exactly.

[00:26:33] Chris Nicholson: Okay, we're gonna dive into audience questions because we love to connect you with our community and bring up the things that are on their mind. So I've got, Andrea Jermanson, I hope I'm saying that right, board advisory members at the Startup Venture Program at Warwick, at University of Warwick. She's asking which use case surprised you most in the way it, by creating value in sports and entertainment?

[00:26:59] Keia Cole: I would say the milestone tracker. I think it's something that you take for granted how that gets tracked, and then all of the work that goes into approving those posts. And so just seeing what that team was able to build, and truly, kind of a cross-functional team bringing our content team together with the PR team. And now on social, whenever I see a player graphic celebrating a milestone, I know how we got there and how we got there quickly. And it's just a fun, it has such an impact on the fan experience in a way that I hadn't even like thought about, oh, we should speed this up and make this faster. So that's one of my favorite ones, actually.

[00:27:49] Chris Nicholson: Yes, and Crush.

[00:27:50] David Simbandumwe: Yeah, one of my favorite ones was Penny, which was really an expense reconciliation system. And traveling a lot when I was younger and early in my career, and having to do expenses on a Sunday night, just being able to do it in an automated fashion just kind of resonates with me.

[00:28:09] Chris Nicholson: Yeah, I'll bet. Okay, we've got a second question. Jason DeLuca, he is at Crossing Point IT Solutions, owner and principal. What roles or workflows at Brooklyn Sports and Entertainment have seen the clearest AI-driven improvement so far? Where do you see the fastest gains?

[00:28:27] David Simbandumwe: So interestingly enough, we've seen a lot of gains in the code AI space. So just being able to go from starting a project to getting something that you can test has really been accelerated with the tools and the availability of them. And just how smart the models have gotten and the harnesses and the ability to generate code and move you from the starting line or a prototype to something that you can actually move to production really rapidly.

[00:28:57] Keia Cole: Thank you. I think we're seeing a lot of benefit on the sales side. That was part of our strategy after doing a big discovery process. This was to pick the area of the business that we thought was most ripe to take on AI transformation. And I think we've seen not only results with respect to revenue and efficiency, but also just that organization starting to build things on their own and having that be a part of how they work. And so having that as kind of a proof point for the rest of the organization to say, hey, if you focus on this area, this is, you know, these are the benefits. And that's letting us kind of move to focus on other business domains.

[00:29:43] Chris Nicholson: Yeah, you got a great feedback loop with hard numbers, really easy to prove the value.

[00:29:45] Keia Cole: Yeah, totally get that.

[00:29:48] Chris Nicholson: Okay, we've got Lana Romanova.

[00:29:48] Chris Nicholson: Okay, we've got Lana Romanova, Python and Machine Learning engineer, aspiring AI researcher, and an educator in Applied Health at Skillbox. So have you seen non-engineering employees become some of the strongest contributors to AI workflows once they deeply understood prompting a process design, which is not something everybody understands?

[00:30:09] David Simbandumwe: Absolutely, and I think for some of our employees it's gone beyond that. So they've started to use AI to link systems together to solve broader problems that are entire workflows instead of just prompting and getting answers back that you can incorporate into your workflow. How do you automate the entire process—so multiple steps incorporating multiple systems to solve a problem? Keia?

[00:30:33] Keia Cole: Yeah, for sure. I think one of the things that was really important for me as we were building the AI Lab was that we really needed to give people foundational training before we asked people to start to build. And so I think David did an excellent job of leading a series of workshops that really gave people the scaffolding that they needed to start to solve problems, especially if people were new to the space. I think the thing that we are experimenting with now is having those workshops be a bit to the level, so really creating entry points for people where they're completely new to this and then other materials that are tailored for prior participants or people that are kind of a bit further along in their journey.

[00:31:27] Chris Nicholson: Levels.

[00:31:24] Keia Cole: Yeah, really.

[00:31:25] Chris Nicholson: So meeting them where they're at.

[00:31:26] Keia Cole: Yeah. And it feels like teaching, to follow up on Lana's point, just teaching people to think about processes and like think algorithmically about what takes place step by step in a workflow is a powerful unlock.

[00:31:39] Chris Nicholson: Okay, we've got Desmond Lavelle. Desmond, I hope I'm pronouncing it right. Lavelle, EVP and Executive Creative Director at BBDO. So he says you mentioned bringing in third parties to validate your AI initiatives. You talk a little bit more about that and what does and does it help in selling internally and optimizing implementation?

[00:31:59] Keia Cole: Yeah, so as part of the pilot phase, we had the three different levels and at each level we communicated pretty early on what the prize would be for each team. And I felt pretty strongly that we should bring in external people so there wasn't an internal bias of we had executives doing it, like they were scoring their teams higher or not. So having some objective aspect to that. So we had somebody from the league who works on AI that we work closely with. We also had a couple of other technical vendor partners come out and serve on this judging panel. So it was really fun for them to get to see what we're doing and get to know the organization a bit better. I think the thing that we learned from that is that all happened in this science fair style demo day. And people really wanted, I think, more feedback and dialogue with the judges about the work. So we probably would do that differently next time. But I think people liked the idea of it not just being an internal thing, but having some engagement with people that are working in the space more broadly.

[00:33:31] Chris Nicholson: So important to feel like learning from industry.

[00:33:33] Chris Nicholson: Totally agree. And what I'm hearing is they wanted mentorship and connecting with people maybe further down the line in terms of their experience, which I see a lot too. It can be very encouraging to hear from people further down the line.

[00:33:48] Chris Nicholson: Okay. We've got Daniel Green. He's lead at Kansas City AI Collective. What are you most excited about when it comes to AI enablement?

[00:34:02] David Simbandumwe: I think autonomous AI is really what's exciting for me. So if you look at a lot of the solutions that we've built so far, they involve somebody actually actively using the AI. But now we're starting to work on autonomous agents that run in the background that respond to Slack messages, and eventually actually build code and process different items. So the ticketing solution we kind of described earlier, there's no reason why we couldn't build that into an autonomous engine that just runs in the background and just presents you with material every day or on a sequence. So I think that's what's really exciting is the fact that this can become a true coworker and work in the background.

[00:34:42] Chris Nicholson: Yeah. Really cool.

[00:34:44] Chris Nicholson: Well, thank you both so much. Your team members too, you've been serving as mentors to me, learning how you're applying it in your organization. I'll bet you're doing it for people in the community too. It's always so instructive and so valuable for us at OpenAI to hear how, to learn from you how our tools are used and how they're helpful. And I think it's valuable for the community to come together and hear from organizations that are forging a path and really setting an example. So I wanna thank you both again. I'm very grateful for your time today. Thank you for sharing all those learnings as well as those demos.

[00:35:25] David Simbandumwe: Thank you, Chris. Appreciate you having us.

[00:35:27] Chris Nicholson: Great, Keia and David, thank you. And I wanna say to the community, thank you all for joining us. Really this forum depends on you and your engagement makes it meaningful. So we're so grateful for you. Thanks for joining us today and we will have more events soon where we dig in deeper into new domains in the AI applications. So we'll see you all soon.