PhotoShelter’s AI Summit:
Is AI worth the money?

I I know this space very well, and I came to work with PhotoShelter about seven months ago. And before that, I was with Texas a and m University for ten years in marketing communications. Communications.

Very near and dear to my heart, and I will I will stop talking and let these other folks say who they are because they’re just a lot cooler than I am nowadays. But y’all go ahead. Cody, you’re up first, man.

Cody Silfies: I don’t know if cooler than you is is true or not. We’ll see.

Hi, everybody. Cody Silfies. I’m the creative director at the University of Vermont. For those who don’t know, we’re a mid sized university with a little over fourteen thousand undergraduate and graduate students nestled in the mountains, the green mountains up here in, Burlington, Vermont.

And in my role, I oversee the visual identity of UVM, including design, photo, video, and, working with internal and external stories and campaigns and and visualizing them and managing all of the assets for all of those things for our office, which is the division of strategic communications, and helping the university as a whole tell our story with those assets.

Michelle Sammet: Great. And, I’m Michelle Summit. I work as athlete relations manager at World Athletics. I think there’s a large US audience. When I say athletics, I mean track and field and road running. I look after anything that is content distribution, education for athletes, consultation with athletes, and it’s really all about how can we make sure that we leverage our athletes as the main assets that they are in our sport.

Sam Craft: Awesome. I am excited to get talking today. When when people ask I say people, when leaders ask about why is AI worth the money, this very much lives where I used to live in my world, specifically with photo shelter because with the first time I ever heard about AI, five, six, ten years ago, whenever it’s been, I had to I had to do these pitches to my leadership to get the money buy in for that kind of space. Let’s let’s talk a little bit about that, cost savings, speed, quality. What what are those things for y’all?

Cody Silfies: Yeah. I I’ll start.

For me, higher education is sort of this complex relationship with AI at the moment. We were kind of at the forefront of it with research and development, and then all of the fear around how it was shaping higher ed sort of creeped in, and now there’s a lot of concern about, what does the future of learning look like with this tool? And us as staff members are kind of caught, right in the middle between these two things.

And whether we’re slower

Sam Craft: Between liability and progress?

Cody Silfies: Between mom and dad. I think as as we, start exploring it, we’re, kind of, right behind the curve a little bit in some ways about how we use it to to work, basically.

And I think we’re finding internally that the question around ROI on cost, is it worth it, really centers around, efficiency of time and solving staffing resources. For folks who maybe have worked in higher ed or are currently working in higher ed, resources, especially in a Marcom team, can be very light sometimes or tight because a lot of the resource of the institution go towards the academic side or the learning, as it should be.

But, but that that resourcing shift is never gonna change. It’s only gonna get tighter, especially as higher ed evolves and the demographic demographic cliff comes and, all of these other realities that we face happen. We have to figure out how we as a Marcom team can provide what our clients, internal and external, and our leadership expect from us, which is sometimes agency quality work with a bare bones team.

And I think we’re finding that AI can solve some of that. We may never have a project manager. We may never have assistant, but if AI can solve some of those things, then it offsets a lot of the work and allows us to work faster, more efficiently, and and really just, prevent burnout of our existing staff who are wearing lots hats to do a lot of things.

Yeah.

Sam Craft: I I think that the three big letters there are ROI. That’s how we always get things done. I know in Michelle’s space with her org, I I think we’re gonna hear some really good represents to that. Yep.

Michelle Sammet: Absolutely. And I think, Cody, you you mentioned human resource there. For us, a lot of it is about scale. When we’re talking about our championships, we’ve got our world indoor championships coming up this weekend. We have about seven hundred athletes competing in that. When we look at our outdoor world championships that we had last year, for example, in Tokyo, we’re looking at about two thousand athletes.

When as I mentioned, we’re looking at how we distribute content, how we capture content, how we serve the athletes, the member federations, and everyone. Everything is about scale. And if we look back at the workflows that we used to have in place, a lot of it was manual, and it required a lot of human resource. We would have photographers capturing work.

They’re editing their work, having to tag work manually, and then also a lot of distribution was happening manually. A lot of this work has been simplified now, and it actually allows for everyone that would normally be involved in this manual task to be able to free up creative time and actually create more content that can then be distributed rather than having to spend half of your night after an event tagging photos and then manually sending them out. For us, the big, big difference has been actually freeing up the workforce to create more, better quality, and to distribute that faster rather than spending half of our time actually making sure it’s tagged and then sent to the right people.

Sam Craft: It is the perfect transition into our next thing because we’re gonna talk about how PhotoShelter AI plays a role in your in the ROI, but also the workflow process and all of those things. And I know for me, in my world, Michelle, you said it the best. I I I pinged it a a thousand times on sales calls even before I worked here. It’s it’s giving you that time back to go create and not sit there behind the screen and and post process. I I don’t think people really realize how much time is lost in those conversations with their creative teams and their creative members that they’re not on on the, on the ground doing those things.

With that, Michelle, I think I think we have some pretty slides here that you’re gonna walk us through and and talk about your process.

Michelle Sammet: Yes. I mean, as I mentioned for us, the game changer has been people ID in PhotoShelter. We were a little bit hesitant to start using it just because we didn’t really quite trust the system because there had been a lot of talk about AI maybe being being skewed towards certain ethnicities, for example. What you see here on the screen is just a small selection of some of the work that we’ve been doing in this space in capturing the athletes.

As I mentioned last year, we had our world championships in Tokyo, about two thousand athletes from two hundred countries roughly competing all around the world. Track and field is one of the most global sports in the world. And as such, as you can see on the screen there, our athletes come from all corners of the world and as such look different. One of the biggest challenges that we have as a sport is trying to serve as many athletes as possible.

We normally would capture a lot of athletes of the winners, of the medalists, so forth, but what we really want to do is be able to really provide every single athlete that competes at the world championships with the opportunity to share their experience with the fan. And, obviously, they can’t capture themselves during competition, it’s our job to serve them with the pictures from that event. What we did last year was a big project that was cross departmental, that was coordinated between us, our communications team, the event presentation team, the broadcast team, IT, and competition.

And we set a workflow whereby throughout the accreditation process, we had photo studio set up. We were capturing every single athlete. Those images were then used across in stadium on the big screens for introductions. They were used in broadcast graphics.

They were used on the website to update the athletes’ profiles.

But and this is where PhotoShelter comes in. We also use those images to train through PeopleID to ensure that the athletes were named and tagged in the system that then ensured when they were competing, they were automatically tagged, and it didn’t require loads and loads of back end work from our editors, from the taggers. Because in the past, whenever we used to work, especially when you have multiple people in an event in a race, it’s hard for every single person to be captured. One of the examples in this space, for example, is the road events, the marathon, or even the ten thousand meters on the track where you just tend to have a much larger field.

And in a marathon, if you have seventy, eighty athletes competing, especially at the start, a regular person manually tagging would maybe tag the front row of five athletes, but they wouldn’t tag every single one in this picture. And what we have seen is just the number of athletes that we were able to serve with content from these past championships compared to previous championships quadrupled just because we were able to identify far more athletes through PeopleID and through automatic tagging. And that’s for us been a game changer that we’re as I mentioned, this coming weekend, we’ve got our World Indoor Championships, and we’re repeating the same project.

We’ve got two FODO studios this time around and two accreditation centers to make sure that we’re capturing all the athletes again. That’s just, one example here.

Then secondly, I touched upon the distribution, and that’s exactly where then a second photo shelter product comes in we work with socially.

And we again, back in the day, I used to have to manually send out photos. I would often have to manually go back and find athletes photos because they signed up after their competition, and the system by that point had already distributed the pictures. Then I would have to manually go and find them in the system and manually send them the content. This is an example of the rules that we have sent up or that we had set up for Tokyo.

We have obviously got the the folder in PhotoShelter, that identifies where the content’s coming from. We have identified which groups the athletes could come from, and then have it set up that any auto or any photo that contains any of the athletes in the photo then automatically sends it to the athlete, adds a caption with the suggested photo credit, gets a text message to the athlete. It adds it to albums, to various tags that then, at a later stage, help us with getting data and analytics and seeing how successful the program is. And one of my favorite features that I’ll I claim that I asked if we, PhotoShelter, built this for me was in the past, we would always have to duplicate a lot of photos in the photo shelter library because the photos would land in photo shelter, but they weren’t tagged yet. And the system would only recognize them once, I was moving them once they were tagged into a new gallery that would then pull it through into Socially.

But since the introduction of the send to Socially function within PhotoShelter, we now have this little apply only when content is sent via send to Socially functionality, which means I can keep the one library or the main folder within PhotoShelter and then just hit that send whenever I know that content is tagged. And then also if an athlete signs up after the fact that they have competed, I just send the entire gallery again, and it just only sends it to those that haven’t received the content yet. For us, it’s massively sped up the process. I remember the days when during the Tokyo Olympics, we were up until three, four at night tagging and manually sending and having to resend after championships, and that’s massively improved, and I get to sleep maybe at one or two AM these days.

Sam Craft: Michelle, that was awesome.

And I I might freelance hire you to do my socially demos on my sales calls because that was a really good example of exactly how those flows work and and what that’s in real time. And I I use basketball for people ID in in some of those spaces. I’m gonna start using track things because I think that’s a much better example of showing forty five people getting ID’d all at the same time. I think that’s awesome.

Cody, what do you got, man? I they don’t tell me what you have, I’m excited to see it.

Cody Silfies: Yeah.

It’s great hearing the way that Michelle’s using it because we’re using, PeopleID in sort of a similar fashion but for a completely different purpose. We, manage an experts database here at UVM with our director of media relations that anytime, a topic comes up and a news organization may wanna talk to someone at our institution about it, he has an easy way to sort of see if we have assets of that individual. Once those faculty have gone through training, I get a notification that they’ve they’re in the experts database on our system, and then I add them to PeopleID. And, and

and I use our in house AI software or our AI platform, Copilot to sort of automate an additional layer of work on this where I could get maybe about twenty names in one batch, and I can use Copilot to just generate the metadata that I need to just put into people ID that once we do have photos of them in the system, that metadata with their college or professional affiliations or subject matter that they are related to is automatically populated and keyworded within PhotoShelter.

And then we create smart galleries for those individuals and, provide the the collection of those smart galleries to our director of media relations that when he gets an ask from a media organization or there’s an interview happening with one of the folks in the experts database, we can easily see if we have assets for them. And that’s especially helpful because for many of the people, we don’t have photos of them. We’ll create the smart gallery anyway because they’re in PeopleID now. And if any of our decentralized communicators upload photos to our system of those people henceforth, it’ll automatically populate that folder with the most recent assets of that individual.

We constantly have sort of, an influx of, potential photos that sort of populate those photos to choose from. And it kind of removes me from the process of that grunt work of, the five emails that we need to exchange back and forth between the media relations person and me, they can go directly to the source, find a photo that they think will work for the organization they’re working with, come to me and make sure that it’s great, and then it’s, good to go rather than me having to sort of find, an asset.

Another workflow, that we often use is the the search is, a game changer for us, AI search in PhotoShelter. When I first joined the team, we had just merged a bunch of different photo shelter accounts into one. And the tagging structure and organizational structure was, Christmas lights all tangled together. And, I often get emails from faculty, staff, random people throughout our institution asking to find the original high res version of, a little clip that they screenshot it off the website or something.

And before, I would have to, put on a detective hat and just, dig for sometimes hours to see if I could find this thing for a person. But once the the AI search function was enabled, I could just say, person in brown coat and purple gloves. Or in this case, I was person in a white lab coat with purple gloves, and we could easily find the assets that we needed if we had them in the in the system. The same with my colleagues in enrollment marketing, they often are looking for vibes.

They’re we want this page is about, community, and this page is about happy students laughing. This one’s about studying, we want something serious and academic. And being able to search concepts sort of pulls in from our really deep, asset library a lot of of, content that they that they can, find on their own without needing to involve me every single time again that I’m not at the center of every process of gathering options for things that I can sort of be the creative director part of it, and the people who are putting together the pieces can sort of do their part of it.

Those those two things in a nutshell have, been the game changer for us with PhotoShelter and helping improve a bunch of workflows and reducing, the amount of Teams messages and emails that I get around back and forth on an image, asset finding.

Sam Craft: I I cannot agree more. I I think the two biggest things that I’ve always loved in my experience with the AI that exist here in PhotoShelter has been the AI visual search, which you just went over very well. I always I I came from agriculture, I would use, brown cows and green grass fields. And, I I tell people all the time, the more I said, the more angry you get at it, the more aggressively specific you get, the better the system likes you and the more it’s gonna find exactly what you’re looking for.

I I tell people all the time, go ahead. Try to break it. You you can, but try to, because it’s gonna it’s gonna it’s gonna help you more. With the fun things that we’ve talked about so far, Michelle, I’ll start with you.

How is AI, how has it changed your org as far as, the features that you have and things that you do? And and if it was gone tomorrow, how would it change your world?

Michelle Sammet: I mean, I don’t think it would we would completely change the way that we work. I think as I briefly mentioned earlier, and I think somebody touched upon it, in in the previous session as well, I think it’s just freed up more time for creativity. One of the big things that we have seen, for example, is that the couple of times where somebody tried to dabble with AI when it comes to actually creative content, it didn’t go down very well, and we didn’t like it. And we we got that feedback from our fans as well.

We have an incredibly talented creative team. We’ve got really, really talented illustrators, graphic designers, writers, and forth. We don’t need AI to help with that. What we have seen is that actually us being able to focus on that and not spend time doing exactly as as Cody mentioned, trying to help the rest of the team find, photos or find certain assets within our libraries or, again, help provide that content to athletes and and everything taking so much more time on that side.

I think that’s been the big game changer for us. And and just to to pick up from what Cody mentioned there, I know for a fact our graphic designers absolutely love the search functionality because they might be needing a specific angle of an athlete because they’re doing a big composition graphic, and they just need this one particular athlete, or they’re looking for the athlete in a specific sponsor kit because we know that they’re going to be wearing that kit in a competition. And rather than having to remember where did they compete, having to go through the website to look at their results, how did they finish in that race, would we have a picture of that, just being able to look for that athlete and look for them wearing that particular color or kit, whatever it is, it’s it’s revolutionized the way that they work, and it just means that they spend less time searching for things and more time creating.

And that’s where I think we would just have to go back to working slower again, and and requiring that additional time to build something. And especially during event time, as I mentioned, we often with championships, we have morning and evening sessions. You’re already spending four, five hours in competition mode, and then all of the work that goes in editing and creating around that competition time, in the past would just take up an incredible amount of time, and you would end up working fifteen, eighteen, twenty hour days just because of so much of the manual work that would be on top of that.

And that’s massively reduced just the the workflow, for the team, and just means that actually everybody’s more well rested. Everybody is still creative on day eight or nine of a championship, and is is having fun with that creativity for us.

Sam Craft: Awesome. Cody, what about you, man?

Cody Silfies: Michelle took the words right out of my mouth. I mean, everything she said, I agree with a hundred percent.

I don’t think if we removed AI from everything, we would still function, but we wouldn’t be as efficient, and we wouldn’t be as creative. I I the amount of time that I, spend in my role doing administrative tasks or writing documents or answering emails, and I find that, a lot of the AI tools that are sort of included in the package of our organizational structure, our IT structure actually help with a lot of that work. And same with PhotoShelter, all of the time efficiencies add up. And they may be a little insignificant on a one off, but at the end of the week, if I’m spending less time answering emails or less time digging through a giant image asset library, I can spend more time working on the creative part of my job, which is what I’m passionate about doing and why I’m to work faster and be more creative with the tools that we have.

Sam Craft: I I think, the the word that keeps coming up here and the word that I tie AI to again y’all over the years has been has been time. How much do I get back? What is it really taking up?

And and look, robots are robots. Right? And because let’s shift back to a general AI discussion. First of all, I wish I would have started this with, the two big letters that are AI because I think there’s a lot of misconceptions on those words.

I I think in our world, creative world, it it means generative or it means creative or it means, hey. It’s just here to help you. And that’s the context that we’re talking over today. But it leads us into the next thing.

Right? Because, when is manual work that we all used to do in tags and all these fun things and athletes and, whatever else, when is that worth it and when is it better to automate a process? And I I think I’ve seen both of y’all talk about the day, especially in, in the socially world with the automated tags and all that kind of stuff. But then there’s the manual work of of, the people ID and making sure this and getting all that going.

What does that look like for both of y’all? Cody, let me let me start with you, man.

Cody Silfies: Yeah. I think Michelle touched on it a little bit. The human component is really integral to our our brand, as an institution. People need to trust higher education.

We’re never using AI to fabricate anything, or we kind of don’t mess with generative stuff at large scale and really relegate it to sort of just helping a workflow around, light photo editing. Let’s fill in some grass here. Let’s remove a trash can in the background, that kind of stuff, but never to change the subject of a photo or just full sale generate copy for anything we’re gonna go out into the world with. We know that our community doesn’t want that.

They’ve been very loud and clear that that sentiment around generative AI or just using AI to sort of make creative is not what they’re ingesting, we’re very careful about that. I think we focus on, the parts of that process where the human component is, integral to creating work.

As I said, for me, it’s the the the piece around I’ll give AI, fixing that grass or remove that trash can, or add that extra inch that I can put it in a publication that’s just a background. That stuff is is easy, but, but the actual creative concepting and sometimes even the pushback on the ask. We get lots of asks creatively for a thing, and we know as creative individuals, that’s not actually what you want. Let’s, digest through it. And if you just gave that full sale to a an AI processor tool, they might just give you what you want, which isn’t really gonna produce a lot of great work in the long run.

I I automate repetitive or, kind of research heavy tasks. I give AI our AI tools a lot of tasks around creating SOPs or guidance documents for internally. The other week, I I got an ask about what is our sort of photography guidance around taking photos of models in our plastination lab, which is where they replace biological material with, a chemical that they can create models out of real human tissue.

If you’ve ever seen, the body world’s exhibit sort of traveling the world, it’s this very

Sam Craft: I’m we we gotta talk more because

Cody Silfies: I’m I’m I’m It’s kinda creepy, but but but, I’m I don’t know anything about, human tissue and what other institutions do or, what the guidance is nationally from all the affiliations for, donating your body to science.

And our AI tools can do some research for us and cite sources and sort of come up with a lot of the the, examples of what our institutions can do and then easily turn that into a guidance document that I can go through and adapt to our institution in a process that takes, an hour. Whereas if I had to do all of that manually, it’d be, a couple days of work to just come up with sort of guidance around that. A lot of our research processes, things that we would hand off that are just, information gathering, easy to hand to AI. But if I’m really thinking about what we wanna do and and what we wanna create, that’s, the human component for me. That’s where we need to have a voice and an opinion and sort of, digest about what we’re trying to communicate about.

Michelle Sammet: Yeah. And I think from our side, there there’s there’s a couple of elements where even it’s the collaboration between AI and the human. One of the things that, for example, we’re often seeing is when we have events, not only do we serve athletes content from our own championships, but it’s also from our kind of one day meet. That’s Diamond Leagues, Continental Tours, World Indoor Tours, where we work with a lot of freelance photographers where they’re brief, they’re getting the content for us, and we’ll often have remote editors that will then edit their work and will still distribute it to the athletes.

But for those events, we’re not the ones generating the start list. We’re not the ones managing the results. We’re relying on third parties like the event organizers themself to provide us with the start list or the results. These come in different formats, in different sizes.

They contain different information. And when somebody has to do any manual tagging or has to do any work in that space because the athletes haven’t necessarily run through the process of going through a photo studio and being identified through our PeopleID. That still requires manual work. And having to manually type the athletes’ names often generates issues with typos and stuff like that, and having to copy it out of a PDF file that might be formatted differently doesn’t really work.

Whereas what we’ve seen a lot is that the taggers and the teams would then run the start list or run the result sheets through AI and say, please just generate me out of the following PDF document, name, surname, comma, country, comma, event, and then it’ll generate that list. And then it means that when they’re editing that work, they can then simply add it to, to that specific tool. It’s also using, again, something that would otherwise take their may just to find shift sift through those results, try and even find where they sit. And I just want to quickly pick up.

There was a question in the chat from Joshua, I think it was, about manually, how we name the the photos and and, the facial recognition part. For that, as I mentioned for us during the the process where we had the photo studio, the way we had set it up is that the photos needed to be edited anyway because we captured them on a green screen they could be used across those multiple assets. In that very initial capture process, the way that the photo was named was with the athlete name and was edited. And in PhotoShelter, The one thing that you need for the people ID is to have the person shown metadata tag contain the athlete name because that’s what’s being read by people ID.

As part of the whole flow or the workflow that we’ve identified, capture of the content, it being remotely edited and and tagged accordingly to make sure that in that metadata person shown, the athlete name was recognized. That was the the one manual thing that had to be done. And then all of that is run through an integration. I think we had it from a Dropbox, all the images automatically being pushed through into PhotoShelter, recognizing all the metadata, and then for us to be able to straight from the folder pull through into PeopleID, and then it automates that process.

I think there are other ways in doing it through CSV files and and that. We haven’t experimented with that. We’ve done a couple of manual ones that we’ve directly uploaded where we can also, for example, add additional tags to that particular athlete. So far, we’re only working with the athlete name.

But in the next step, we’re probably also going to start looking at adding additional metadata to the athletes. It then contains things like the country, like their discipline. That’s going to be the next step for us. We’re figuring out how to do that best.

That will probably be, again, a manual task. But for us, that tiny bit of manual work then saves so much time in the long run, that it’s absolutely worth it in that initial setup process. And for us, it’s, yeah, saved hours of work so far.

Sam Craft: I think that’s a nail on the head. It’s it’s it’s, the actual setup process, specifically what we’re talking about right now, Michelle, what you described, is not it’s not cumbersome. It’s not overtimely. If you just sit down, you do it, and you set it up, and you walk away, it’s done.

I mean, and adding is is little here and there, but you’re right. It will save you so much time on the long run of not having to identify these things manually and and automatically adding keywords and all that kind of stuff. Y’all, we’ve got about five minutes or so. I think that’s right.

We’ll kinda keep on track, but I think these are kinda we we got a couple questions left. They’re pretty easy, though. Looking ahead, are there any other things that that that you would like AI to help solve in y’all’s world?

Michelle Sammet: I mean, for us, the next one has been video. We’ve started experimenting with that during the past indoor season.

It’s not quite as reliable as images we’ve seen so far, but that also depends on the quality of the footage that we’re receiving. As I said, it’s not always just our produced events that we’re distributing and capturing content from. It’s also, where we have different host broadcasters where we’re then ingesting their feeds. We’re a little bit at the mercy of of what the quality looks like there. But, again, even with that, we’ve seen that PeopleID has helped us with that process. But, yeah, we’re still, working with the team on on figuring out, for example, how close does the the recognition I can’t remember what the exact term is, Sam. You might know this.

I might put you on the spot there, but you can set in in PeopleID just how sure the system or AI has to be in order for

Sam Craft: it to apply the auto tag?
Yeah. The accuracy of it. Yeah. No.
And the reasons, keep it pretty high because we’d rather be on

Michelle Sammet: Oh

Cody Silfies: Uh-oh.

Michelle Sammet: I’ll just talk over him.

Kassandra Costa: I’m just gonna jump in, and I’ll keep the conversation going as best as I can without knowing what the questions were that you guys put together. I think, yeah, I’m thinking about the people ID.

Sam Craft: I’m back. No. No. I’m back. I’m sorry. My VPN reset. That’s on that’s that’s my bad.

Kassandra Costa: Alright. You’re back.
I’m gonna. Head back out.

Sam Craft: Thanks, Kassandra. You’re the best. Sorry, y’all. That that’s totally my bad. I did not know it was gonna reset on me. Michelle, you finished your question. Cody, did you get the answer?

Michelle Sammet: Yes. I was going to say, it’s the accuracy, for example, on video that that we’re playing around. But, yes, video has been is for us is kind of the next step and just, again, knowing one of the things that I don’t think I have mentioned, but for us, that that is is crucial is the volume in terms of the turnover between events. We have, as I mentioned, two thousand athletes in our championships, but then two years later, when we have our next world championships, I would say only about fifty percent or so of those athletes will actually be in the next championships.

We also have under twenty championships. We have cross country championships. We have world relays. We have road running championships, and there isn’t that much crossover between our athletes.

Actually, the volume that we capture and the volume that we’re working with is huge. Anything that we can do to kind of simplify that process and help, with that is is massively making a difference for us because, again, otherwise, it’s very hard for the team to remember faces and to be able to do something off the cuff rather than, having to than if you can just set it up manually, basically.

Sam Craft: Sure. Cody, what do you got?

Cody Silfies: Yeah. I think, the tools are evolving so fast. I I try to think about what are the what are the pipelines in my world that are the longest and how my AI help with some of the to compress some of that time. We think about video a lot here, and it the appetite for video is just, increased, increased, increased, especially across all of our social media platforms. And that just happens to also be sort of the product that takes the longest time to produce. And how can AI help us with that?

And not gen again, not generative, but, what parts of that could be benefited from, from an AI tool? And that could just be, project management. Whose responsibility is what? What are the steps that we need coming up with an appropriate timeline for things, helping with scheduling of subjects.

All of those things are very helpful, with some of our AI tools. But I think about, approval workflows also. The amount of times that I have, three different approvals out to various clients at our institution, but they’re, living in my head waiting for the answer back. And if I could just put that aside and let a robot sort of handle that or check back in with them if they haven’t responded, then that would free up more of my time to focus on the creative work that I’m doing rather than having to chase people down.

I’m hopeful that some of these tools will help in that area.

Sam Craft: I I I can’t talk about it because that’s not my place, but there are exciting things coming this year in video AI space with PhotoShelter. I’m excited to be able to share that at some point this year. With that, y’all y’all kinda answered the last question. Really, to stay on track, I just wanna kinda wrap up and say thank y’all so much for your time. And I think we probably got time for maybe just one or two quick questions if Jeremy and the other power that be grant that. If there’s any if anybody’s got a quick question, let’s let’s try to get one or two in if anybody’s got one. Maybe?

Jeremy Berkowitz: I can jump in here. Yeah. I can jump in here. I just wanna say, I think thank you all for this awesome session, and thank you, Cassandra, for jumping in momentarily. Appreciate it.

If anyone does have questions, please let them or, jot them down in the chat. But for the sake of time, I always like to ask a question about, advice. If the do you have any advice for teams who are looking to kind of implement these tools or kind of take the approach of speeding up their processes? What’s, a quick piece of advice we could wrap up with? And I’ll start with Cody.

Cody Silfies: For me, I think it’s to AI they add AI. AI is in everything. They shove it into every product. It does various things depending on where you’re using it, and it can get a little confusing about what it actually will do when you buy a product. I think my advice is to, really identify the areas that you’re hoping that it’ll help you and and to just make sure that what you’re looking at does that because you could very much end up with a product that does a lot of stuff, but it doesn’t actually help you with, the thing that you actually need help with. I would the clarity is key, I think, sometimes in just being sure that, you’ve identified the areas in your workflow or your organization that could benefit and, and aligning that with a tool.

Jeremy Berkowitz: Sure.

Michelle Sammet: And I think, yeah, one of the main things is also where do you have expertise, and where does AI maybe not have that expertise, and what are the areas that really can be automated. I think one of the big things whenever I have also played around kind of in my spare time, is actually realizing that when it comes to some of the intricacies of our sport, sometimes it’s a rule change or sometimes it’s a more recent result or sometimes it’s a a story about an athlete that if you’re trying to look into something, actually, you’re not getting the right answer. And if you have the expertise in your team or if you know someone, rely on that.

But when it comes to something that can be automated, that doesn’t really require knowledge as such, I think that’s really where it’s worth investing. And as I said earlier, sometimes it might require a little investment. For example, us hiring someone to do the manual tagging that very first time around, but it then helps the entire team massively in the long run.

Jeremy Berkowitz: I love it.