.png)
Light Talk Podcast
Light Talk Podcast
Flynn Talbot: AI For Lighting Designers
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
You can feel it in every project: the spark happens early, then the real grind begins. I sit down with Flynn Talbot to talk about what happens when AI stops being a toy and starts taking meaningful chunks out of the lighting design workflow, especially the parts that slow studios down the most: documentation, schedules, spec sheets, and the endless back-and-forth that turns a concept into something contractors can actually build.
Flynn shares the “lightning bolt” moment that convinced him AI was ready, then we unpack his pivot toward helping lighting studios adopt AI in practical ways. We get specific about automating the luminaire schedule, pulling product data from manufacturer URLs and PDFs, and why a purpose-built tool can reduce errors, speed up revisions, and still leave the final judgment where it belongs: with an experienced lighting designer.
From there, we zoom out to the bigger questions shaping the future of architectural lighting design and lighting consulting. How do we prevent a race to the bottom on fees? Should we disclose AI use to clients, and what does privacy look like when NDAs collide with cloud tools? What happens to junior designers if AI handles the “boring” work, and how do studios build guardrails that protect craft, quality, and brand standards? If you care about AI workflow automation, value engineering defense, and keeping the soul of lighting design alive, this conversation is for you.
Subscribe, share this with a lighting designer who is feeling the time crunch, and leave a review with one workflow step you want AI to fix first.
More from the Light Talk Desk
Watch my vlogs, industry dialogues and project tours.
Listen to enriching stories, deep insights, and expert opinions in my podcast.
Learn directly from a master, absorbing the knowledge, techniques and the artistry of lighting design through my masterclass.
Sign Up to the AI Fundamentals course for Lighting Designers.
Download and delve into my works 'Light Talk, a Year in the Life of Light' and 'Light Talk 2, 16 Years in the Life of Light' - follow the journey of lighting design through his books.
Check out lighttalk.space
The Creative Ratio Problem
SPEAKER_00By implementing AI into your studio, you can, you know, I'm sure a lot of people can agree sometimes it feels like the creative part of the part that feeds your soul is 10 to 20 percent at the beginning, and then it's just all this work to get it delivered the way that you can see it in your mind. But if you're a lighting designer and a lot of your work is digital data processing, that's actually a blessing because if we can compress that down and change that ratio between creative and documentation, even by 10 or 20 percent, that it allows you to do what you said you enjoy doing, or it's a part of your process, Martin, of actually holding a light fitting in your hand and making sure you're happy with the build quality and the optics and all of these variables before you specify it on a project and put your name on it.
Flynn’s Pivot Into AI
SPEAKER_01Lynn, welcome. Welcome to the podcast. Hey Martin, thanks so much for having me. So you recently stepped into uh the AI space with your AI-empowered uh lighting designer, a course that probably will empower today's lighting design designers to embrace AI, which I think is an interesting career move. As previously you were working as a more immersive and experiential lighting artist. You have clocked a lot of hours, locked no, sorry, a lot of years in lighting design. So this is a really interesting shift. So I would like to start with uh understanding your journey and how you came to move from the lighting designer that I knew before to the AI-empowered lighting designer of today.
SPEAKER_00Yeah, that's true. It has been an interesting pivot. I can say that I'm still doing a little bit of lighting consulting, but it's it's dialed down from like 100% down to maybe 10%. I think in in the webinar that I held recently that you you joined me for, and thanks for your support on that, by the way. What what I shared there is, you know, when I understood the power of light early in my career, you know, I I wanted to play my part or or participate in creating a better lives world, like a better illuminated world. And I've liked to think I've I've done that in my way throughout my career. And it occurred to me more recently after you know watching the AI space and seeing the developments there and implementing, starting to implement some of that in my own studio and and understand like the power of it and and the scope for change there, that if I was to support, yeah, on the one hand, there was this very exciting new developments in the AI space. And then I was also, you know, looking looking at my career and and my mission of creating a better lit world. And the more conversations that I had, the more I saw that people were, you know, tinkering with Chat GPT a lot, but they hadn't really implemented it in their in their businesses. So I thought if I made a pivot into supporting lighting studios and and manufacturers to to do that more, then I could actually amplify the effect of of my mission of creating a better lit world. So I decided I just had a natural gap in in projects, finished a couple of them, and I had some space, and I just thought now's the time I gotta, you know, I've got the luxury of space and time to dive into this. And yeah, it was kind of probably a a bold and interesting pivot, but I also felt with AI that I had a feeling within myself that was the same as when I discovered lighting or the the the power and the magic of light in the first place. So yeah, I just decided to take that lead.
SPEAKER_01So one of the conversations that you had was with me. We met in Perth at my favorite City Beach Cafe sometime last year, I don't remember when. I I was at that time also already uh diving into AI, and uh I I could distinctively uh feel that you had an interest in AI, and uh I I could yeah, I could I could feel that interest. So I'm I'm really wondering at what point of time uh was there a moment that you thought, oh, this is it? It was there something specifically that triggered that, or is it just the fact that you had some time to yourself to dive into that? But was there something specific that said, yeah, I uh I need to get into this, this this is it, this is this is gonna be so important for for the future. It's gonna change our your life and and and our lives in general. Was there such moments?
SPEAKER_00I think probably our our conversation came up at a point where I was starting to understand the the potential impact and how our all of our landscapes are gonna change. But honestly, at that time I hadn't made any decision around anything. I did make a core decision not to follow the work of of anyone, and I wanted to like dive into it myself and uncover my own, yeah, not necessarily carve my own niche in the space, but more just to like enjoy my own journey of discovery with it and try and work out what would be meaningful to me and then have conversations with you know, yeah, outside of a a few conversations around this interview and and the webinar, you know, you and I hadn't been in touch so much and since that meeting at the Beach Cafe.
The Image Edit Lightning Bolt
SPEAKER_00Yeah, I think there was one moment that really hit me. Sometimes in life I feel like you you get hit by a kind of a a lightning bolt or like a lightning bolt. And that one for me was when I was working on a project and I I've I've had I was in this this moment where we've all experienced where you've you know you you created your concept, your presentation's coming together, it's like maybe 90-95% complete, and then you're just looking for that one mood image that's gonna just tell that story of one detail of your project. And I found this great image, and uh that's the one I showed on the webinar, and it it was a grid, a grid ceiling of timber with these short little cables with like bare lamps on there, and I just wanted to have a space, it was gonna be a dining space, so I just wanted to have these one fixture grid on this timber ceiling, and it was it was quite perfect, except for it was like a dark mahogany timber veneer on the ceiling, and it was supposed to be off-white. And then I just it was right at the moment where Google's latest image, you know, Gen AI, we'll just call it nano banana, because that's what everyone knows, came out, and then I dropped it in there, I said make the ceiling off white, and then it did it, and the image didn't look fake like an AI generated image, it just tweaked the image that was already there, and then I was kind of like, okay, that's a solid use case. It just saved me a bunch of time, and that's a pain point that we've all experienced many times over. So I can say that that was definitely a light bulb, light, lightning bulb kind of moment. And then I thought, wow, like I've been tinkering around with it. I mean, I definitely wouldn't class myself as an expert, but at that moment I realized okay, this is like kind of getting serious now. Probably time to like dive all in on things. I do have a track record of you know being excited by innovation and wanting to be at at the beginning of of the bell curve. I don't like kind of coming into things at the end so much. Yeah, that moment really stuck with me. That's why I put it into the webinar because it was true.
SPEAKER_01Yeah. Yeah, I remember I saw that uh in in the webinar. I thought it was very interesting. Now you I think you're you're you're a Perth native, if I'm not mistaken. So now you're you're in the UK. Is the location relevant to your AI journey?
SPEAKER_00The move back to the UK was kind of well, it mainly it was a personal one, you know, for my my family and my kids. We came back to Perth. It just honestly didn't feel like the right fit, so then we bounced back to the UK. But just having been here, now I can get the train into London and meet people, and you know, I'm much more connected with the UK audience if I look. People that message me and sort of conversations that we're having on LinkedIn and where people are following me, it's UK first, and then Europe and Australia and the states. So that's just organically where my network was because I've been here for quite a long time, and you know, majority of my exhibitions and things like that have been here. Although my architectural projects have probably like 90% or more have been in Perth. So yeah, it's an interesting mix and kind of fascinating that you find yourself there of all the places in the world where you can where you can land.
SPEAKER_01So definitely a strategic move. Now, I know you from Finn Telbot Studios. Now you have sort of incorporated Kelvin lighting, I think you you call it. Is Kelvin is your your vehicle to promote AI, if I'm not mistaken. Tell us a bit about that.
SPEAKER_00So yeah, after having lived this this career for my whole life, I'm well aware of you know the pain points that we face.
Fee Pressure And The Adoption Gap
SPEAKER_00I also find kind of quite globally, even in really uh you know buoyant markets like the you know, Middle East, for example, just conversations that I have with people in in my network is that you know clients are very well traveled, costs of construction always seem to be going up, and people, you know, clients are generally expecting more, but it's kind of people feel a bit of a squeeze on on fees, you know, across the board. So uh would you agree with that?
SPEAKER_01Yeah, yeah, yeah, yeah, yeah. Fees are always under pressure. So now the question of course is how AI can help us to increase fees. Yeah.
SPEAKER_00So I I thought also I'm I'm answering your question in a really long-winded way, by the way. But I also found in in my discussions around AI and other studios that there's there's a there's a huge gap between what's possible now and actual AI adoption of where people are at. So I'm sure we'll get to this more in the conversation, but Kelvin is is my vehicle for using AI in a way that's really relevant now to solve pain points that exist right now. I just thought if I'd built if I built an app, a product that was really high quality and kind of the best that I could make now and launch that soon and get people using it and improve it and improve it throughout this year, 2026, and just then that would be a a smart a smart move to take away pain points, help lighting studios, uh do it in a way that was really intuitive. So one thing that always stuck in my mind, yeah, a lot of a lot of the a lot of the space that I see that AI can help is you know less in the creative imagery of your project, a visualization piece, although that is very impactful and can save people a lot of time and help them, you know, creatively iterate faster. I've put my focus more on what I would call the boring side of lighting design, where you've got you've had your inspiration, you've like been creative, you know, you can see your project in your mind, but then you've got to create this endless string of documentation to actually deliver that project. And it feels like the lighting industry has become really technical. If we look at compared to other industries like landscape architecture or interior design, we've got more data that we need to crunch. And that side of things, I often felt kind of oh, just at sometimes I'm looking at a fixture of spec on a screen, and I was like, can't I just click my fingers and just have that in my spreadsheet or in my luminaire schedule or
Automating The Luminaire Schedule
SPEAKER_00whatever? This this thing around the luminaire schedule, and again, was in the webinar, but I looked back at a few projects. One of them had like over a hundred, I think it was 102 fixtures on a decent sized project, and it would just take like two and a half days to fill that out. And I just thought there's got to be a way to AI that right now in a like really clean interface. So that's the first product that I created because I thought if we can demonstrate something that works well, used to take two and a half days, and now it could take you maybe one to two hours with all the tinkering and changes and all of that, then that would be a pretty solid place to start. And then from there, as a foundation, we could build out of that. So that's what Kelvin is. It's a combination right now of consulting, but mainly it's this app. And then through all the conversations, yeah, then I just focused a lot on this luminaire schedule tool, and I thought if I can make that really good, then that's a solid foundation on which I can build this year. So, you know, I'm very lucky now. I've got I'm gonna be launching the beta soon. Not sure when this video is coming out, but it's gonna be around the time of light and building, hopefully a little bit before. So I'll have my first group of beta testers coming through, which, you know, I hope you can be one of them. And then I'm just gonna keep putting everyone's feedback into a big melting pot and just making it better and better. So Kelvin, this is a really long answer to your question, but Kelvin is essentially a lighting co-pilot through your project from really the the beginning to end. But I've focused more on the on the data crunching side initially. But I want to the more beta users that I get in there, I want people to vote on what you know which features we should build next. Because I want to kind of, you know, I can build things that I know are going to be useful, but I also am aware that I I don't I don't know I don't know everything, and I certainly don't know everything about every market. So I want to be really responsive with it in that way and build something that's actually useful and tailored like specifically to our
Pulling Specs From Links And PDFs
SPEAKER_00industry.
SPEAKER_01So that also brings a question about uh lighting manufacturers, because obviously projects can be in different countries, they might have different manufacturers, different suppliers. Do in order to do that that selection, that product selection, and create that that cut sheet, do you need to put in manufacturers or do you do that? How how does that work? You might have explained it in your webinar, I can't remember, but I would imagine that there might be some sort of uh pre-selection needed in terms of light fittings that are solid, that are good. I mean, I I have in my studio the habit of I want to have played with the light fitting before I specified it. I I find it very hard to specify purely from a catalog. I want to play with it, I want to see it. But that's maybe my my human experience part that that does that. But how how do you combine that AI selection process with the selection of light fixtures manufacturers? People may have different preferences, different suppliers or manufacturers they like. How does that work?
SPEAKER_00I caught most of what you said there, Martin, but unfortunately it was broken up a little bit. But what I got was, and I think this is something also that AI can help with a lot by taking a like compressing the data documentation side, allow us to spend more time with light fixtures in our hands and actually doing the tests that lighting designers love, like doing the shootouts, you know, three or five or ten fixtures or whatever, actually seeing the optics on a facade. I was speaking to one guy in in rehab who who did a facade linear fixtures. He did a shootout with 16 different uh manufacturers, which I thought was absolutely amazing data to have done. I was kind of like, whoa, I never went to that much effort. But wouldn't that be great if everyone had the time to do that? But in terms of specifying fixtures, yeah, the the tool that I created in Kelvin right now is when you know what the fixtures are. So it's kind of like, okay, we've done our we've done our tests, we've tried different samples, you know, we're clear on what we want to, you know, here's L L6 or AA10 or however you want to call it, you're clear on what those fixtures are, and now you just want to wave your magic wand and just have them all in a in a link or a spreadsheet or whatever to share with people. So that's that's the point of that where Kelvin helps right now. And if you need to change things, that's fine, it's all all doable as well. But and the way that you do that is you either paste in a link for manufacturer. If some of the bigger manufacturers that have bigger, more extensive websites where they have a unique URL for each product, you paste that URL in and basically will extract all of the the data, the the images and all of the specs of the product. Right now it's like at 80-90% because every lighting manufacturer's website is different, so it needs a little bit of custom code for every manufacturer. But from my side, it will only work for lighting designers if it works for every brand. So it was just a huge process that we have to go through. But effectively we'll allow you to specify anything. So you can either paste a URL, upload a PDF spec sheet, or you can just manually fill it out because it's essentially it's a spreadsheet in your browser that just happens to be tailored to lighting design. Did I answer your question?
SPEAKER_01Yeah, yeah, yeah, I understand now. Now you also obviously you have identified quite a lot of pain points in the lighting design workflow.
Mapping The 46-Step Workflow
SPEAKER_01I think you identified more than uh 50 or close to 50 different pain points in your you call it what 6D schedule with within the different uh processes that we go through in our workflow. So this is obviously one of them. Are there others one other ones that you are already focusing on to help with the AI tools that uh we have to our disposal?
SPEAKER_00Yeah, so I identified, and it's not everything, but it's kind of like the core the core procedures or the core design process from start to finish, had 46 key steps. And some of those were able to be broken down into sub-tasks, or we'll call them tasks. And then I was able to fit those under six pillars. So you but the first one is kind of pitching on projects and winning work, so that only applies to you know studio owners and and more management, you know, and then there's concepts, there's kind of uh you know design development after you you're iterating, then you're moving into design, and then you're all getting ready for tender and documentation, and then you're you're delivering the project, which are bundled a whole bunch of stuff together, which is you know, dealing with all the the value engineering and the the requests from information from site and then you're creating handover documents at the end. So when I look at all of these tasks, it's uh that's actually quite valuable just to look at and remember all of the stuff that we all have to do to deliver great projects. When I started compiling, I feel like a lot of people that I talk to, they know they instinctually know all of these processes, but not everyone has actually created, you know, like SOPs for their business, their staff haven't really sat back and looked at that whole process and created kind of a formula to it. They just a lot of people are just diving into that creative process with excitement and then navigating through and then ending up with a finished project at the end. But I feel like there's some value to just take the eagle-eye view of the whole lighting designers process and then understand that probably 80 to 90 percent of it can be touched or optimized or something by AI, either in a huge way, you know, taking something that takes a few days and getting it to take an hour or less. I have been to you know, make dot com or N-A-N, these like tools that you use to order create AI automations without code. I've been to some of their build-a-thon events in London and they had people up on stage. Yeah, one guy's company worked for a manufacturing company, not in lighting, but I think they were making like cherry pickers and farm machinery, and he said he took a process that would have been a few people, you know, over a period of time, but it was roughly like 40 hours of of stuff of work and iteration and editing, and he reckons it he got it down to 45 seconds, and that was just how long it took for the automation to actually run. And then his company changed his role and made him like digital automation expert. But he said at the beginning he was just kind of lazy and trying to do his work faster. I think that's a pretty amazing example. If we can think across all different industries where that kind of stuff's gonna happen, and we think about our actual lighting design process and what that looks like, you know, in terms of ROI for a business, you know, it's this game changing stuff to be had. But as I said earlier, there's there's a huge gap in terms of what's possible and what's actually being implemented in a in a reliable and tested way for lighting design studios right now.
Turning AI Into Studio ROI
SPEAKER_00But what I'm kind of the message I'm trying to share is that by implementing AI into your studio, you can, you know, I'm sure a lot of people can agree sometimes it feels like the creative part or the part that feeds your soul is is 10 to 20% at the beginning and then it's just all this work to, you know, get it delivered the way that you can see it in your mind. And it's like that for a lot of, you know, kind of everything creative really, but in a way we're lucky and this is actually just coming to me right now, maybe if you're a sculptor, AI can maybe have less impact on your work. But if you're a lighting designer and a lot of your work is digital data processing, that's actually a blessing because if we can compress that down and change that ratio between creative and documentation even by 10 or 20%, then it allows you to do what you said you enjoy doing or as a part of your process, Martin, of actually holding a light fitting in your hand and making sure you're happy with it, the build quality and the optics and all of these variables before you specify it on a project and put your name on it. So yeah, I think it's a I think it's a really exciting time. I think AI is being super underutilized and I think basically every studio should have at least one person who's going all in or working with a AI consultant to just really understand it. Or people should buy your course or buy my webinar or just they should be soaking it up as fast as possible is what I can
AI Co-Pilot With Human Checks
SPEAKER_00see.
SPEAKER_01So let's dive into AI and creativity because there might be some misconceptions from designers of what AI really can do and where do you begin using AI in in the creative process and and how do you how do you master that how important will it be to master that AI process and have the oversight of what AI is going to produce for you because obviously we can produce 100x times the the number of options and and solutions but somewhere around there we need to find the ones that are actually realizable buildable and eliminate any hallucinations that the AI tool may produce and obviously if you prompt better you might get better better outcomes but there might be some some yeah some some worry maybe some some scariness for people who dive into the AI world on how to use AI in the creative process and and I would love to hear your point of view and in how AI and creativity can can marry together the the first step is to understand that it's not going to do all your work for you and you still need to have a masterful eye to check what it's doing.
SPEAKER_00But one thing you can do is build a couple of AIs one to do the work and one to check the work that's been done. So when you're seeing it you're lowering the risk of hallucinations coming through and then you are you can also set things up in a way where it's highlighted the changes have been highlighted or the key points. So one conversation that I had is around in the UK like tech subs or value engineering these kind of substitutions that come in most pretty pretty much every project goes over budget and takes longer than it's supposed to take is pretty pretty standard part of that process you know as we all know lighting's getting installed at the end of a project so it's often unfortunately budgets are cut in lighting and then builders and contractors are you know looking to create valued engineering documents and they can always supply stuff cheaper and they'll get you know they'll get the ear of the client and try and get this stuff through. But it's never apples for apples and it's easy to create well easy but it's possible to create a really good workflow just to use this as an example where you have your existing luminary schedule and a value engineered schedule that's been put forward to create an automated workflow where it's checking and highlighting and either qualifying or disqualifying fixtures that either meet or don't meet all of the the specs that you know are important for all of the reasons that you know either safety or aesthetics or you know meeting codes or regulations or whatever it may be. You've already clients already hired you for your expertise to put all this effort into selecting these live fixtures and there's no one better to find similar ish fixtures that still meet the design intent than the lighting designer. So part of our work now is to defend all of this cheaper copy stuff well not copy but cheaper potentially inferior products that that are not going to allow the project to be delivered with a designer's vision upheld. So yeah creating a workflow around that getting an AI to compare whether everything is you know apples for apples and but then highlighting all of the the areas where it's not but then of course you know someone still needs to check that so I think one of the biggest shifts is we need to understand that our roles are are changing from doing all the manual documentation ourselves to doing less of it and checking the work of you know an AI just say the word colleague because this is yeah where things are at really it's not really a tool it's becoming more of a kind of an AI colleague for your lighting design process. And sure setting up guardrails around prompting and all this stuff will help you to get better results but we've got to understand that it's not you can't replace human thinking they can't know all of the reasons why you chose a specific fixture or why you want to mount it in a certain position. There's always so many nuances on buildings and looking at the image behind you there, Martin like some AI can go sure let's put some uplights on that facade but they're not going to know exactly you know the placement of the fixtures there because you've been standing on site and you've been on a ladder or scaffold and you've done the test or whatever. There's many ways where it's not going to be able to have the full context or be fully informed. But yeah I think the biggest thing and and you know conversations like this are are really great to help people to understand that our roles are evolving and shifting a bit. If we're not going to do all of the documentation ourselves and we might say oh that's great that's reduced that time by 20% or 50% here and there we do need to be aware that 10 or 20% needs needs to now be spent on on checking the work that's actually been there. But you can you can set up a series of hurdles so those those checks have
Fighting Value Engineering With AI
SPEAKER_00already been kind of validated with by a couple of steps before you see it. So yeah that's a possibility too but I'm certainly not saying that AI is here to you know replace your replace you or your workflow. And that's actually one of the questions I had right now like what's your take on the industry around AI right now do you talk to many people that that feel that you know it's gonna kind of take over their take over their job?
SPEAKER_01You're asking me a question I was supposed to ask the questions but no I mean you you might have followed a little bit what I'm what I'm saying I'm saying we need a new mindset. I'm saying that AI is here to stay there's no way that we can just wait and see AI is already here we need to embrace it we need to work with it but it like you say it's our co-pilot it's our assistant but what I what I miss in everything is the human in the loop the experience because you and I have have a lot of experience together you more than 20 I'm more than 40 so I've been on site I I've been there I've done that so I just need to see a site or see a lighting installation and I can immediately translate what works and what not works and and I'm sure you have the same thing. Now if you I don't know if I understood you properly but I thought you said that you were also gonna create an AI tool to check the AI. I'm not sure if I got that correct but anyhow the point being that we need we need the human with the experience and in the new generation and that's where I feel there might be quite a big risk that the new generation that doesn't have the experience that hasn't gone through the actual project loop of a couple of years from start till finish understand what happens at the end when you implement lighting and lighting design understand what kind of problems contractors come across on site what problems there might be with supply issues. If you don't know that you can't anticipate that in the earlier stages of design and creation so that's where my worry lies at the moment with AI the human in the loop and the experience if you have the experience you can create the trust and you were just mentioning the client comes to you because they trust you yeah that's because you have that 20 years of experience of doing projects but the new generation who might just start up might not have that experience and if they rely on AI to create lighting solutions how secure is that where's the oversight from an experienced person that has been there so that's a little bit where I'm at at the moment I'm I'm talking about three points that yes AI can bring you the speed to to create many opportunities and give you more time back but your experience is the the the element that will bring the trust for the client and then you as a lighting designer has to have to provide the integrity to make sure that what you deliver is actually compliant is buildable etc so that's what I'm trying to push and trying to make sure and that's part of the the the the guardrails that you were mentioning the the infrastructure that we need to provide and create that provides us a safe environment to develop AI tools but I think we're all all sort of seeing the same thing AI is here we need to embrace it we need to master it it's not going to replace it I'm saying I said in my last presentation that AI is not going to replace you but it will be you will be potentially replaced by somebody mastering AI. So if you're not into it and if you're not diving into things like your webinar or my course or other there's quite another quite a number of other people also offering AI courses I think that's crucial if you want to survive.
SPEAKER_00Yeah 100% I I did a quick post some some weeks ago that was like like smaller modern lighting studios and I'm in thinking
Training Juniors Without Losing Craft
SPEAKER_00about this in a more deeper context based off what you just said smaller lighting design studios that have experienced founders and kind of know what they're doing can now leverage AI and compete with bigger studios and and the other part of the post was bigger studios that that are complacent and not willing to change may start to lose some market share to more AI enabled younger more nimble studios if they want to just stay in a really manual you know stay in their lane with doing everything manually basically but your point around like what's it like to be a junior lighting designer now joining a studio right now or joining or you're still finishing your studies and you're going to join a lighting studio next year really fascinating to think about. So yeah I mean because we had to do everything manually that forces you to learn things and it forces like every line on that luminaire schedule or whatever whatever part of the process that you're working on is all of that it's like you you've designed the concept you've held the light fixtures in your hands you know you've you've documented it all you've drawn the CAD plans you've drawn the sections and the details and all the things all of that is sinking into you and if now if you're not doing parts of that I guess the responsibility then lands on the more senior staff and the founders to make sure that the next generation is getting the right kind of experience and they're actually learning and going to be able to carry the torch because people are always gonna build buildings you know real estate property development we're all passionate about the built environment no matter what kind of buildings they are or where they are so that side of things is is always going to continue but there needs to be yeah definitely I mean this is another level where lighting studios need to evolve and part of what they do now is to make sure that they're not just as many people like yourself and and myself I've had interns and staff with me and and freelance designers or all kinds of different people in different parts of the world there's you know you want to pass on the knowledge but now the knowledge has to be passed on in a in a different way and maybe the question is also for the juniors and they would have already been doing this but it's like here I've come here's three possible options for this space and here's the reasons why and whether they used AI for that process or not they're still going to need to be able to know those reasons why and and improve them and I think that's potentially a critical piece and then have a dialogue with more senior staff and to just make sure that is actually understood.
SPEAKER_01Yeah but this is a really interesting thing I would love to understand what you would say give as an advice to new the new generation of lighting designer but at the same time there's also the older generation like like myself and we have a lot of legacy thinking the way we used to do things so it is a small anecdote what what how I sort of learned my my skills when I started I started with Signify or Philips it was called at the time and in in in the design studio that I was working in terms of learning I went to the elderly that is called the the the seasoned lighting designers and I asked the first one I I was given my first job and asked the first one how would you tackle this problem? How would you light this in this case was a facade lighting and he gave me an approach and a design concept that I said well I would probably do it this way and then I went to another guy and asked the same question and he came up with a different solution and I thought oh that's interesting. I didn't realize that there was more than one solution facade lighting this is the way to do it. And then actually went to a third person and he came with another solution again. So it taught me that there's first of all many solutions to challenge but also that I had to learn and make something my own I needed to pick what I liked and then sort of create something that would be I guess my signature lighting design move on that facade. So now the question is with AI can we trust yes you're right we have to ask the seniors in the company to give advice and that's the the experience based thinking but if you also dive into AI and ask AI for opinions how much of that is actually truthful and based on proper design thinking and how much of it potentially is made up and we might create lighting designers that use potentially unverified advice.
SPEAKER_00What's your thinking about that quite a fascinating space like junior designers are they're gonna need to come up with they're gonna need to understand all the different layers and come up with concepts on their own and if they don't learn to do that then the foundation of understanding light in a space and all the all the possibilities we've never had more light fixtures available to us ways of lighting things than we do right now. It's almost overwhelming how many brands and new fixtures come out every year. I mean I think it is actually overwhelming. So we don't really need more light fixtures but you have to be fully aware of all of the possibilities and I've said this in a few meetings in my life whenever I look at a space same as you kind of feel like a bit like Neo in the matrix you're kind of flying through testing like in your mind within milliseconds you're like yes that could work no no no for these reasons yes no filtering through all these different creative iterations at lightning speed and that's a really important skill to have because sometimes you you're on site in front of a client and a builder team and they expect something hasn't worked for any number of reasons and someone expects some pretty good solutions on the fly. So you're a human uh you still need to be able to come up with that stuff. So there is definitely a problem if juniors are only using AI to kind of do some Tinder style swipe through lighting concepts they're not actually learning the foundations of of lighting in a space and they might be able to tell you why it's going to work and why it's not and here it meets you know meets the codes in these these areas. Yeah this is why I think these discussions are really good. So I think they're each studio is going to tackle this differently but it yeah it raises a really really powerful question and maybe maybe the conversation is not just around guardrails in terms of prompting and AI use. Maybe there's all senior staff also need to put guardrails in place around which part of the design process AI is able to be used within their studio to make sure that their juniors are building foundation in the right way that's going to serve them in the long term. And maybe like I can remember when I was younger for example reading through Luminaire schedule that a colleague had created and noticing all the human error just typos not like a lot but just a little bit here and there and then finding out later that's because they used that same light fixture on a previous project and they thought oh instead of manually typing it out I could just copy and paste that in and you know and that kind of makes sense. But then because of all the nuances and variables on a light fitting there you know inevitably there was one detail that they didn't notice at a quick glance you know the optic was slightly wider or something but on the surface it all looked the same. I always typed every luminaire schedule from scratch and because I never wanted to have those human errors because they can turn into really big problems down the line when you order 20 light fixtures from Italy and they get air freighted to Perth which done a number of times you want to make sure they're the right ones. So yeah I think you're just circling back to your point around yeah I think there's work to do for studios to define yes AI can touch many tasks within our design process but as a studio oh I've just got some magic light beam on my face as a studio we're gonna make a decision that we are going to implement AI here and here and here and we're gonna enjoy kind of like the time savings there but maybe we're not gonna implement it in these other areas because we want our juniors to build that foundational knowledge and you know of course AI everyone in the lighting industry generally pretty smart and innovative people and AI is a very hugely flexible so that's gonna look different for each studio and based off the conversations that I've been having a lot of people have their own processes and yes there's elderly in a number of the studios that have always been doing things a certain way and I think there's a there's a risk of the risk goes both ways of like not helping your juniors have the right foundation but then also not implementing AI and staying stuck in your lane too much. I think everyone needs to navigate that that middle ground and what does that look like for them?
SPEAKER_01I mean that might be easier for smaller studios than for bigger studios because where where is the oversight and in the end you know the the brand of the the studio stands for something you you you know you you produce things under Flint Talbot and people expect a certain quality and a certain creativity from you. I'm I'm today I'm part of NUTI huge company with with more than 150 staff keeping the oversight on the AI process I mean everybody agrees we need to integrate AI but then it also brings with it quite a responsibility Responsibility of maintaining the quality that people expect. And as you said, very rightfully, there is a chance that there might be something creeping in, an error somewhere, and that's not spotted and it's sort of amplified along the way. You know, there's there is a responsibility as a studio to make sure that whatever comes out of the studio complies with the brand image, complies with the standards and the qualities that people expect from that studio. So there's definitely a lot of uh work to be done, and yes, we're gonna gain a lot of speed in in creation and in and in you know fast tracking the the pain points, but at the same time, we probably need to put in more hours in terms of checking what comes out of all these iterations and all the documentation that we produce. So on that point, what how do you think the future, I mean, looking five years ahead from now, for instance, how how would the the future studio which is driven by AI look like? You have any vision of that in your mind, how that would
The AI Studio In Five Years
SPEAKER_01look like?
SPEAKER_00I think first of all, there's like transparency with clients. I think by that time, I think there's hesitation from a lot of studios to to say that they are actually using AI now. Some I know some are doing it, but not very many. So I think by that time, AI integration is is gonna be so widespread across basically many industries that I think it you know the AI influence or optimization is gonna be more known to the client. I think one thing we've got to be really careful of is that this kind of ship is is steered in the right way, and that's not like a race to the bottom on fees, and clients are kind of like, oh, you're an AI studio, so should be charging me only 40% of what you used to because AI is doing all your work now. Like that would be obviously a really bad place for everyone to end up. So I think at the end of the day, as you said, your clients are paying you for expertise and creativity and problem solving. So we need to we need to set up now the right foundations for for that to continue forward and for yeah, AI to help where it's useful, but not to detract from human expertise and creative vision. Yeah, it's what's the landscape going to look like? Quite fascinating. So both of us were in the lighting industry before LED was on the scene. Um we've seen huge changes in that space. And I think now, you know, we're looking at something equally as massive that's happening kind of 10 times faster. And a lot of people, even in the AI space, you know, I've I've been learning a lot and researching a lot and sitting in on other webinars with you know Sam Altman, co-founder of OpenAI and all these people, and they have all these research teams and they kind of think they know where things are heading, but even they don't totally know, and they you know openly say that sometimes. So yeah, on the one hand, they said there was more people who were willing to pay for Chat GPT than what they thought, but actual adoption from from humans and from businesses has been less than what they predicted. So yeah, I think there's some people are gonna be really leaned in and they're gonna find the right foundation. I guess they're probably gonna stuff up on a couple of projects. There's probably gonna be some errors if they're relying on AI too much, and then they'll hopefully learn from that and and dial it back in and evolve what their roles actually are. There'll be people that have great wins in the visualization space, and they'll be allow them to be kind of creative, but creative 2.0, where you're able to iterate and try things faster, people who will create more systematized workflows and they'll be able to maintain the quality of their presentations and their output. I think lighting presentations, by the way, people are gonna stop sending PDFs to clients. That's gonna feel really shit and boring, probably. In five years, people uh clients are really gonna be expecting a lot more than that. So that's a pretty exciting place. You know, like there'll it'll just be an immersive video, and Martin will be walking around within the model and explaining, you know, this and that. Much more interactive ways of presenting your work, which is very exciting. But yeah, a lot of it's unknown, to be fair. I think it's a really exciting time right now. I think there's definitely a lot of questions that need answers around you know client privacy and data sharing and all of this side of things, which is a huge, huge point. But also I think even right now people are signing NDAs, but they're still, you know, using Claude or Gemini or or whatever, yeah, they're already integrating AI in there, whether they've fully acknowledged it or not. I think, yeah, I like to think with a lot of things as a sliding scale between fully human and fully AI, and every every studio has got to kind of slide that scale and and find the right balance. But I do like to think that if the right conversations are being had and the right industry leaders like yourself are in place and helping kind of integrate AI into workflows of studios that will find organically the right balance and one that is able to kind of amplify our creativity, allow us to ultimately design better and present our projects better. And you know, and if one time AI comes up with an idea that even you or I together hadn't thought of, but a client's paying us to do that, and we had created that AI based off our own brand and all of our own documentation and past presentations and all of that, then ultimately, you know, we cu curated that experience based off our own expertise in the field. And and you know, to be honest, I think that's that's fair as well. As long as the main yeah, you said human in the loop earlier, as long as there's still many humans in in the loop, then buildings will still get built and real art fittings will be screwed to walls and mounted correctly, and everything will, you know, will continue. It'll just continue in a very different way.
Disclosure, Privacy, And Client Trust
SPEAKER_01You mentioned an an interesting point which I've had in several discussions, is whether to disclose the use of AI or not. There's some people that say, Oh, you should be open and transparent and you should tell your client that you're using AI. Others are saying, Listen, we've been using Google search and and Pinterest and whatever, all these other apps in the past. We never disclose that. So there there is something for both sides to say. We also have the education where, you know, education where students now are using AI for their task, and and and uh when they have to submissions, they have they use AI, whether to disclose that or not. So there's a big discussion going on. But I think in the future it will be taken as a given that you you know you're using AI. And if you don't, that's your problem. But people will sort of expect it probably as it becomes a bit more mainstream. We're probably at the very beginning right now, even though there are people that have been using AI for a couple of years already. And and but it's a discussion that originally I thought, oh, if you use AI and you don't disclose it, you're cheating somehow. Now I'm thinking, nah, it's it's you know, everybody is using AI, so why should you have to specifically mention that? People will notice it anyhow. So yeah, it's it's a bit of a you know, we we're probably at at the edge where it will become so common that nobody will really ask about it and it won't be really an issue. Are you are you actually disclosing the use of AI in your consultancy? Or or is that just something that you just integrate as a normal step?
SPEAKER_00I mean, I haven't done yes, but as I I also haven't used AI that much on projects because I just had a natural kind of finishing of a couple of projects at the same time and a space in the last like kind of three, four months. I was using it just in a kind of basic behind the scenes way before that I think a lot of people are already using it there. But yeah, I guess on the one hand, this raises a kind of ethics question. On the other hand, I think, yeah, as you said, everyone's already using it to some degree. So hmm, it's an interesting one. Yeah, I don't know that there's a clear answer to that. I think it's kind of a personal choice. Maybe some people will win projects because they market themselves as a like AI optimized studio, and they're able to maybe some clients prefer that, and they're maybe able to get more technical answers to questions faster because of that, and maybe that helps. I think there's definitely going to be clients that are pretty against it as well, or expect, as I said earlier, like you do your fees to be less if they think AI is doing half half your work or something. But I would say if you were going to disclose it, I guess yeah, going back to the ethics thing, if you are really AI empowered and you've got many systems in place and you know that it's optimizing 30% of your business, I I think it's a personal choice. If you were gonna disclose that you're using AI, I would have a kind of company-wide document or a link or something that explains the use of AI within your company. That would probably be the best way to do it, and then you know, get some get some help to put that together. Because I think, like right now, you and I are having this conversation, and this will go out to some of the lighting designers in the community, and we're educating lighting designers, but someone also needs to start educating clients and property developers. And who's who's doing that? They're they're already aware that basically everyone else in the team is using AI to some degree already, even if they don't want to acknowledge it. So maybe there's a point soon where each studio needs a kind of AI legal document that comes along with their fee proposals or somewhere sits next to the terms and conditions on their web device.
SPEAKER_01Yeah, I mean that that will be the point if you do a fee proposal or you produce, you have your stages of work, you probably need to sort of shift from time-based billing to value-based building somehow, to explain that you're actually spending still the same amount of time, but you're converting that in a higher quality and a higher added value of your designs that you're you're providing to your clients. So there's a lot of thinking now to be done on how to approach that. I don't think we need to to hide the fact that we use AI, but we need to now start educating our clients what the additional value is, and that's what I'm trying to promote right now, trying to show where the added value is in using AI and what the benefits for a client is and why they still have to pay what they have been paying us, or maybe even more. So I think that's the discovery journey in the months ahead. Or how do we make this count for the company and how does that not diminish our services but add value to our services? I think that's very much the the way I see things going and developing.
SPEAKER_00Yeah, absolutely. I I really like that. I'm just imagining, you know, a page on every studio's website, you know, how we use AI, and there's a video there, and it just explains the design process of what it used to look like and what it looks like now using AI and the amount of definitely value-based fees. Client doesn't really care to pay people to sit there to plug in like low-value data entry. They they want to buy your experience and creative vision on a project to make their make their investment look and feel and function better. That's ultimately what they want. And of course, they want that for as cheap as they can get. But if they're building a proper building, they know that they have to invest in good lighting design to a certain level. And then if everyone's still being invited to, you know, pitch, you know, multiple studios will still have to pitch on the same project. And then if a client can easily understand here's here's their pitch, here's how they're using AI, here's how it's help helping optimize some of this low-value data entry stuff, and then here's you know, three key examples of how they've used that time to do better lighting design. I think that makes a lot of sense and it's quite easy to you know give very clear examples of that. And that's probably the way forward. When does something like that need to happen? Studio could I think everyone should start thinking about that right now, probably by the end of this year, just based off what I've seen about the you know, Chad GPT being maybe 10 times more powerful or so by the end of the year. So we've got on the one hand, AI's accelerating like this, but then we've got all these people like us and clients that are in a lot of ways kind of resistant to change and and don't want to have to change their businesses kind of too dramatically. So there's a there's a the curve for that is a lot flatter than the the curve for AI speed and intelligence and power and everything right now. So yeah, at some point by the end of this year, certainly next year, I would think that would be pretty common on most studios websites or part of their presentation.
SPEAKER_01I would like to to wrap up the discussion
Guardrails To Protect Quality And Fees
SPEAKER_01a little bit. Is there anything that we did not discuss that you would like to discuss, or maybe in another phrasing, if you could have things your way in the implementation of AI, is there anything that we'd like to see happening or implement it?
SPEAKER_00I think based off a couple of the key points that we covered here around the future of lighting design, uh what does it look like to be a junior in the space? What does it look like to, you know, not have a race to the bottom on fees globally? I think it'd be important to you know, a few things I would like to see is in industry leaders kind of standing up and representing lighting designers by creating some, you know, the kind I guess we'll call it the guardrails for AI use in lighting design and just educating on that so all all clients understand how it's like AI is impacting every industry in the world. So do we just do we just open the floodgates and let it come in and impact lighting design in in any way, or should we actually put guardrails on that too, for the entire industry to kind of yeah, safeguard the space and make sure that light is you know is still AI is being integrated in a way that helps lighting design and helps kind of maintain the future of the industry. I think that's that's kind of it seems like a lot of documents and kind of boring, but it seems also kind of critically important. And yeah, the educating of clients in that space, more more discussions, more panels with spaces outside of lighting design, you know, architects, property developers involved, and just kind of like uh yeah, spreading this thinking outside of our small bubble into these bigger spheres. I think should be pretty important. And then yeah, somehow getting some kind of consensus for from you know, uh people like you and other experienced people might might call you an elder. But I just think getting wise people together to create those guardrails for the industry is probably gonna be a really smart thing to do. And then within that studios can decide, you know, if they if they agree or what their version of that looks like, but ultimately would be to kind of safeguard the fees and you know the profitability of of businesses into the future, which you know there's many people that that have jobs in lighting design, mainly you know, lighting designers are very important to you know stay afloat, keep keep the businesses going and you know, keep the industry afloat for the you know the lit quality of the world. So yeah, I think instead of putting a head in the sand on a lot of things, we could be more proactive and more leaned in and steer steer the ship in the right direction.
SPEAKER_01Yeah, I think we have a very similar drive here. People say, Well, why don't you retire and just go fishing? But at this point in time, you know, I I wanna the the reason I'm I'm diving into this is because I I hope that we can keep the soul of lighting design still alive, the whole creative process, the human in the loop. There's a fear, certainly if if you know you and I have been in this business for many, many years. You you feel that things may get lost in in this whole AI development. And I think that's why I think it's really important to to dive into this and and make sure that we create those those guardrails, that we create the infrastructure, so that the real true soul of lighting design still remains and is not taken over by some AI avatar who uh yeah also might might claim to have that experience and and and design foresight. So thank you very much for this discussion, Talbot. Flynn. It's been a really uh enjoyable discussion, and I think we could go on for hours, but I think we've captured very much what you wanted to say. Where can people reach
Where To Find Kelvin Lighting
SPEAKER_01you? You you had a webinar that people can maybe still review or purchase. I can I can put it in the show notes so that people can still reach out to you if they want to know more. Otherwise, I look forward to seeing you and present in in uh Light and Build in Frankfurt.
SPEAKER_00Yeah, thank you so much, Martin. It's been a great discussion. Do you the only thing I wanted to add to it before we wrap is like also embrace the parts of AI that are exciting as well and the things that will make us like design better. I think it's really important. Like set up the guide rails for the industry as as a whole, but then also embrace the new ways that we can be lighting designers. I think all of that is really important. But uh yeah, just that one final thought. Yeah, thanks so much. Always a pleasure. Follow me on LinkedIn. That's the best place to keep up to date with things. Check out kelvin.lighting. So that's my new new company, a new app building there. I look forward to inviting people into some of the beta testing and then the first product release, and very excited to deliver the the first first piece of the puzzle out of these 46 tasks of the lighting designer, as as I kind of defined them, just starting on that one, getting that one working well. By the end of this year, there'll there'll be many more, and it's it'll be valuable for studios, whether you're a one person or you got 50 people or 150 or however many Nelty has right now. Yeah, whether you're a super huge studio or you're a one person that will be valuable for you. So yeah, check me out. LinkedIn on Kelvin.lighting, and yeah, real pleasure. I look forward to seeing you in Frankfurt. So much. Thank you, Martin.