Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
The GenAI revolution gets visual, with Sam Pickover
24 Plays17 days ago
This episode was recorded on the morning after the 2024 US election. Historians may just have a few things to say about it in years to come. One reflection will no doubt be on the impact of images and videos on the campaign. Whilst we’ve had several social media elections now, this was perhaps the first one where social media was combined with the power of AI to deliver visual content. The results were in turn unnerving, powerful, positive, and terrifying.
The impact of AI is, and will be felt no more viscerally than in the domain of visual media. After all, videos and photos impact us on the most human of levels. And organizations, like politicians, are still trying to come to terms with it.
To discuss the fusion of AI and visual experiences and what it will mean for businesses, Will is joined by the VP of Strategy at Cloudinary, Sam Pickover.
Get in touch with Cortico-X here.
Follow Cortico-X on LinkedIn here.
Recommended
Transcript
Introduction and Podcast Context
00:00:00
Speaker
Cortico X is an experience-led transformation business that partners with clients and technology companies to drive digital acceleration. We are experienced activists passionate about elevating everyday human experiences through the belief that what's best for people is what's best for an organization. Reach out to us for a chat. A link is in the show notes
00:00:34
Speaker
is in the show notes. Hello and welcome to Speaking from Experience from Cortico X, where we speak to the people with experience, of experience. I'm Will Kingston. We are recording this podcast on the morning after the 2024 US election. Historians may just have a few things to say about it in years to come.
AI's Role in Visual Content During Elections
00:00:55
Speaker
One reflection will no doubt be on the impact of images and videos on the campaign. Whilst we've had several social media elections now, this was perhaps the first one where social media was combined with the power of AI to deliver visual content. The results were in turn unnerving, powerful,
00:01:14
Speaker
also do and terrifying the impact of ai is and will be felt no more viscerally than in the domain of visual media after all videos and photos impact us on the most human of levels and organisations like politicians ah still trying to come to terms with it to discuss the fusion of ai and visual experiences and what it will mean for business i am delighted to be joined by the vp of strategy at cloudenry
00:01:45
Speaker
andpico sam welcome to Speaking from Experience. Thanks very much for having me, Will. Looking looking forward to our chat.
Cloudinary's Role in Visual Media Management
00:01:52
Speaker
As am I. I won't drag you into the swamp of political conversation. So instead, let's let's focus on visual experience and we'll start with Cloudinary. Tell me about it. Yeah, sure. Cloudinary, simply put, is a provider of visual media software for brands. Translated, it makes it easier, simpler for companies that want to create, want to manage, and then ultimately deliver visual content, images, videos, 3D models, whatever it is to to their customers. Yeah, as you say, this is something that that touches a lot of people. we We're seeing a huge amount of it, obviously, in the consumer domain. Now you touched on the the way it's affecting political discourse at the moment, but yeah really and we're only just at the and of precipice of the impact of of some of this stuff in the in the enterprise space. So yeah, I think anything anything you see when you're shopping online, you're browsing some media news sites or logging onto social apps, yeah you're often seeing visual content and often that's powered by Cloudinary or a company vico like Okay. I popped a photo on Twitter this morning. That wasn't too difficult. What's the secret sauce? What does Cloudinary do that makes it makes it special? Yeah. So for Cloudinary, as I said, maybe one kind of mental model for thinking about what a company like Cloudinary does is thinking through the stages of a the kind of creation and delivery of a visual asset. So there's the kind of creation and manipulation of it. There's then managing it internally. And then finally, there's actually delivering it you know in a way that is and some useful, optimal, visually appealing you know to the kind of end consumer on a phone, whatever device it is you're looking at.
The Evolution of AI in Visual Media
00:03:39
Speaker
And there are sort of a couple of things that Cloudinary does that's sort of trying to make that that whole whole process process easier easier firstly we play in the space of doing that for companies that are operating at serious scale so you know if you're a we're not we're not sort of competing with the adobes the canvas of this world places that are trying to you know make it easy to create single images however if you're wanting to manipulate tens of thousands or manage you know in case of the online marketplace there's millions of images and videos and that sort of thing that's that's sort of where we play so yeah that's kind of one thing is the scale problem and then as i said the that process of getting through that funnel of kind of actually creating manipulating and managing internally and ultimately delivering still a really cumbersome laborintensive process. Lots of actors involved from the creative end right through to the kind of technical end. And so a lot of room in there as well to just make that make that process easier and ultimately empower the people who care about the of substance of the visual media to give there put more power in their hands to kind of deliver the message that they want to deliver to and to customers, end users, whatever it is. So a bit more complicated than one bloke putting up a photo on Twitter. That create, manage, deliver framework is interesting. Has Cloudinary always gone for the end-to-end approach or have they started at one side of that framework? Talk me through the journey. Yeah, sure. So Cloudinary has been going for a little over a decade now. And yeah, as you sort of were touching on, very much like any any good software as a service company started started with one kind of niche part of that and then expanded from there. So where Cloudinary played typically was at at the end of that funnel was about just trying to take images that had already been created curated done and just make them appear quickly and simply on any device at any you know in any low file size so things could load quickly but really you know pretty kind of basic stuff in in a way started life as sort of tooling to help developers manage that process and then has expanded from there and sort of slowly but surely serving a kind of a greater range of company problems and sort of expanding fairly organically back into into that funnel got it we will return to maybe some of the lessons that you and the rest of your team has learnt as you've evolved that proposition in a bit. Before we do, I want to just quickly look back to where we've got to to this point in time. So in some respects, text and AI was the first frontier in in this world. And now we're seeing how this same story is playing out with respect to video and images. What are your reflections on on that first wave around large language models and the fusion of AI and text and potentially some of the things that we should be keeping in mind now that video and imagery is is increasingly part of this story? Yeah, interesting question. And um ah not to a lot to unpack in there, I guess. I think you're sort of right to distinguish between these kind of two these two parts of the the generative AI revolution that we're living through is so, you know, absurdly hyped at the moment, not without some basis. You know, there's the sort of LLM text side, and then there's also all the kind of amazing stuff that we're starting to see in terms of image and video generation. And yeah, I think it's important to distinguish but between those things because they're in sort of slightly different places, I think. you know we're We're seeing amazing kind of cutting-edge you know, things that can be done at at either extreme, but in terms of driving towards like enterprise adoption, adoption, it's actually really practical and usable for companies. We're obviously a long way away from mainstream adoption from both. We're still only talking about what, like a couple of percent, three, four percent of sort of total enterprise software spend is kind of going to this space, but that is skewed towards use cases on the kind of tech side, you know, stuff you hear about all the time, the customer service, you know, chatbots and marketing copy and that sort of thing. we're We're even, we're a little bit further back and at a kind of pretty exciting kind of precipice, I'd say on the, on the visual side. So yeah, look, to give you the the simple potted history, I suppose, of that of the pathway that's been taken so far on the visual side of things, similar to kind of AI more broadly, it's gone from a series of pretty targeted specific use cases, you know, not not powered by the kind of generative models that we've seen.
AI's Impact on Content Creation and Media Landscape
00:08:23
Speaker
So going from using, you know, sort of targeted machine learning models that have allowed for things like, yeah, for sort of basic manipulation of images and videos, you know, oh, I've got to take a video on my phone. It kind of goes all over the place. And I want to then just focus down on the subject matter and, you know, make your, you know, make that the subject, you know, make my GoPro footage look look better or something. We're now in the kind of earliest waves of the Gen AI stuff, which is in general, what we're seeing is kind of use cases and things that were broadly already present in the market. Stuff like, you know, increasing the quality of your images and videos and that sort of thing. That's now the kind of stuff that's being used because it's you know low barrier to entry for companies. It's pretty simple. What we're then going to shift to is going to be a ah kind of starting to play far more with the content of images and videos. The tech's now there to it sub things in and out of of images and videos, to change backgrounds, to start to manipulate what's there and what but that's still something that's not broadly being used at kind of scale for businesses and so yeah we're we're sort of on the on the cusp of some of those those transitions there and then beyond that and we might spend a bit of time chatting through it later there's some really pretty disruptive and sort of exciting changes that could be happening in the space, in the market as a whole that could come about because of this stuff. Well, that's ah that's a lovely segue. Let's look forward and start thinking about what are some potential scenarios that you see in that third horizon that you just alluded to? What are some things for us to expect and what will that what will that look like? Sure. um Before thinking through some of these scenarios and sort of how how this world of kind of enterprise visual visual media is going to going to change, I think it's often helpful, I find, to think through some of the ah kind of natural first order or kind of primary implications of of the tech that we're that we're working with the text change and then you can start to think okay well how's this going to affect the market competitive and dynamics you know and sort of think through those kind of tertiary you know secondary tertiary kind of effects i think a couple of things in in mind i think firstly in terms of actually creating and changing, manipulating assets, the one thing is these pictures and images and stuff. The effective marginal cost of content creation is heading towards zero. It's going so much cheaper, so much easier to do it. I mean, obviously, at the moment, playing with these models, they're still computationally super intensive. It's still quite expensive. But like we're also seeing crazy decreases you know year on year at the moment in the terms of the cost of compute. So that stuff's going down. I think the other thing that excites me personally, I think, excites a lot of people is we're talking about the capabilities is required to actually create you could have really cool, you know visually arresting exciting images you know the stuff that we are consuming day in day out every every you know almost every minute of every day depending on how bad your social media addiction is the cost of the like requirements in terms of technical skills and competency to create that stuff is goes going way down as well as you know effectively some democratizing this of creative capability.
Changing Roles in Marketing and Creative Industries
00:11:47
Speaker
And so I think it's important to take those kind of two things. There are probably two big things that are generally sort of positive and exciting and then couple that with some other changes as well that I think we can probably imagine. So before you get to those other changes, on that second point there and around how this makes it just considerably easier to create a restive creative content, does that present a risk for people in say marketing functions or social media managers or any number of different creative roles in large businesses? Are their jobs under threat in this new world? It's the considerably more than and a million-dollar question, right? How how that will happen, how how value will will shift as a result. I think there's a couple of things that are semi-intentioned that are worth thinking through. I think one is in classic any technological revolution for you know kind form we've got, the fact that we are just going to see a proliferation of content we're going to you know we're already seeing it right but there is the possibility of just creating more of doing more means that in some respects yes the role will change but because we're going to be doing so much more that you know people are still going to be in the job i think secondly you can probably assume that the role people need to play is going to change the kind of basic technical mastery of your photoshops and things like that. Whilst, you know, always going to be there as, you know, people who really want to be precise with what's happening. You're talking about just people playing more of a, I know, like if my mental model of it is shifting from being more of a kind of, to being more of an editor, I suppose, someone who's, you know, kind of critiquing the work of the AI and and and changing that. I think finally, and and probably the thing that kind of gives me hope that these people, that people in these kind of roles and creative roles are really not going out of a job anytime soon, as We're still talking about aesthetic judgment, about standards. It's not to say that like an AI couldn't do this as well or better, but it's still just a judgment call. And so ultimately people, I think, will still be involved in that process, as I said, more as editors, kind of curating and coming up with something that will be coming up with the kind of design language, the vision for the what the you know the campaign could be and that sort of thing. Maybe working more and more collaboratively with an AI. But yeah, I think ultimately still still going to be there, at least for the the conceivable future. Yes, you're almost moving into that philosophical question of whether AI can create great art and great poetry and all that sort of stuff. And that's, I think that's still an open question. In any case, you've told me that in effect, what this will do is it will make creation cheaper and it will make it easier. Pick up the story from there. What does that mean? Yeah. I mean, it's that in a nutshell. And then the last thing we're probably worth keeping in mind is and something that's going to be a little bit different to some of the past kind of technological booms that we've seen with you say cloud and SAS and and that sort of thing is that we're going to, because of the nature of the tech that we're using, its ability to ingest unstructured data, its ability to kind of generate and create workflows that aren't kind of as rigorously defined by rules, you're seeing going to see kind of a shift in the kind of traditional advantages that companies who provide this software have. I mean, I'm a strategy node and nerd at heart, and in know, as a kind of in SaaS in particular, you often rely on a couple of that kind of core incumbency advantages. If you're so if you're a successful company, you require on, you sort of rely on being embedded in the workflows of your customers. So you're sort of hard to hard to hard to leave. you You benefit from, you know, having lots of data from your customers that are kind of hard to shift off. Think of your teams or your slacks or whatever who have you know all this this amazing treasure trove of company data and it' kind of makes it's reasonably hard to sort of shift off those platforms as a result. that Those kind of incumbency advantages are going to change pretty significantly as a result of you know AI's ability, the suite of AI technologies we're seeing now is an ability to ingest and understand pretty unstructured data.
Future of Visual Media Creation and Management
00:16:14
Speaker
So with those kind of things in mind, some kind of broad principles that you can pretty you can debate how much of an impact they're going to have or how strong of an impact they're going to have, but they're definitely there. I think you can then kind of envisage a couple of futures in this in this space and ah probably best to do that in, I think you sort of articulated the framework pretty nicely before, of you know, the kind of creation, the management, the end user experience of of visual media. I think all three of those are going to see some pretty big changes. We've been kind of dancing around the first of those, right? I think the first is the the ability and um ah to create visual media and manipulate and and to transform it as you need. That's obviously we're going to see a massive shift as we were talking about the kind of shift in responsibility towards the requesters of content, the making people more editors away from the kind of pure kind of creative role. Won't dwell on that one too much. But I think one thing that is really kind of interesting as you think further and further ahead is how do things like like aesthetic standards change in this world where it's super easy to create content? Are we going to be you know in a world where actually, you know, when every knock and company that's dedicated to flogging knockoffs on Amazon ah can create content that looks like it's been made by you know Nike or something, how it is how how do we start to react differently to visual media as a result? Yeah, some super interesting questions. And also, how can brands still how can brands aim for distinctiveness and differentiation in a world where in the past, Nike had an enormous marketing budget to be able to create that level of distinctiveness and differentiation in their marketing. And now, you know a kid in their basement may very well be able to produce a similar quality of output to what the Naki marketing department could do. That's a real challenge for marketers, I would have thought. Yeah, absolutely. yeah absolutely As I said, it sort of shifts that that focus. As I was alluding to before, it becomes less about just the technical proficiency and competency and and you know money you're willing to spend on your on your some creative teams to create and you know the absolute top-end photo shoots and to be able to reshoot it in 100 different ways, to have it you know sort of just right. Yeah, it means hopefully a kind of elevation of the importance of the the message and the the aesthetics and the style and whatever it is else it is that you're trying to appeal to and have people respond to. The second limb on the management side, talk me through some of the changes we're going to see there. Yeah, so internally, as I was saying before, there's obviously a huge amount of manual effort, time, energy invested in trying to manage you know manage all this vast wealth of content that is being produced by by organizations. And so you're starting to see here the same kinds of changes that, frankly, we're seeing in almost every enterprise software domain. You know, you start to shift towards more text-based interfaces and, you know, through your LLMs.
AI in Content Workflows and Brand Interaction
00:19:30
Speaker
And we're going to have to think about how do we manage all these variants of images and and that sort of thing. So that's kind of the table stake stuff. You know, it's interesting in this space, some debate about how valuable text-based interfaces are going to be, because, you know, ultimately, if you're playing with images and videos, you you often want to be able to kind of interact with them visually as well, not verbally. So there's the stuff there. I think the longer-term view here is, again, probably not dissimilar to some of the other kind of enterprise big enterprise workflows that we're talking about changing. It's shifting towards kind of more more autonomous yeah agents, one of the better word, AI driven- agents who are able to drive a lot of the workflows a lot of the processes around kind of your your media around moderating it seeing if it's appropriate getting it online taking it offline and decommissioning it all all this stuff you're starting to see kind of groups of aid you know ai driven agents probably able to do this more autonomously, which, yeah, again, begs the question of what kind of software and who what kind of providers are going to be the winners of this? Is it going to be specialist provider but specialist providers or is it going to be the big players who are offering generic models and you know you the open AIs of this world or whatever who are able to jet provide a general capability that will be able to do this for you. Obviously, maybe sort of sort of a bias as a provider, so hope someone who works for a company that's firmly in the former of these he camps. But I think this, is of the various enterprise domains, I think this is a space where specialist software is more likely to come out on top for the reasons I mentioned, because you actually want all these additional UI driven tools, like visual tools, not just, you know, generic text-based interfaces to be able to play with your images and your videos and edit them and that sort of thing. That's that kind of stuff, I think is, it's going to lend itself towards kind of specialist providers, I think. You mentioned that AI enabled workflows will be running more autonomously than in the past where it is managed. In the past, it was managed end to end by by by person. Is there still a role for the human in the loop to use that like slightly hackneyed phrase now, as you get more AI-enabled workflows in that middle section, in that managed section? Yeah, I think for the next conceivable while, frankly, until we're on the cusp of you know ah not of general intelligence, you know I think we're looking at humans in the loop still being in, humans still being in the loop for for a lot of this. I think, firstly, because as I said, we are still talking in many cases about ah aesthetic judgment. we're talking we're We're not talking about something which is like, can bot create a script for a customer problem, you know, a customer service request that is yeah good enough and solves their problem and and moves them on. we're We're talking about how you want to represent your brand, how you want your company to appear to the outside world, which, you know, increasingly we do with images and videos and visuals. And so, you know, I think humans will remain in the loop for that kind of creation and sort of deciding what where what kind of direction we go. So that there's that. And then I think in terms of the kind of the internal workflows and the management and that sort of thing. Again, I see this as as a case of sort of us really progressively moving towards, you know, more and more autonomy, agents doing doing more and more. But yeah, it'll be a ah longer journey there. I think if there's any lesson learned from some of the kind of previous waves is that larger enterprises in particular can be pretty risk averse, justifiably so, given the level of scrutiny they can be under. And so you know it's it's not going to be something that that changes in in the next year or so, I wouldn't think. Let's move from create to manage and now on to the end delivery to the consumer. We're an experience podcast. How does the customer experience look different in this new world? So there's there's a a couple of ways this could go. and I think all of them are pretty interesting. I think one and is probably something ah not dissimilar to what you're see you would see internally for the organizations that are serving this kind of visual media. You're going to be seeing customers as well that increasingly interact with providers, with vendors of e-commerce, media sites, whatever, through intermediated through their, or potentially intermediated through some kind of you know AI textual interface in some way. you know I think it's not ah not too big a stretch to assume that if you're got you're going to, much in the way you can already use a Google shopper or something that you might end up having kind of an aggregator you know you're kind of and so an aggregator and an agent that is ah spend effectively disintermediating your the customer from the kind of end website that they want to interact with so you the AI is making the the judgment about what it pulls and what it serves up to you as a user. And so there's a world where, as consumers, you're you're not direct interacting directly with yeah with these these end websites. So I think you're probably you can sort of imagine or envisage a transitional state, one that's broadly already happening now, where you have, you know if you're going on to Google or ChatGPT or whoever your complexity or whoever your provider of choice is, that it's scraping sites and kind of aggregating and pulling stuff together and serving it up to you. I think what you'll see increasingly is websites that are optimized for that. Indeed, not just, you know, making life easy for for an AI bot that is coming over and scraping them, but is perhaps even interacting almost, you know, agent to agent with those to try and serve up the right kind of content. So that's sort of one kind of bucket of change is that you you might not be going to your your favorite website and seeing, you know, and going to their kind of curated online shopfront experience, you might sort of lose that. You might, it might sort of, companies might lose that kind of ownership of the end user interaction. That's probably one thing. And then the other, I think, again, is you're seeing as a result of stuff I mentioned earlier, the cost of production going down, of being easier to produce way more of these, you're going to see a vast proliferation. Think of a variance and of the types of images you're seeing. You're going to see people will experience images, videos that are far more personalized to them, to their needs, their standards, their background, whatever it is. You're going to see stuff that is potentially even going further down the line, more interactive or dynamic, because again, it's just easier to create this stuff on the fly you can start to create a how-to explainer for your next product you want to buy in the voice of an influencer that you already like and can answer questions you have dynamically about how the product's going to work and that
Challenges and Strategies in AI Adoption
00:26:43
Speaker
sort of thing. So yeah, I think potentially a pretty radical change for the consumer and their experience. As a business, how do you go about proving the ah ROI for something, for an experience like that? So take the example of we've got an initiative in place whereby we're going to implement a particular tool which can communicate to our customer in the influencer's voice that they particularly resonate with. How do you go about saying, well, what's the value of that? How do you demonstrate the the value? Or do you just say, well, that sounds kind of cool. Let's give it a crack. Yeah. I mean, the question around ah ROI of these sorts of initiatives is a, yeah, and this particular kind, I think, touches on it on a broader point of just in general trying to prove out the value of these but these kind of AI, generative AI experiments that companies are launching. I mean, at the moment, we're seeing huge wealth of, and were we're seeing ah so many of these use cases kind of going and it being experimented and then not ah not as many. I mean, I think I've seen some survey data that's saying from between 25 to 35% of these experiments are actually being implemented. Most most aren't. And you know I think in part that is is because proving the ah ROI is is is tough. So a couple of thoughts on that. I mean, the the direct answer to the question is, you know, you can look for any of these kinds of experiments. You can look, you know, particularly stuff that you're doing online. It's it's relatively easy to set up the A-B test and have the counterfactual and just sort of say, look, you know, here's our conversion rates with and without that sort of thing. It's sort of trivial, but, but not oh, not trivial, trivial, but you know if you're getting enough volume through it, it's doable. I think what is interesting here is, ah again, how do you is it how how do you sort of set up in such a way that you can firstly de-risk the adoption of these sorts of use cases and the sorts of vendors you're using and and give yourself give yourself some confidence over over the longer term that you're going to you're you're going to get there. So yeah, a couple of things. I think probably first and so foremost, there are a couple of sort of sources of risk I think that's worth managing. I think for companies experimenting here, one, as I mentioned, is this vendor risk. I think trying to find a vendor with whom you can have a pretty ah an honest conversation about their roadmap and the pathway that they're taking so you can actually sort of buy into the journey and see that it's kind of aligned with it with where you want to go is one thing. I think, two, finding a vendor that's willing to be, that is gives you a certain flexibility. I think this is something that ah you know my company Cloudinary has certainly learned in the kind of previous waves of AI-driven tech that we've tried to deploy to market is that you know early on, you want to you can't be too opinionated about what these solutions will look like. You actually just need to provide, it can often help to provide early adopters, the tooling, the the basic, the kind of building blocks that they can use to then stitch together solutions that for them are going to be most valuable. And over time you see those and you build up and then you can start to say, well, right, actually, you know, 10, 15% of our customer base is going after this solution in this way. Great. We can turn that into a product and sort of go from there and turn that into something that we can then take to other companies and go there. So I think for companies, yeah, having that kind of flexibility is also super important. It's a good segue to my final area of questioning, which is around lessons for leaders. What are some broad pieces of advice would you give to senior leaders who are grappling with this AI wave in visual experience and what it means for their business? Yeah. So for leaders who are thinking through this as a premise, who are the generally who generally owns this responsibility in most organizations? Is it the the CMO normally? So in the visual media space, yeah, it can sit in a couple of places, but yeah, it's it's typically sort of a CMO or sometimes you know for more kind of technically or minded companies, it can be a sort of senior product leader as well. But yeah, typically the the places we see who are kind of owning or, you know, sort of a leader of e-commerce as well, sort of a digital head chief digital officer, that sort of thing. these are These are people who often kind of own this this agenda. Yeah. And so, look, in terms of advice or things that,'m thinking through, that many in this space are thinking through, I might sort of break it into into two buckets. like I think a little bit firstly about the kind of challenges and and things we need to think through as any organization that's trying to implement AI internally, sort of it's for its own operations and that sort of thing. And then I also think about companies that potentially, you know, vendors in this space or where their core product and core capabilities are at significant risk of disruption. um And I think it's sort of both, you know, sort of considerations are are slightly different. On the former, I mean, we we started to touch touch on that a bit. It's about how do you find ways to de-risk your initial investment by you know starting small with you know use cases, ideally um often ones that you can kind of co-build with your your vendors or your service providers, whatever it is, who are helping you on your AI journey and sort of going from there. Ideally going with ones that have a compelling roadmap for the future so you can kind of go along with their journey as well a little bit. You know, your they're going to help drive. Ones who give flexibility, who aren't forcing you to sign into kind of a big monolithic sort of offering, but ones that sort of give you a little bit of discretion to kind of pick and choose and go after the use cases are going to be highest value for you and your organization it's probably probably one sort of big piece of advice and then the second and the other within there i would say again kind of ah alluding to this well coming back to this point you mentioned earlier around human in the loop and trusting these outputs and that sort of thing is is don't underestimate the the non-sexy stuff
Human Intervention and Future Planning
00:33:08
Speaker
here. like Don't underestimate that actually in an enterprise context, yeah know for most big businesses, yeah risk-averse, big groups of users who care a lot about a brand that has been hard-won over years. Often the stuff that's going to, in the next, I would wager, next two, three, four, five years that's going to be important here is going to be less the technical ability to solve a use case, but it's going to be how do you hit the right level of how how do you get enough confidence in the output that you can actually deploy this meaningfully to your customers? So how do you get the right levels of human intervention? How do you have the technical ability to check your outputs and you know automatically make sure that you know you're not getting you know things that are going to go off-brand or or sort of damage you in any way? So yeah, I think they're probably two big things that I'd keep in mind for companies that are implementing AI. And then coming to that second bucket I mentioned about about companies that are facing disruption or at this kind of inflection point where were ah could either things could either go really well and you ride the wave or or yeah know end up disappearing into obscurity if you if you get it wrong. Look, you know this is in some respects no different, I think, to some of the other big transformative technology technological waves we've seen in the last couple of years, whether it be the yeah rise of, whether it be previous AI booms, whether it be the shift to mobile, e-commerce, whether it be SaaS, that sort of thing. I think that maybe the only thing differentiating is just the pace at which it's happening. It is meaningfully happening a couple of times faster than, say, the adoption rate of SaaS or something. Here, I think all you can do is try and think through, as we've done a little bit today, some of these some of these scenarios and try and anticipate, I guess, ah sort of firstly understand these kind of key drivers of uncertainty you have about the way your market is going he going to evolve. Try and characterize those in a couple of compelling, you know viable scenarios that you need to anticipate. You have to choose one. You have to choose one as your likely base case that you are then going to base your roadmap and your investment strategy on. But be super, super alert to what you are looking for from the
Podcast Conclusion and Listener Engagement
00:35:31
Speaker
other scenarios. And if it looks like your market is heading that way, make sure you've got the wherewithal, the flexibility to try and pivot towards them as well. Sam Pickover, thank you for coming on Speaking from Experience. Yeah, my pleasure, Will. Thanks so much for having me. Thank you very much for listening to this episode of Speaking from Experience. If you enjoyed the show, please leave us a rating and a review. And if you really enjoyed the show, reach out to us for a chat. A link is in the show notes.
00:36:10
Speaker
you