Latest / The Interline Podcast / What Business Spend Data Really Says About AI Adoption

Transcript

What Business Spend Data Really Says About AI Adoption · The Interline Podcast

• speaker-0: Hey, welcome back to the Interline Podcast. If you're listening to this, the odds are that you're spending money on AI. At the time I'm recording, 55% of American businesses are, and those figures are probably reflective of what's going on in the EU and the UK as well. Then let's add in the fact that the Interline's audience is made up of fashion professionals who have a vested interest in staying near the cutting edge, and you can see why it would be selective. Now, to be clear, you might not like AI. Or you might not be seeing a return on your spend. It's a different story. And it seems clear at this point that cultural attitudes towards AI and business spending on it are on divergent tracks. That's actually one of the main themes of our 2026 AI report. So if you haven't already, I'd encourage you to grab your free copy of that for additional context for this conversation. But while paying for AI might be a binary on-off state, what you're paying for. what results you're getting, what capabilities are the most deployed across the huge surface of image generation, coding, general knowledge, cowork, and so on, and which vendors you're buying it all from, as a much more complicated mix. And it's also a mix that it's kind of been hard to find numbers to substantiate. There's all manner of consumer surveys about how people feel about AI on the street, how they use it. And our AI survey from the AI report I just mentioned is focused on extending some of that sentiment analysis and maturity benchmarking straight into fashion. But when you're looking at a potential transformation on the scale that AI advocates are promising, you need bigger than consumer, you need bigger than fashion, you need the broadest and the most technically robust statistics you can find. For us here at the Interline, and for a lot of media, the AI index from RAMP. Has become the go-to source for that kind of economic data and for expert analysis of it. That data set is behind a lot of the articles you've read about AI adoption, AI spend per employee, and about the rapid ascendance of Anthropic to lead in enterprise market share after OpenAI had dominated for so long. That index is put together by RAMP's Economics Lab, which is a four-person team embedded in a much larger financial technology company. That team includes an economist, a director of applied science, a head of analytics and data science, and a lead economist. So not the kind of team you usually find working for a fintech company, in other words. Today's guest is that lead economist, Ara Karazian. He used to lead economic research at Square, and you might have seen him in the New York Times, on NPR, or in the Financial Times. He's become, in a pretty major way, one of the most visible faces of the economic side of AI. And As companies across fashion continue to frame AI less as an open-ended possibility horizon and a money sink, and more as a capital expenditure that needs to deliver a quantifiable return, more and more brands are interested in the commercial and the behavioral side of things that Ara and his team cover. So I asked him to come on the show when we were putting the finishing touches to our AI report for the year. So our readers and listeners can get the most complete picture we can offer of AI in fashion. Both from directly inside the industry, from our survey and our analysis, and from a much higher and broader level from the economic research that Ara and his team have been doing. So here we go. Okay. ⁓ Aura Farosium, welcome to the Intelligent Podcast. speaker-1: Thank you, Ben, for having me. speaker-0: Not all. I'm excited about this one. we have we've referenced data that you have had a very direct hand in producing multiple times, I think, over the last year or so. So I've been looking forward to this discussion. Now, we start these shows pretty universally by doing two things. We ask the guests about what their day to day lives look like, and we ask them to define something. ⁓ your day to day is super interesting to me in a couple of ways. The first is that you work for a tech company that does spend management. So ⁓ accounts payable, employee cards, expenses and so on. None of which, like no offense, is is outwardly that exciting. And it's not that thrilling unless you work in finance or HR. But you're also you're a fully fledged economist of the kind you normally find. I would expect people like you to be working for banks or advising on monetary policy at think tank or something like that. So that's one piece that I find interesting. The second is Weirdly, we've had a bunch of guests recently who are also in the media business in a way, despite you not thinking that from their job title, because you produce written analysis that people then bookmark and come back to and put a lot of stock in. I've heard people say that your research and the AI index in particular accounts for something like third of all ramps, like non platform traffic. You're routinely cited on the big news. There's a clawed connector I have ⁓ that l live and claude cowork and code code where I can go and pull live data from the RAMP AI index. There's there's just a ton about your day that I'm interested in. So tell me how you got here, one thing. How do you end up as an economist working for a tech company? ⁓ and tell me what you actually do when you sit down in the morning. speaker-1: There's I mean, the joke that we used to make about this job is that when I was hired, it's because the rates were way too low that the tech company could shell out to hire this kind of researcher that's not even focused on the product. But once you think about our product and you think about the data set that we might have access to, I think the value of this job becomes very clear. So Ramp is a corporate card bill pay ⁓ product. It's an AI finance platform more broadly, but if you think about the core data set we might have access to, it's all coming from the business spend that's occurring on our platform. So that's also the data set that I work with. So think about the universe of business spend essentially anything except for payroll. Now you think about other parts of the economy ⁓ that already have access to great data sets. You know, you see consumer credit card companies routinely producing research about ⁓ consumer economic trends using their unique data sets. You see all these sort of housing platforms on the internet ⁓ reporting on the price of a house in this or that city. So in our economy, there's generally a lot of great private sector data sets that are used to drive important decision making. You know, you buy a house, you go to, at least in the United States, Zillow, if you are ⁓ applying to a job, you want to check the salaries online, you can do that with something like Glassdoor. There is no data set, at least public data set, that tracks where and how businesses spend. And you can think of that as the main value proposition of the research that we do at RAMP. We collect this data. Everything we published is aggregated and anonymized, so there's no privacy concerns with whatever we release. But we're producing research about how fast moving businesses spend, ideally with the goal of helping other businesses ⁓ make similar decisions ⁓ and make those decisions with a better data set. So lately in the age of AI, that means how do you buy AI for the first time? This is a new innation technology for which there isn't a great playbook for how to integrate it into your very complex business. And when you go online, it's also the kind of sector that attracts a lot of pundits who may have a profit incentive for the kind of advice they're trying to give you. So one place we've tried to start is: hey, we see how all these businesses are paying for AI to start. Can we publish those kinds of trends to then help other businesses get on those same adoption curves and then drive similar growth? Through AI adoption. Who are they buying AI from? What's the difference between OpenAI and Anthropic? And then what does it really mean to use AI effectively? Is it just a chatbot for everyone in the workplace, or is it using something more advanced like a coding agent? And so that's that's the general ⁓ focus of my work is how do we publish this kind of data about how businesses are operating today, how they're changing. And then can we ideally make some recommendations so that other firms can adopt those same changes and and grow faster. speaker-0: So you've sort of punched there that I a hundred percent agree with. I think in fashion as well, there's been a long history of companies kind of wrestling with the idea of best practice. ⁓ this is this is very kind of idio idiosyncratic to the industry, but ⁓ it it it's it's a market made up of brands that have had a long time, some take some cases centuries for but free maze lots and so on, operating as islands and doing things their own way. technology responds better in some cases to doing things in a best practice way. There's more of an argument to be able to say, well, everybody is doing this and this is why they are doing it. And it is the most effective way to do things. So you can benefit from that. ⁓ so I I think the data set you produce is incredibly valuable, as we've already talked about. I think for fashion in particular, it's this shift to AI is a an important sort of juncture point because it represents that difference between The way the industry has historically thought about its own processes and its tools and so on and what is now a onrushing sort of technology wave and treating that as an opportunity to say, Well, if everybody is doing things a certain way, that's probably the best practice. ⁓ and Vashion has had that kind of push and pull relationship with best practice that I don't know that necessarily other industries have had. speaker-1: Well, I agree with you. I mean, you often see the same trends in tech though, too, where large software companies, you know, when we're talking about B2B SaaS, right, become very entrenched in a specific sector or for a specific business process. Even if it's not if it even if it's not necessarily ⁓ the most up-to-date tool for a task, there does tend to be a lot of vendor lock-in and there does tend to be a lot of stickiness. And you know, ultimately part you know, particularly when you're thinking about these technological adoptions in terms of firms' operations. ⁓ smaller firms are much more nimble. They are much more able to ⁓ quickly move between different software vendors or adopt new technologies fairly quickly and to onboard all their employees with new technologies. That is ultimately what enables the same startup culture that has driven the tech sector for now a few decades is that Large companies are rarely safe when they stop innovating. You know, but we shouldn't have this impression in our heads that because the tech sector created these technologies, that they are immune to ⁓ the kinds of changes that that ⁓ that may affect the market share of even the largest players. speaker-0: That's interesting point. ⁓ and you just mentioned vendor lock in there. That was actually the term that I was going to ask you to define. So I ⁓ there's a thing that it means to me as somebody who's been around for a couple of turns, ⁓ of technology and fashion, like sixteen years or so. The way that traditional enterprise tech has been rolled out over the last sort of previous decade or so, ⁓ has been defined by long custom implementations with a bunch of service days, i.e. consultant and advisory time baked into it, ⁓ supported afterwards by maintenance contracts. but very heavily kind of like we will customize and bespoke things for you with the goal to building up ARR as much as we can as a as a technology vendor. And to do that we will architect all of this in a proprietary data model, we'll have to build bespoke integrations and portals for your suppliers and so on. Like you get the idea. That's that's what vendor locket is meant to be. that's how it gets its hooks into organizations at a cost level as well. Like that model that I've just described is very expensive. ⁓ companies in fashion spend way too much on building systems of record to ever move away from them, even if over time you actually end up with less data in there than you thought you would have, you know, a bunch being abstracted out. For ⁓ kind of people building workarounds and things. Like long story short, when you chuck a million dollars at making something work, you're suddenly very incentivized to use it as the platform to make everything work from there. Now the opposite side to what I just talked about there would be somebody saying, Well sure, but it that the the technology was really important. So we had to do we had no other choice. Like we had we had the the technology was that vital to the growth of the company and the outward market and the world was changing, we we had to do it that way. That's what kind of gets me to AI here. Because I think, like I said before, if if this is the most pivotal tech is changing what it means to be a company moment, ⁓ then it feels like technology companies sh w should be championing it the bit to do that run that playbook again, to lock up their share of it, right? And that's that's that's where my mind goes when we talk about vendor lock in. You said, I think it was kind of ⁓ not long ago, but like last month or so, that unlike traditional software, AI doesn't really have vendor lock in and that the most advanced AI shoppers you see from your data run something like eight AI vendors at once. I'm gonna hazard a guess that the big A the big AI labs, your Google, OpenAI, Anthropic and so on, would like to get that down to one or two. What is it that you think makes this that unlikely? What is it that you actually think makes AI different when it comes to vendor lock in to what I think of it as in traditional software. speaker-1: Well, it it's think about the way that you use AI in your day-to-day life. Right. Like there's no there's not much of an ad all all these traditional features of a product or software that a vendor can build to keep you on the platform doesn't really exist in AI. You know, it's not like a Salesforce where you, you know, it's an extremely complex technology that, you know, involves, you know, building integrations through all of your workflows and training all of your employees to use exclusively Salesforce to to to sort of attract customer interactions. AI is not like that in that it's actually pretty easy to switch between different models, if not advantageous to be using multiple models at the same time. ⁓ I mean, you can s not to not to say that there isn't advantage to having high market share and holding that ⁓ and and to, you know, ha being early in the marketplace may sort of entrench a firm as one of the The defaults. That really is kind of the story of OpenA and Anthropic is that OpenAI was for a long time this consumer default, in that, you know, when you thought about AI, you're thinking about Chat GPT. And because AI is this new and nation technology that businesses are buying for the first time, as well as consumers, if you are a business buying AI for the first time, you're probably going to buy open AI. At least that was the case from 2023 through most of 2025. ⁓ open AI's, at least in our data center app AI index, we found its adoption rates shortly after launch, at least to enterprise, shot up to about 23, 24% as of May 2023. Those were the the sh the sort of quarter of businesses in the United States were already paying for open AI. Most businesses had never heard of Anthropic, probably weren't aware of Google's offerings in that area. And then it sort of hovered around there and grew up to about 30, 40%. Until Anthropic started to make significant inroads into that market. Anthropic, originally popular with technology companies and particularly early AI adopters, though in a much smaller way, started to grow its market share by several percentage points month over month, which you don't see in most sectors. And then toward the end of 2025, beginning of 2026, skyrocketed in terms of adoption rates. Ultimately passing OpenAI in May 2026 and becoming the most popular AI company, at least amongst businesses. So, you know, you have a story here of a company, OpenAI, that was the first mover, became the consumer default, in many ways became this business default. No one really expected that to be toppled by another AI lab. Maybe we would have expected a large, you know, heavily Resourced firm like Google to take some market share. But it ended up being Anthropic, which was a small nimble firm, ⁓ was able to develop products that were especially geared toward enterprise, was particularly successful in activating those early adopters and evangelists of AI at a firm to then spread AI to more firms and then within the firms they were already at. And then also and then Finally becoming ⁓ the number one player in the market. speaker-0: In your in your experience then are most companies running both now. ⁓ 'cause I'm guessing it's not a zero sum game in most cases. It's not a case of okay, well we were using codecs and now we're switching exclusively to using Clock Code or what have you. And and I'm I'm guessing a good amount of companies or a good amount of people have those two icons side by side in the dock. Well speaker-1: Most companies that are using at this point, yes. Most companies that are using one are also probably using the either. But remember that every f in the same way that every firm is on its own AI adoption curve and that there's some firms that are more advanced than others, you see those same dynamics within a firm too. So, you know, the engineers may have been using anthropic for a long time. That's typically how we saw anthropic proliferate through our data set, is that when it started at companies, it was primarily being used by the engineering teams. But then over time, other teams start adopting these tools and technologies too. So the marketing team, the sales team, ⁓ product team shortly after energy engineering is generally what we see in terms of the flow of adoption. And so other teams may start using multiple models. And that's generally what drives then the rising costs of AI across a firm is that a firm buys the technology. But it's extremely variable as far as it's as far as a tool that is being used. You know, each person as they adopt AI is going to drive more spend on their own part. And then as AI proliferates through a firm and more employees at the firm, particularly in non-technical roles, start to adopt AI, then you see even more increase in spend across the firm. So both as a function of the intensity of an individual's usage and then the number of teams. within a firm that are using it. Yeah. speaker-0: And we we see we see that firsthand. So our ⁓ our kind of repeat monthly billings with Anthropic, but we have token budgets ⁓ and token spend across Open Router, and also across OpenAI. So yeah, I I I understand that. And and for the these things are used for different functions. I think if I was one of those vendors and I was chasing Waqin and I wanted to try and run that traditional technology playbook, I would be leaning into connectors to the extended enterprise and skills for specific disciplines and specific industries. And I think we you know we we see that with Anthropic, right? You see and and also OpenAI, but Anthropic comes to mind as as having a specific offer for the financial industry, for instance, or having a specific offer for the legal industry, which encompasses skills, plugins, mm connectors to libraries like Lexus Nexus or have you, if you're if you're in law. That to me feels like the way to get people into your ecosystem and keep them there is to be able to say, Well, we have connectors that the other place does not. speaker-1: Well, and it and remember, that means that much of the development of AI, at least as an industry, is gonna come not from making the models themselves better, but making them perform better in production for an employee by giving those models access to data sets they don't already have access to. You know, it's like literally being able to connect to your enterprise workflows is ultimately what's gonna drive the productivity gains that we hope to see from AI. speaker-0: Yeah, I think I think that's right. And I think there's some of those bits there that you can rely on most people having. You know, every company is either a Google workspace or Microsoft has with some variation in the EU for companies that use Proton or what have you. ⁓ but then it gets much more heterogeneous from that. Not y you can't assume that everybody is a sales force. There's a lot of different CRMs out there. You can't assume that everybody is on Shopify, although it is the dominant kind of e-commerce platform. So I think I think you're right. I think that that level of how well it works for your enterprise is definitely more important than frontier mobile capabilities. speaker-1: That that is also constantly changing. And I think that's ultimately going to be great for the businesses that are adopting AI in the same way that it's great for consumers. Is that this is an increasingly and extremely competitive industry where similarly month over month that we see these ⁓ changes in market share, we also see the companies trying to sell AI, ⁓ employ their competitive tactics to try to get the others customers. So it's not just about releasing better models, but like we routinely see these kinds of price cuts that are enacted so that, you know, you can, you know, make sure you're offering the most price effective model for for any given customer. And so that that that is gonna favor firms that are model agnostic. You know, the businesses that benefit most from this are the ones that do not employ vendor lock-in because they don't need to and you don't generally want to. Given that there are so frequently gonna be new model improvements and price cuts in this market that allow you to take advantage of the best model at the lowest cost. Now late lately that's meant employing what you know, you just mentioned open router. You know, the vast majority of firms on our platform, even of the tech forward firms, aren't even using open router. So for listeners, open router is an LOM inference. Platform. It's essentially one of the platforms businesses can use to get access to models outside of OpenAI and Anthropic. So you can use OpenAI and Anthropic, but in prime, but you can also access have access to some of the open source models and some of the models offered through, you know, if you've seen the buzz about China, Chinese models. So they often are cheaper. But on our platform, you know, less than 10% of cust companies are even using those. ⁓ And so OpenI Anthropic are relatively less inclined to then respond to those competitive pressures from open source models and Chinese competitors. But as that share goes up, you're gonna see their market share be threatened as well. And then OpenI and Anthropic will start enacting some more price cuts in response. That's ultimately gonna benefit the firms that are using multiple models. speaker-0: That's a th that that's a good answer. I will say, and I'm sure a bunch of listeners feel this, the the usage and pricing page ⁓ or area of the club desktop app is the most confusing thing I think I've ever seen in my in my time in business. ⁓ like sometimes I look at it and I'm like, How did we get here? How did we get to the stage where I have f five different sets of limits and different sets of usage credits and incentives that are being pushed in my face for fifty percent additional usage by a set period. ⁓ I will say that there's there's some arcane stuff going on with ⁓ with model pricing across those those big companies. But as you said, that's because they are they're fighting fiercely for share of spend. How much visibility do you have into the composition of AI spends? We just talked about I think the total amount of it. so if I think about fashion, right? If you ran a pure token in and out. speaker-1: Yeah. Yeah. Yeah. Okay. speaker-0: analysis and again for listeners, tokens are roughly represented if we think about text ⁓ of words. It doesn't quite map out like that something like point seven five of a word represents a token or something. But if you ran a pure token in our analysis profession, most of the usage in our industry would be visual. It would be image and video generation. in part that's just 'cause we're a visually geared industry. and I interviewed the CEO of Photoroom a couple of months ago and ⁓ he put it this way, which is that fashion sells images and then it has to fulfill the promises of those images with product. but it's I think it's also a matter of AI maturity or at least the perception of AI maturity. We ran a survey at the ⁓ in the springtime, the results of which be out by the time this episode is out. ⁓ much narrower data pool to what you're dealing with, but ⁓ in it people tell us that they believe image generation. ⁓ both for in-house use at the beginning. So you come up with a design you want to present it to colleagues, you want to put mood boards and things together. And then downstream in campaign photography and product detail page e-commerce imagery. Is people perceive that to be way more proven and reliable than say retrieval augmented generation across an extended tech estate that includes ERP and CRM and PLM and all these other things. ⁓ how much visibility do you get into the kinds of AI that companies are spending on? Like w do do y can you see what the capabilities are? Do you just see it at the pure kind of token spend and subscription spend level? speaker-1: No, we do. And that's that's been the focus of our research going forward. I mean, when we started this work, we were focused on adoption rates because AI was so ⁓ early in its implementation at most firms that the only question we really could effectively answer was, hey, does this firm use AI at all? now as the sector has developed, we can start measuring the intensity of that adoption and use that to separate firms in terms of their place on the adoption curve. So I mean the most the most costly share of AI spend is API usage. So that's generally going to correspond to usage of coding agents, but it's also ⁓ the use of AI to label ⁓ large ⁓ corpus of information. So you can imagine the way Ramp often uses it, the way RAMP's product uses it is, you know, you take a receipt, you upload it to the platform for your corporate card reimbursement, and RAMP's AI will Go through that receipt, label it, and file that expense for you. So that's sort of that kind of product experience that we use day-to-day in many AI native products. You don't even necessarily think of it as using AI as a consumer. That's going to be one of the biggest categories followed by coding agents. ⁓ And then some of the other really advanced forms of AI adoption, particularly in consumer companies, are ⁓ AI, as part of it is used for customer service operations and sales. And that can take a lot of different forms. Everything from you may, you know, you're talking to a customer service chat bot that at this point had become quite advanced that many of us have probably spoken to in AI and we didn't even know it. ⁓ they have voice agents that are quite advanced now too, that they might, you know, use to try to close your inquiries before they send you to a a human ⁓ agent. ⁓ image and video generation is increasing in size. ⁓ but it is a very small share of actual AI spend. And I think that, you know, you can you can imagine why it's not extre it's not particularly scalable in the same way that using AI for a customer service operation is where you know you're you're deploying it across the entire organization handling s of calls a day, ⁓ whereas image and video generation, particularly for advertising, tends to be much more ad hoc and team-based. speaker-0: Thousands of Yeah, no, that seems right. one thing I just want to pin down, I think, about your the data that you have available to you. It's US only, correct? ⁓ and I so we're a UK complication with heavy EU readership, although the US is our biggest market. So I just want to understand the geographical limits of it. But I also just want to quiz you a little bit on the enterprise nature of it and the kind of fast growing company nature of it. Because ⁓ fast growing company could mean different things to different people. Just run me through what you consider to be the like how you delimit the scope of the information that you have, like geographically and in terms of company size. It's a speaker-1: It's a question that we get often, you know, is our data set representative of the typical firm in the United States? And it's it's not. You know, we don't the you know, the businesses that tend to use our platform tend to be quite tech forward, no matter how you cut it. You know, not not to say that they're always in the tech sector. Actually, most businesses that use ramp are not in the tech sector, but it's that it ramp is itself kind of an AI finance platform such that, you know, the manufacturing and construction and consumer Firms that use our platform also themselves tend to be quite tech forward and likely to adopt new technologies. Now that is actually somewhat of an advantage for the research that we are trying to produce. AI itself is this technology that is going to be invested in by tech forward firms. So it benefits us to have access to this data set that looks at what they're spending money on specifically. Now, in our work, we will control our analyses. ⁓ You know, when it makes sense to show our results relative to the typical firm. ⁓ when we ⁓ we recently wrote this paper that ⁓ examining AI's impact on the job market. So we use data from RAMP ⁓ to study to group out firms that use AI versus firms that don't. And then within those firms that use AI, which ones ⁓ which ones are using AI particularly intensely versus, you know, in a very sort of light way, maybe like chat GPT subscriptions for everybody. And then we got workforce data from a company called Ravellio Labs, and we joined these two data sets. So we could see now these firms use AI, these firms didn't, what happened to jobs over time. And it's there that we got these very compelling results that showed, you know, despite a lot of the talk going around about tech layoffs and ⁓ you know, people are being fired because of AI. Firms that use AI heavily are actually growing faster. They hire 10% more over the two years following AI adoption. But it's Concentrated. It's essentially only happening amongst the firms that are using AI intensely. So, you know, what that means is that most firms that are using AI today aren't really seeing all that many benefits at all. Because for most firms using AI, they're not using image and video generation or AI for sales and customer service. They're not using the coding agents. The vast majority of firms that you're using AI are still on. fairly simple ChatGPT or Claude subscriptions that are given to all the workers of the company, but are not necessarily connected to all the tools. And if you've used those very basic forms of technology, you can see it's kind of nice to have. But it's very hard to see how that might be so productivity enhancing that it's going to vault the company into the next generation of productivity growth. And so a lot of what we study is, you know, what's it going to take to get more firms Not just on an AI adoption curve, but on a high intensity adoption curve. Now, the rest of our data set, what I'm actually very excited for is that RAMP has now launched this summer it launched internationally, at least in the UK and EU. And so speaker-0: Chris, we are Scientist. You're available in the UK. We ⁓ we don't we don't use Ramp, but we the we we are paid by ⁓ speaker-1: In yeah, in ⁓ in ⁓ a couple of years, I'm hoping we'll get ⁓ a long enough period of time in our data set that we can start reporting results that compare the performance of firms in the US versus firms in the EU. and ⁓ particularly in terms of AI adoption, its impact on the labor markets. speaker-0: Perfect. And we we see the same thing with our survey, to be fair. We've always we've always said there's a degree of pre selection that happens whenever we talk to the market or have the market talk to us, which is people who read the interline, a tech publication profession, are already people who are fairly deeply into this space. ⁓ so we we recognize but we also see it the same way as you, which is you're better off looking at the the high intensity folks and the more committed companies than you are at the people who are just kind of coasting. ⁓ speaker-1: You might speaker-0: What do you d just to kind of go back to the vendor lock in side of things, just very, very briefly again for a second. Do you see consolidation coming in the market? Because it's a strange thing out. So you basically have two major suppliers, I would say, for most companies. ⁓ Google is a weird one because technically speaking we do have we do pay Google for AI, as I'm sure a lot of companies do, but it's baked into workspace and we would not be paying it separately. You just you you can't not. So ⁓ so that's so it's in it's in there for that reason. But ⁓ on a selective basis, you basically have open AI and anthropic. I ⁓ you hinted before at the idea that ⁓ there's you know, you think about Chinese models, think about other companies, there's maybe an expansion of the market coming. And then is there a contraction? Like ⁓ do do you think there's a long term market where there's a wider cohort of AI companies that can actually survive out there? Or do you think this is destined to be a two horse race for the foreseeable future? speaker-1: I I I don't think it's gonna be a two-horse race. I think there's gonna be many more competitors. And that's because I think that AI will the firms that use AI most effectively, the the strategy in terms of buying AI that's going to win out is going to be the model agnostic strategy. That it behooves you as a firm to both experiment with many models to figure out which one works for what task, but also it behooves you as that firm because. ⁓ you're going to want to use the most performant model for a certain task at the best price. And there are going to, there are already some technologies entering the market that allow you to, open router is actually one of them, right? That allow you to ⁓ automatically and intelligently direct tasks to the model that makes sense. And this is a really hard market to keep track of, too. I mean, if you're using OpenAI and Anthropic yourself. You know, you're just using codecs or claud code. Let's remember that ⁓ they don't make it particularly easy for you to get the highest ROI from your model usage. And they have no incentive to. You know, they make the most money when you use the most tokens. And so if you're using OpenIR Anthropic for your projects, you know it you you kind of have to be extremely informed on AI models as they develop and on the market so that you can make sure that, like, ⁓ do I need to pay the most expensive model for this task or can I send it to the cheaper one? And that's just not that's I don't think that's gonna be natural for most people to do. I don't think most people are capable of doing that in such a fast moving market. And I don't think Open and Anthropic today have the incentive to create a product themselves that will intelligently route you to the most appropriate model at the most appropriate time. Well, there are going to be increasing numbers of companies that develop that kind of technology that businesses will buy that route you and your tasks to the most appropriate model without you having to do anything. And that's going to support a future that is much more model agnostic, for which there are many more models besides open and anthropic, and that may include open source models too. And by the way, other American model companies, Google. speaker-0: Mm-hmm. ⁓ and and they do NVIDIA as well. You get this there's there's there's a lot of companies out there that that have their own frontier LLMs. But then when you widen the net, as you said, there's a lot of companies out there that have smaller specialized models for vectorization and embedding and things like that and text speech and what have you. speaker-1: And then think about the international complications of this. You know, if you're a company based outside the US, we just saw that the United States has the ability, this is with the Fable ban, the Anthropics model ban, has the ability to impose export controls that within minutes cut off your company's access to a certain AI model simply because you may not be based as a company based in the United States, even if you are otherwise a US ally. At least you're based in a US ally country. Now, if you are then therefore a European company that is building using AI, you might be very hesitant to adopt OpenI and Anthropic and rely your and and build your business on their tools. Because you might have the technology pulled out from under you without any notice. So what do you do if you're one of those companies? You probably are going to explore ⁓ providers outside certainly providers outside of OpenIn Anthropic, if not outside the US entirely. ⁓ and ⁓ and and distribute your your efforts across many firms so that you are not tied to one. speaker-0: So if you think about that practically speaking, then so the the current the contemporary example of that would be the OpenAI ⁓ unreleased model. ⁓ I was gonna use the word hack, it's the wrong word. The the open AI unreleased model ⁓ finding its way into production databases at Hugging Face. Hugging Face not being able to use US developed closed weights models because of guardrails and having to turn to ⁓ open weights and Chinese models in order to to foil that. ⁓ It feels like we're barreling towards a bit of like a hyper balkanized future of AI, what you're what you're describing there and what what they're looking at. ⁓ what should have d just just like a rational actor, just a regular company in in this sort of space, what practically can they be doing to do that kind of risk spreading and diversification? 'Cause we've we've said there are alternate platforms as open router and so on. ⁓ from a pure like what do we what do they install on their employees' computers? Like what is it that actually gives them a way to spread that risk. speaker-1: Well, you know, it's hard to say today because there aren't great options, at least on the frontier, that don't rely on open an anthropic. So I think today, I wouldn't tell a firm not to use open an anthropic. You should be using open anthropic, you should be using the frontier models, you should have your employees experiment as much as they can ⁓ with the best models that are available. and also if you are in charge of procurement. particularly for a company outside the United States, you should be looking to implement these kinds of platforms that would set you up to be model agnostic fairly quickly. And so that's going to be ⁓ you know, platforms like Open Router today. speaker-0: And if we think about the Chinese model side of things then, so Moonshot's Kimi K three is the is the current ⁓ the kind of the Chinese model du jour as it was, ⁓ and is the one that seems to be getting a lot of attention. Now, I feel like we've done this dance before. If we if everybody thinks back to the initial ⁓ Deep Sea Car One was it release where all of a sudden ⁓ there was a big ⁓ dropout in the bottom of speaker-1: ⁓ speaker-0: tech stocks and things and everybody's saying, well, okay, this this fully democratizes the capabilities. It's ⁓ it's the end of the business model for open AI and Anthropic. That did not pan out with that initial deep seat release. Do you think this this turn is different now? Do you what is it that you think that American companies, international companies can get from models developed overseas that make them more attractive? And do you think it's actually going to move the needle this time? speaker-1: Well, it's similar in that like last year when Deep Seek had its first sort of viral hype moment, ⁓ it exposed the main feature of open anthropics dominance in the United States that isn't working for businesses, in that they are themselves not incentivized to build highly performant models that are also cost effective. And they have to date not built A product infrastructure that allows firms and that allows their customers to use their technologies more cost-effectively. And so this latest, you know, the last time, last year when it happened, and DeepSeq had this big hype moment, you know, even though the vast majority of firms weren't switching off OpenIn Anthropic, it did spook them. And they responded by putting out some cheaper, highly performant models. And those ended up being significantly better than what was otherwise out there from DeepSeek. And so You know, we didn't hear about Deep Seek for a while. Now it happened again where not just DeepSeek, there are many more Chinese models that are highly performant. and you know, I I wanna say that it's as easy as the American model companies enacting some new price cuts and releasing some cheaper models. But it's not it that's gonna require much more than cheaper models. It is going to require a a rethink. From OpenIon Anthropic, assuming that they are affected enough competitively to respond in this way, to start revising to start offering products that allow firms to better measure the ROI and monitor their spend. Today that is not offered. I mean, the fact that there's no automatic model switching offered in Codex or Cloud Code, ⁓ you know, in any in any sort of in in the sort of way that that companies are asking for just does not exist in any sophisticated way. And so You know, there there's changes around the product that you know businesses are clearly asking for, but that open I and I open I and Anthropic are not able to meet. Now, I think the ⁓ the mistake would be thinking that the threat is only gonna come from Chinese model companies. It's not. It's also gonna come from Google. It's also gonna come from ⁓ other AI companies that are ⁓ starting up in the US that are trying to offer, you know, alternatives to To the two top auto companies. speaker-0: Okay. Now one of the statistics from the AI index that we've used a couple of times and that I personally find the most interesting is the the token spend per employee. ⁓ and I forgive me, I haven't looked at it in if it's been updated in the last week or so, but the last time I did, I think the upper bound for what companies would spend on an individual employee on a on a token basis was something like seven and a half thousand dollars per employee per month. That's the absolute like high end. The top ten percent I think was a bit more of a reasonable kind of six hundred dollars a person. ⁓ but the median firm, like your average company, which which we've been talking about, spends about eleven dollars per employee on ⁓ on the token side of things. I know regional pricing is different, but that's that's about or slightly less than a single chat bot's cheapest tier here in the UK. So that's two two cups of coffee, it's like a six hundredfold difference. How do we look at this? Do we think about this in the friendly way, which is that companies spending seven thousand dollars a month per employee can't possibly be getting that much value and that they're just like lighting cash on fire. Or do we look at it the other way and say the companies that are only spending eleven dollars per employee per month just giving somebody deep seat to Chat GPT or whatever are are missing the boat. ⁓ and that the market is gonna outpace them. 'Cause you know, I look at this from our experience. It's very easy to burn a hundred pounds, a hundred dollars on automating some admin and then looking back and saying, Did I speaker-1: It's speaker-0: Is that necessary? Did I need to do that? ⁓ but then you also look at, you know, world's biggest retailers, or Europe's biggest retailers, Orlando, ⁓ going all in on image generation as the future of content creation. That is not cheap. It's potentially cheaper than traditional photography. It's not cheap. so I'm left a bit conflicted over do ⁓ are the biggest companies just trying things here, or do they know something that the median market does not, which is that you need to become that high intensity, high frequency company to really start to get the value and get ahead? speaker-1: So I I mean, I think about it from my perspective as a researcher, which is, you know, when we were considering this metric to track spend per employee, you know, we kind of had to break it out when we looked at the dis when we looked at the distribu distribution of spend. Because, you know, although most of our metrics are reported at the median, you start to see how it falls apart for AI, ⁓ because there's such large differences between firms in the top one percent and even the top ten percent versus the median firm. So our goal with this kind of research is to not not not not even suggest, hey, here's how much you should be spending on AI, but at least give some benchmark for the wide range of of spend a firm can have. ⁓ I mean, look, it's all gonna come down to the actual ROI for a firm. Now we we went through a couple of different phases in in the tech sector as far as what good AI adoption looks like. You know, I think actually One of the unfortunate phases that we're still in is that firms still do not know how to talk about the value of AI. N not that they don't know even know how to talk about the value. They're really not sure how to measure it within the firm itself. That's why we end up seeing all of these different metrics thrown around, firms measuring themselves by how much they spend on AI, measuring it by the percent of workers at their firm that use AI. That's not the worst one. You know, I I've seen similarly like the share of code at our company that is written by AI. And none of these metrics give you any sense of the quality of work that is being done, or whether or not the work is being done in a cost-effective manner, or whether or not you know whether or not any of it was worth it. So I will say the kinds of metrics that I've heard that are quite impressive as far as measuring AI's impact, and I think are Are generally the way most firms should go, or the ones that are tied to outcomes that are felt by customers. So I know I talked to a tech CFO, but like kind of involved in the consumer space, who set an actual goal, minimum threshold for the share of customer service inquiries that need to be addressed by AI completely with no human touch by the end of the year. So that I thought was great. You know, you're not saying, ⁓ I need we need this many customer service agents using AI, but you're saying that, hey, we need, first of all, we need to have a certain level of customer service operations that can just be done by AI completely, because that's going allow our human agents to then focus on more inquiries. And then you can think about the other metrics that you would define. You know, maybe even for your human agents, they can use AI to complete customer service calls faster because they can look up information and answers. So you can define a benchmark based on that too. Like the the average amount of time it takes for a human agent to close a call. See if that comes down over time as well as AI adoption proliferates through a firm. These are much better metrics that are actually tied to the impact AI can have on a firm and it can have on your customers that that I think resonate more with the market and resonate more with customers, especially, rather than us hearing all the time about the promise AI can have because ⁓ more code is written. Or we have more data center jobs available. speaker-0: They are harder metrics, but they're also metrics that bring me to my kind of ⁓ very near final question, which so you mentioned earlier, the kinds of companies that you capture in your data are fast growing and they are also growing in terms of headcount despite being ⁓ high intensity adopters of AI. If I think about call centers, the which is the you know, the center you've just been talking about there, if if that company succeeds in that metric by the end of the year and has a fixed amount, But maybe it's a majority, maybe it's not. A fixed amount of its customer service interactions handled fully, autonomously, without human touch, that has to lead to net job loss in that segment at the very least. Because presumably those queries are handled by people today and they will not be handled by people in the future. And I think the thing for me that I keep coming back to is fashion by and large is not made up of fast-growing VC backed. well capitalized companies. It's made up of companies that trade on pretty thin margins. Companies that are in the SME market in particular, a collection, a bad collection or two away from going out of business. These are the kinds of companies that if they scale their AI spending in the way that we're talking about here and they they they become these high intensity folks and they set these kind of concrete metrics. That spend comes from somewhere else. ⁓ It's it it there's no kind of magic money tree to fund these things. Do you think there's maybe a different world that you're looking at with these fast growth, well capitalized companies versus your typical kind of brand or retailer when it comes to the outlook of the relationship between AI and jobs? speaker-1: AI is going to affect my job. It's going to affect other sectors too, and mine, yours, and and sectors that we don't even know to consider yet as far as AI's impact on them. I I definitely fall on the side that people are overestimating AI's impact on the job market as far as job loss. First, because we haven't actually seen any broad job loss from AI at all. And so these predictions. Are not really based on anything quantitative or observable. And the second is that I really am not compelled by any of these suggestions that AI is going to develop any differently from any other technology that has also automated work in the past. Now, AI will definitely lead to some industries declining in jobs. We may not know to what extent, and we may not know what those industries are. ⁓ But I have my instinct is it's it's probably going to. happen in industries that we aren't thinking about and we don't even expect. There tends to be, for example, a lot of focus on white-collar work and how AI is going to automate the work of software engineers and lawyers and accountants, and therefore we'll have we'll have fewer lawyers and accountants than software engineers. Now, we have the result from our data set that shows actually there's a lot of job growth in white-collar work, especially at firms that adopt AI intensely. But then you can think of the similar impacts that other technological innovations have had in the decades past on those sectors and similarly seen AI, computers in general made the work of a lawyer much easier and much faster. You know, go to a law firm. They historically have had inside the actual office like libraries that ⁓ lawyers had to go to physically to open books, to find case law, to refer to them in their legal briefs. And now we have computers and web searches that allow you to build those briefs theoretically much faster. think about the hours and hours of time that are saved per day by not having to look up case law in a book. Any does anyone there's no suggestion out there that lawyers suddenly because of the advent of computers in Alexis Nexus were able to start going home at 4 p.m every day, right? Because their competitors also had access to that same technology. Very true. And therefore they also started building their reports much faster. And therefore they found ways to make their legal briefs even stronger or better and do other human oriented work that technology could not do. And that's why they all still go home at like midnight. speaker-0: So that's also why I'm not a lawyer, to be clear. ⁓ and I think the other thing as well. So the putting a computer in a legal office, okay, fine. So you replaced the job of the archivist or the librarian or the philologist or what have you. and you you created an army of paralegals and other things that that that were that had the computer as an axiom that that you needed to that technology revolution to create those other jobs. I think if I frame it through a fashion lens, a lot a lot of focus is on the creative work. So particularly ⁓ like campaign photography and things like that, where up until now you needed a photographer, you needed a stylist, you needed a makeup artist, you needed somebody on lighting and so on, ⁓ to put together a campaign. Shoot, now you don't need any of those people. You can do it all with AI. those jobs initially were also created by the advent of photography. They were particularly created by the advent of digital photography, which made all of this ⁓ scalable and and so on. So I see where you're coming from. I think it's like it it it does impact jobs, but I I I buy the idea that a lot of those jobs were created by technology in the first place. ⁓ and I think it's maybe a natural evolution from that point of view. Let's end 'cause we are over time on a prediction. If I asked you to guess what you would see in the AI index if we rated this interview a year from now, what would your best guess be? Do you think AI, at least as a paid product, is still only going to be serving just over half of US companies? Do you think those two American labs are still going to dominate? Do you think that medium brand we've talked about is going to pick up and become more high intensity? Do you expect to see spend per employee go up? Just give me your potted overview of what you see happening. speaker-1: I think we're definitely going to see AI adoption rise to virtually 100% of firms within the next five to ten years. Probably closer to the five side. ⁓ the average firm will become more intense in its AI adoption. I also think, unfortunately, firms that are less inclined to adopt AI at all will probably suffer from market competition that is able to grow faster with AI. And those firms may end up going out of business. And that's going to happen, not to be a doomsday prediction, but that's just generally what technological innovation does, even in sectors where you wouldn't expect there to be massive productivity gains for that firm. I mean just as an example, we can go we can think about doctors' offices again. You know, computers and like, you know, didn't necessarily change the job of a doctor all that much. But every doctor that you go to has access to a computer. And every doctor that you go to probably has a Google Maps listing and has maybe done a little bit of digital advertising here and there. You know, the firms that never adopted even the sort of basic technology of a computer or a Google Maps listing, those are probably the firms that went out of business, even though the technology itself was not that beneficial to the growth of the industry itself. So we're gonna see something very similar with AI, that AI is enabling some firms to grow faster, even if it is not enabling them to produce more themselves. You know, it just ends up being this growth engine in ways that we do not expect. speaker-0: What ways do you expect the high intensity firms to change in that five year period? Like what what does a high intensity firm and I know this is difficult to say across business models and things, but if you stay at that frontier as a company and you commit to this, what do you see being the benefits that you look back on in five years time and saying, that was advantageous for us and that has changed the type of business that we are? speaker-1: Well, I think a lot of the growth is gonna come from adoption outside of the main models. ⁓ I think our research, if if there's one headline point I I would point to in our research over the last two years, it's that ⁓ you know, you don't really see that many productivity gains from AI as we know it, as most consumers at least understand it. You know, and that the free subscrip free or relatively cheap subscriptions to Chat GPT and Claude. Are not particularly productivity enhancing. And some of the most exciting areas of the development for AI are happening outside of the hype cycles. I use customer service operations as an example often because it may be a little bit dry as far as ways in which AI can improve our lives and improve a company's operations. But it is one of the most commercially promising versions of AI, and one that I think many people would be excited about as far as a change that would affect their lives. you know, helping customer service operations at a company move faster, helping the job of customer service agents close out more inquiries and help more customers and ultimately leading to a better experience for more customers. I like to focus on versions of AI where you can clearly see the ROI ⁓ you know, both for the company and for the customers that it serves. So I look for examples of the technology that are ⁓ able to prove their value there. speaker-0: Good. Well, I think everybody is in the same boat. ⁓ that's probably the focus of our AI report that should be out around the time people listen to this as well. Aura Coresian, I've I've taken up a little bit more of your time than I intended, but this has been a great conversation. I appreciate you joining. And that's the end of my conversation with Ara. I got a lot out of this one, and I hope that came across. Now, I do tend to agree with his projection there that we'll see total AI adoption within five years or so. ⁓ you can infer that from basic chart reading in the sense that speaker-1: Thank you so much for having me. speaker-0: If we got to fifty five percent uptake in three years, another five should really capture the rest. What I still think isn't clear is what the adoption is going to look like, especially in fashion. Our AI survey this year focused very much on AI maturity by sector and on how people feel about the technology. I'm starting to think we should run a separate vendor focused data collection exercise as well. So we can understand what capabilities people are spending on and which providers they're using for it. Because Frontier LLM spending. Is obviously very concentrated on ChatGPT and Claude, but I think the picture would be far more blended if we did a sector-specific slice for fashion. I also want to emphasize that just because AI has the potential to create jobs that didn't exist pre-AI, like R and I just talked about, there's no guarantee that those are net new jobs or that they're going to be applicable or available to the people who work in traditional fashion disciplines today. It's not always the case that you can upskill to be ready to fill. A role that replaces your current role. What is obvious though is that AI adoption across every sector and in fashion specifically has a lot more twists and turns to make over the next few years. And I'd really like to have our back on, I think at some point in the future, so we can compare benchmarks again across the broad suite of data and across our fashion specific one. For now, thanks for listening. If you're a fan of these weekly interview shows and you haven't checked out our other show yet, The Edit, ⁓ Go do that because that's where Grace and I spend twenty five minutes or less each week talking about the news headlines. It's a breezier, more accessible, more fun format designed to fit into your commute, but without sacrificing the expertise that the Interline is built on. Whether you end up listening to both shows or not, I'll talk to you again really soon.