UBS AI Podcast - CEO Series - Ep. 1 (Arvind Jain, CEO of Glean)
The desk believes that advancements in artificial intelligence, particularly driven by large language models, will significantly impact workplace productivity and reshape industries. Per the full note source, the conversation with Glean's CEO, Arvind Jain, highlights the transformative potential of AI in enterprise settings, underscoring how Glean's solutions facilitate efficient information retrieval. This echoes broader themes of digital transformation observed across various sectoral analyses and positions AI as a critical catalyst for productivity gains moving forward.
What the desk is arguing
The desk posits that AI's integration into business processes represents a seismic shift that could alter competitive dynamics within multiple sectors. Specifically, tools like Glean, which streamline knowledge search in the workplace, signify a pivotal shift akin to past technology disruptions. This perspective is reinforced by Jain's insights into how AI can redefine knowledge management within organizations.
As Glean leverages transformer-based models, the desk notes the growing acceptance and implementation of AI-driven capabilities across industries. This shift is evident in increasing investments and strategic prioritization around AI technologies, suggesting a momentum that could drive market valuations and operational efficiencies to new heights.
Where it sits in our coverage
We currently track a consensus target for the EUR/USD at 1.075, with a range of 1.04 on the lower end and 1.12 on the upper end. Specific targets from notable firms include: - jpmorgan: 1.10 (Mar26) - bofa: 1.04 (Mar26)
Considering this AI-driven thesis, the desk believes we are positioned at the higher end of the spectrum, potentially forecasting further upside as AI adoption accelerates.
How other firms see it
Firms such as jpmorgan and citi align with this forward-looking view, emphasizing the impact of AI on efficiency and productivity. In contrast, bofa remains skeptical, focusing on the risks and limitations associated with rapid AI deployment.
Key related currencies to monitor include EUR/USD given the broader implications of the Eurozone's digital transformation agenda and USD/JPY where technological advancements are also expected to play a significant role in shaping economic evaluations.
What the calendar says
There are currently no high-impact events scheduled for the next 30 days that could influence this discussion, allowing for a focus on ongoing sectoral trends without immediate catalyst distractions.
How firms align with this view
Aligned with the desk view
Contrary positioning
Key takeaways
- 01AI's transformative role is expected to enhance workplace productivity significantly.
- 02Glean, leveraging AI, exemplifies effective knowledge management solutions.
- 03A growing momentum in AI adoption could merit bullish positioning in related currencies.
- 04Firms are divided on the implications of AI, with varying forecasts reflecting differing confidence.
Market implications
Watch for the USD to strengthen against the Euro if increased productivity through AI leads to better-than-expected economic outputs. Any fluctuations around the EUR/USD level of 1.075 could signal market shifts in response to evolving sentiment on AI adoption.
Risks to this view
A slowdown in AI adoption or regulatory pushback against tech implementations could derail the positive market outlook. Additionally, broader economic indicators failing to show improvements driven by AI integration may necessitate a reassessment of bullish positions.
Welcome to Layer by Layer, our UBS AI podcast, where we explore the fascinating and evolving world of artificial intelligence. How it's transforming industries, redefining what's possible and reshaping the future. Join us as we deep dive into the minds of visionary creators and thought leaders at the cutting edge of AI.
I'm Ulrike Hoffmann-Bojari, CIO for Global Equities at UBS. Today I'm thrilled to be joined by Arvind Jain, CEO and co-founder of Glean. As the leader of Glean, Arvind is transforming the way we search, connect and work.
Arvind brings a wealth of experience from his time at Google, where he made groundbreaking innovations in search and infrastructure. And as the co-founder of Rubrik, a pioneer in data management and cloud solutions. Today we'll dive into the vision behind Glean, Arvind's thought on the future of AI and workplace productivity.
Arvind, thank you so much for joining us on our first episode of Layer by Layer. Let's get started. To kick it off, what problem does Glean solve?
Glean makes it easy for people to find information at work. Whenever you have any questions that you need answers for, and the answers are somewhere in your company's data and knowledge, but it's hard for you to find where exactly it is. That's the thing that Glean makes it easy.
You can think of us like Google or chat GPT inside your enterprise, one place where you go and ask any questions and we'll find the right information and bring the right answers to you very quickly. What I find interesting is that you founded Glean exactly two years after the Transformer model was published in 2017. So to what extent were large language models the so-called prompt for what you wanted to build?
The language models actually played a big role in what motivated us to start Glean. In 2018, when we started to think about this idea, we saw the real transformative power of these Transformer-based models in the search teams at Google, which is where I and most of our initial team came from. We were seeing how good these models were at understanding information.
In 2018, Google actually put some of these models straight on the entire web in open domain for companies like ours to use. That was one of the key catalysts for us starting Glean because we felt that using these models, we could very deeply understand enterprise information and data and build a really good search engine using that. It seems like you used Gen AI and the tools that Google offered at the time to reinvent enterprise search.
Is that the right way to think about what the value proposition is with which you started Glean? Yes. You can think of it that way.
Search as a problem hasn't been solved in enterprise. When we started, we had never seen any good search product in the enterprise, and there were quite a few reasons for why it was so hard to build search over your enterprise information. One of the things was made easier by these new Transformer-based models because they allowed you to understand content more semantically.
But there were a few other factors which were also important for us when we started, especially the SaaS transformation that the industry went through that not only made the problem more acute for our customers, but it also actually made it easy for us to tap into all of that data and knowledge that sits within those individual SaaS products and made it easier for us to build a turnkey search experience. But today, you can think of Glean exactly a modern search and AI platform where we actually go very deep into your enterprise, build this knowledge graph, which essentially is an understanding of how people work together inside an enterprise and how they work with the knowledge inside your enterprise, and then combining it with a RAG-style architecture and generation capabilities of language models as well as reasoning capabilities of these models, you're able to actually go and create an experience which goes well beyond the traditional search where you will type in a few keywords and we will surface some information, like some documents back to you. Today, you can go much further where you can converse with the system, you can ask questions in natural language, and we will not only find the right documents that actually will answer those questions, but then we'll have AI work on it and actually generate concise answers to those questions that you have.
So it's a much more powerful and modern experience of search inside the enterprise. So you mentioned a few portions of the stack where you use large language models, it seems, building a knowledge graph and RAG, and then also the conversational interface that helps with answering a query. If you maybe step back and say, at what areas of the enterprise stack that you built, do you use large language models and what capabilities, retrieval and also generation?
So first thing, when you think about enterprise knowledge, every company has their unique sort of jargon, their language. You have code words, there are a lot of things just specific to how your company works. So one of the things that language models help a lot with is when you train models on your enterprise data, it starts to sort of understand your speak, your language, it starts to figure out what are the key concepts, projects, entities within your business, and that understanding allows us to then match content at a conceptual and semantic level to questions that you have.
So that's one area where this language model technology comes out to be very helpful, is in building a strong retrieval system, which is more semantic, more conceptual. And then the other thing that the models do a really good job at is when you have complex questions, they have the ability to read large amounts of knowledge and actually reason through that knowledge and figure out how to answer those questions for you. So this sort of enhances the end user experience for our users, because when they have questions sometimes, they don't want to go and read those long documents, and they want a quick, short answer.
And the models have a really good capability of generating and synthesizing those responses. So that's just a simple thing, like when you're actually doing question answering using Glean. But Glean actually goes well beyond those capabilities now.
You can actually ask Glean to do work for you. You can ask it to analyze content. You can ask it to generate reports.
And all of those capabilities also, again, is where we are putting together your enterprise knowledge along with the capabilities of the models. So what we're doing is, if you want to analyze some data, first our retrieval stack will actually figure out where that data is. And then you can actually use the model to actually write code on the fly and do complex analysis on that data.
So ultimately, these models allow you to really extend the capabilities of what a user can do with the product. It goes beyond question answering to doing actual work where AI is helping you do a lot of your work for you now. So Arun, it seems like you're moving from enterprise search into creating a broader workplace AI platform.
And as you're moving, as you're transitioning towards creating these autonomous agents, is it right to think that you're sort of moving for more of the retrieval aspects of large language models and their capability to also leverage the generative abilities and especially of course, agentic reasoning, which seems to be talk of town these days when you think about abilities of large language models. That's correct. In fact, you can think of Glean in many ways like, you know, it's like chat GPT.
You come in, you ask it to do some work. And depending on like what kind of work you want Glean to do, it's going to actually like first, of course, like, you know, tap into all of your internal private company's data and knowledge in a safe and secure way, but also tap into the world's knowledge, like if you need to do that and then and then it goes and starts to solve those tasks for you. Sometimes, you know, it's question answering, sometimes it's, you know, taking a complex task and figuring out how it's going to actually go and solve that task in a more agentic fashion where you will take a complex task, you'll figure out like, you know, how to break it down into a series of steps, again, using these, you know, language model technology.
And then you'll execute each one of those tasks, you'll actually bring some sort of self, you know, reasoning and an auto correction as you sort of network through that task. So, so absolutely like, you know, the product has evolved from, you know, again, that core capabilities of helping you finding information to bringing you answers to actually doing work for you. And how quickly do you see we will see these autonomous agents?
I'm sure for some smaller tasks that's already happening, is this what will be, again, the main focus of investments and R&D going forward? Yes. So I think we will see a lot of progress on building, you know, AI agents that are, you know, that are going to automate a lot of our business processes, although in terms of like, you know, how these agents are built, there are two different concepts, you know, one where, you know, as a, let's say that I'm a business process owner and I want to automate, you know, a complex, you know, business process that I have.
And so I can work with AI and I can sort of describe to AI that, hey, this is how I should work on this task. These are all the different things I do. And you can sort of put in that investment and help the AI, you know, system actually build that agent based on that guidance that you're giving it.
And you sort of collaborate with it in terms of, you know, getting that agent built. And then once you have this agent built, you know, then it can actually run that business process autonomously, you know, like, you know, like every week, for example, if you know that's how you were off and, you know, that work had to be done. And so that's sort of one approach.
And we think, you know, that's the right approach, like right now that actually works, you know, well, you know, as long as human is a human in the loop that actually, you know, puts, you know, like, you know, full effort into building that agent, like, you know, without, of course, writing any lines of code. And then from there, like, you know, the other the other extreme is where you just have a task and you don't help AI in any way and you make AI sort of learn and figure out automatically everything on their on its own. And there we see like a lot of challenges today.
Like, you know, we have not seen good results where you give open ended tasks to AI and hope that, you know, like, you know, that you're able to sort of autonomously create an agent. So agent creation, like, you know, we feel like it's good to have humans in the loop for that. And then once agents are built, I think they can, you can like for a lot of use cases, you can let them operate completely independently.
So we have talked now conceptually quite a bit, but obviously you have a lot of fans in your user base, not least the CEO of the largest U.S. company, AI company or U.S. company, Jensen Huang. So I'm curious, what is the ROI that your customers are seeing by using Glean and what are some of the tangible outcomes your customers have achieved with Glean? In terms of ROI, so number one, you know, AI is all about efficiency and we have done like a lot of surveys, our customers have done a lot of surveys, you know, that show that, you know, we typically save about two to three hours of time every week across, you know, your entire employee base.
That's one of the things. So it's time savings because a lot of like a lot of times you're looking for things, a lot of time you're trying to actually get, you know, quick answers to your questions or trying to actually find like, you know, who you can connect with in the enterprise, you know, who can help you on the task that you're on. And all of those things become much faster with Glean.
So there's a sense of like, you know, significant time savings and like for specific teams like, you know, engineering or, you know, support, like, you know, those time savings are actually like even much greater than that. So that's one thing. So there is, you know, these time savings, you know, which ultimately some companies, you know, feel comfortable like, you know, using that as a metric to figure out like what's written on investment of Glean for them.
But I think but time saving sometimes, you know, can be a soft metric because like you don't know what happened, you know, with the time that you saved. So you have to look, you know, look at the next level of metrics. And so but like function by function, that's what like, you know, our customers do today.
They look at like the hard metrics, business metrics that they have for that function and see how we're impacting them. So as an example, you know, one of the largest telcos in the US, they have more than 50,000 customer care agents using Glean every day to help, you know, with answering questions, you know, that their customers have like, you know, help resolve tickets. And Glean has shown 47 percent improvement in that case resolution time across teams there.
And that's like, you know, that then directly translates into millions of dollars of like, you know, hard savings because like, you know, when you think about large customer care organization, if you're able to, you know, resolve twice as many cases, you know, that directly impacts like, you know, what your cost is going to be because you can reduce your overall cost of, you know, you know, you know, or your headcount plans, you know, by by a half. Right. So so you can like, you know, function by function, like for engineering, you can see like, you know, are you running, writing more lines of code or, you know, more and more commits in that are your teams are uploading.
So that, you know, for when you go into specific functions, you can look at those existing metrics and see the movement. And and our customers are seeing, you know, pretty significant movement, you know, on those metrics. Same for sales, like you can actually prospect more, you can actually be more effective, close more deals.
So, yeah, there are all these metrics, you know, across different functions. I know that we've seen our customers see real value from Glean, and that's why they're deploying the product company wide. It has gone well beyond, you know, the POC stage.
You mentioned these different sources of productivity improvements. So when you look at one of the superpowers of Gen AI, it's certainly writing code. And there, there's a lot of evidence data that shows that we see that productivity improvements of 56 percent.
I think this is one of the numbers that is mentioned in a recent study. Did you think that is going to also be true for other use cases that we see really what is an unprecedented level of productivity enhancement going forward? I think we're going to see more and more, like, great, like an impact from AI on our day to day work.
Productivity increases like today, like in pockets, you can have that level of improvements. But overall, like, you know, we're in very, very early stages of the journey of how much impact AI is having in the enterprise. But if you go like this, look back, if you sort of think about what's going to happen in the next five years, I believe that the way we work is going to fundamentally change.
In fact, I feel that like the majority of our work in the future, you know, is going to be done by, done by these AI agents and assistants that every employee will have, you know, personal, personal agents and assistants, you know, that, you know, that you have, it doesn't matter, like, you know, who you are, like in terms of like, you know, today, like, you know, you have the senior executives in a company have some level of assistance available to them in forms of like, you know, EAs or chief of staff. But in the future, like everybody who works will have amazing assistance available, you know, through this, you know, concept of an AI team that you have as an individual. And I think most of your work is going to be done by, you know, these AI agents and you'll be able to up-level, you'll be able to, you know, like, you know, be, you know, twice as more efficient or even 10x as more efficient than what you are today.
Most of the work that you're doing today, you won't be any more like in five years from now. And I'm just curious, how do you use AI internally? Do you use a lot of these agents already inside Glean?
Yeah, so we actually, first of all, of course, we use Glean very heavily inside Glean. Everybody uses it to search for information, ask questions, get like, you know, basic tasks completed, you know, within Glean. We also use some other third-party AI tools, like specifically our engineering teams, you know, they use GitHub Copilot, like for coding assistance.
And then, like, you know, our team, every department within Glean has built, you know, specific prompts and applications for their own department on top of the Glean AI platform. For example, I'll give you some examples. For our sales team, you know, there's one of our, in fact, there's, you know, one of our sellers, like he's not a programmer, he went and built a customer 360 application on Glean.
It's, you know, like, and I can just, you know, like you go to that application, you give it a name of a customer and you get this really comprehensive view on that customer, including like, you know, what opportunities are open, who's working on that, any issues, you know, that this customer has reported, what are some of the recent discussions where you get this, you know, very comprehensive, you know, view of like, you know, the state of what that customer is. There's, you know, there's another application which helps you write a really nice follow-up email post, you know, having a customer meeting. The engineering teams, you know, they have built interesting applications.
Like one of the really cool ones is that every time somebody uploads any new code into our code repo, we have an AI agent built on top of the Glean platform that actually does the first review of that code and finds mistakes and errors and issues in it. And that way, like, you know, we're ensuring that, you know, we're uploading, you know, like our code quality remains high. So, yes, it's actually very interesting that once you start to actually, you know, give this capability of like building applications to people who don't have to actually code and deal with like, you know, clunky sort of system integrations, it's like, you know, you can, like, you know, it sort of really accelerates, you know, like how many things you can automate in the business with it.
So, but I will also note one more thing that while, like, you know, we are having good success with people building a lot of applications and workflows that they're automating with AI, it takes a lot of effort to actually get that going inside a company. You have to have evangelists, you have to have a program, you have to push, like, you know, one of the things that I did, even though we are an AI native company, you know, initially, like, you know, we were not, like besides using Glean, we were not actually being very active in terms of automating a lot of our business processes. So we had to do a push to set quarterly goals, you know, for our executive team to actually, you know, get that process going.
But once it got going, we started to see really good results from it. Yeah, that's a really good point. You need not only the technology, but also the technology mindset.
We do believe that as well, that that's going to be the differentiator, the combination of the two. And one question I had for you is just on mode, given that AI is so capable, especially in coding, how do you define your mode? And in particular, maybe how do you think about competition?
Because it was interesting that OpenAI quoted you among its top competitors. I'm curious, who do you see from your perspective as your main competitors? Good question.
So first of all, you know, our approach is to not think so much about competition, but, like, think more about our users and what value we need to deliver to them. And I feel like, you know, we are very, very far from, like, you know, bringing to the users the kind of product experiences and technology that can be built with today's technology. So, like, there's a lot of work for us to do.
We think, you know, we are well ahead of any other player when it comes to this domain, which is, again, by thinking of Lean as being that, you know, that chat GPT inside your company, like, you know, one place where you can get all kinds of assistance, you know, from AI. If you look at that market, I think, again, like, you know, I think today, like, you know, it may be the case that, you know, we're helping you with, you know, with about 10 percent of your work, we are helping you automate that. I think you have to take that 10 percent to 90 percent over the next few years.
So there's a lot of work that we have to do is what I'm trying to say. So and I think, like, you know, like our mindset in general is to look forward, not like not sideways, not behind. So so that's that's that's number one.
Like, you know, I would say, like, you know, from a point of view of, like, you know, how we execute, like, you know, we are thinking user first, not competition first. In terms of more like, again, I feel like the when you build products, like, like you don't think about, like, you know, how you protect your advantage, at least that's not how I think, you know, like there is like I think technology building any technology is hard. It takes a lot of time.
And and I think that's the opportunity that we see, like, you know, we built a lot of technology over the last six years, which is which are things like, you know, building these really deep integrations with all of your enterprise systems, understanding, you know, content and knowledge within those understanding, you know, these complex permissioning and, you know, governance architecture that you have on all of your enterprise knowledge, and and then being able to use all of that to deliver a safe and secure AI experience inside your company. That's it. That's, you know, like all those things will take a lot of time to build.
And I think that that will serve as, you know, that time advantage for us. It's not it's not something that we can use that and forever, you know, feel safe because we've built it. Of course, other people will also build it over time.
You just have to stay ahead. And that's like one of the things where we are multiple years ahead of competition. The second area is around actual search, because if you think about making AI work in the enterprise, all of these AI models are they're not trained on your enterprise knowledge, you know, they but they are they have really good reasoning capabilities and they have, you know, they obviously also know their general intelligence based on all of the world's knowledge.
But to actually make them work in an enterprise, you have to somehow combine your enterprise data with the power of these models. And one of the core ways you do that is using search. And and search is very hard because if you don't like given any task or question, if you don't pick the right information, you can actually give the wrong data to the AI model.
And of course, it's going to actually give you back the wrong answer, you know, if you don't give it the right things, you know, as input, you know, the output is not going to be great. And that's that's another big advantage of clean is that we have built a really powerful search experience, we deeply understand your company's knowledge, we understand what information is authoritative, high quality, written by subject matter experts. So this, you know, the knowledge graph that we've built, you know, we understand your people and your data.
That's another thing, you know, which is which has taken us a lot of time to sort of mature and build and we feel is going to serve as a as an advantage for us. But but but but coming back, like, you know, all of these things, you know, like, you know, that's the technology world works like, you know, you build, you know, you build things first, other people will, you know, follow fast on that. And for you to stay relevant for you to, like, add value to your customers, you just have to keep thinking, you know, keep that focus on the customers and figure out like, you know, what are the next things we need to do?
I was I was talking about this concept of, like, you know, we want every individual, you know, in any company, you know, to have a really amazing AI team around them. It's a team of assistants, companions, co-workers and coaches, you know, that actually not only do 90 percent of your work for you, they're actually proactive in terms of how they help you to actually get, you know, make you better at the things that you do, like, you know, by, you know, coaching you like, you know, on things that you've been that you do. So so to sort of build that AI team, there's so much more work, you know, that we need to do on that.
So so I think that's sort of what we feel is the opportunity for us is that, like, you know, there are there are amazing opportunities to transform, like, you know, people's work and work lives. And and I think, you know, by staying focused on innovation and, you know, keeping up the speed, like, you know, we hope to win that way, not because, you know, we have created some more. Yeah, it makes sense.
And when you said, you know, staying ahead is key. It was interesting that OpenAI just announced model context protocol. So it seems like that is another explanation for why they cited you as competition.
Yeah, I mean, I think the well, we like, by the way, like we feel that, you know, we partner with, you know, with them and with many other companies. I mean, our philosophy, again, is to is to actually leverage all the great work that, like, you know, that companies are building on and and then, like, you know, do the work they know that others are not focused on doing. So that's how you sort of add value.
That's our core mantra. So we we actually, in fact, you know, OpenAI, you know, their models are one of the key model technology that we use, you know, in our product. So so so that's I think that's our view on them, you know, as well as, like, you know, any other model company, any other cloud provider, like we get to use amazing innovations that they make and bring it to our customers.
I just found your comment really interesting on on agents and how how they may evolve. Do you think ultimately the agents will be tied to individuals or to an enterprise? And could that model change over time?
Because like you said, different people have different skills and need different things. So to what extent is that model going to be specific to a person and maybe owned by a person, subscribed by a person versus an enterprise? There's certainly two parts of data that that's going to be relevant over time.
So you're absolutely right. Like there are going to be agents which are automating, you know, certain business processes, as an example, inside your company. And then there is this concept of a personal agent, you know, that is yours, like as an individual.
It really understands you deeply, you know, your work life, your goals, like what you need to do this week. And this personal agent is the one that proactively helps you with your work. And so, yeah, so we see these as two different things.
Our focus, we want to actually, of course, you know, contribute to both of them. But a big part of what we're doing is, you know, building that personal agent or a personal AI team around every individual at work. Very clear.
And Arvind, let's close it off with your thoughts on some of the key AI debates. So first of all, the limits of scaling laws. What are your thoughts on that?
Does every next generation of model require 10x more compute to achieve similar improvements going forward? So the way this technology is headed right now is that, first of all, whether you need or don't need, the way people are training models is, you know, is following that scaling law, which is they're actually, you know, trying to see what is possible if you actually increased, you know, everything 10x. And you're going to see, like, you know, how much impact it's going to have, what kind of, you know, additional intelligence capabilities, you know, these models will develop if you were to actually go and, you know, increase the amount of resources available to them 10x.
So that's the model. So first of all, like that's actually a desire. Like people want to actually train larger and larger models.
But there's also this thing that, like, you know, ultimately the improvements is not just coming from scaling. You know, there's also like, you know, real technology development around like, you know, which is sort of complementary to the core model development. You have to use, you know, a variety of different techniques, you know, outside of the model on the periphery to sort of, you know, you know, bring a lot more improvements and which is sort of in addition and stacks up on top of like, you know, the core capabilities of the model.
Even if you think about Gleam, like Gleam is a compound AI system. Models is actually only one part of our product. And by, you know, doing a lot of additional work around like, you know, like figuring out how to use these models effectively, we're able to solve business problems in a more intelligent fashion.
So yes, overall, I think we will keep seeing that, like, you know, the next generation of these models will keep using in an order of magnitude higher amount of resources. And probably like, you know, hopefully, you know, we'll get like, you know, that it turns in like, you know, intelligence as, you know, as part of our scaling too. And then secondly, main drivers of compute costs going forward, we talked about agentic reasoning, timescale inferencing, where do you see training, which of the vectors is going to be most important for increasing compute?
Yeah, I feel that in the future, more of the compute costs are, you know, going to be on the inferencing side. I hope, you know, that's actually important because I think we want to make sure that AI is having a big impact. And so like training models, like, you know, that's costly, that takes a lot of resources.
But once it's trained, you want to make sure that it's used much more widely, you know, in our day-to-day work and our lives, you know, than what it is today. Like AI has to play a bigger role. And so therefore, from that perspective, you know, I feel that, you know, without any doubt that over time, like most of the compute costs are going to be, you know, on inferencing side.
And that ratio is just going to keep, you know, getting more and more skewed. Makes sense. And then do you think we're going to see emergent superintelligence or is it ultimately something that's going to be constructed in some shape or form?
I might say that that's something that I'm actually, I don't feel qualified enough to actually comment on that. We are, I think the way we think about like AI is that like even today, these models that we have are, like in some ways you can say they're limited in terms of their intelligence and what they can really do. Like they're also not consistent, they can make mistakes.
But the technology that's available to us today itself is so powerful and we're not even scratching the surface in terms of actually utilizing, you know, this power today, you know, to bring improvements to like, you know, our products and business processes. So I feel like, you know, like the thing that I'm most excited about is that even if you don't see, you know, any big, you know, advances in the level of intelligence of these models, which is not going to be the case, by the way, but even if that were the case, like we've got another 10 years of like using, figuring out how to actually use this amazing power of the models to build very meaningful products. And I think like, you know, increasing the impact of AI overall, you know, in businesses and in our world.
Great. This was an awesome ending. Thank you so much.
To our listeners and clients, thank you for your time and engagement. If you have enjoyed this episode, you may subscribe to UBS Market Moves, available on Apple Podcasts and Spotify for additional content. Until next time, where again, we decode the future of AI layer by layer.
UBS Chief Investment Office's investment views are prepared and published by the Global Wealth Management Business of UBS AG or its affiliates. The views and opinions expressed in this material by external guest speakers are those of the author, speaker and are not those of UBS, its subsidiaries or affiliates. Accordingly, UBS does not accept any liability over the content of this material or any claims, losses or damages arising from the use or reliance of all or any part thereof.
This material has no regard to the specific investment objectives, financial situation or particular needs of any specific recipient and is published for informational purposes only. For a full legal disclaimer applicable to the independent investment views produced by UBS, please visit our website at UBS.com forward slash CIO dash disclaimer.
Sources & References
How we cover this story