EP 520: IMB Think 2025 - AI Updates that could shape Enterprise Work

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. On this show, we talk a lot about consumer AI. Those are the AI chatbots that we as individuals go in and log in and start using, right, the ChatGPTs and the Geminis and the Claudes of the world. But I understand that that's not always how work gets done at work, right? Because I know even in our audience that's not always technical, many of you are working in enterprise organizations. And even if you might have access to those tools, maybe on a team or enterprise plan, there's a good chance that you're tapping into a much more a more robust way to work in specifically with generative AI. And there's a good chance that to do that, you're using IBM's products. So that's why today I'm excited to talk about what was just released at the IBM Think conference in Boston and some AI updates that I think could shape enterprise work. All right, what's going on, y'all? My name is Jordan Wilson, and I'm the host of Everyday AI. It's for you. This is your daily livestream podcast and free daily newsletter helping us all not just learn what's happening in the world of AI, but how we can actually leverage it to grow our companies and our careers. So if that's what you're trying to do, welcome, you're in the right place. It starts here on the livestream and podcast, but where you're actually going to leverage what you learn is on our website. That's youreverydayai.com. There you can not just sign up for our free daily newsletter, we'll be recapping all the AI news for today and everything we talk about on this very episode. But you can also go listen to more than, I don't know, 510 now episodes from some of the leading, some of the world's leading experts in AI, sort of by category. So no matter what you care about. All right. So make sure you go do that if you haven't already. All right. Before we get in and we talk about what's new in IBM's Watson AI platforms, let's start off as we do most days by going over the AI news. A lot going on as always. So first, Amazon is reportedly developing in AWS an AI-powered code generation tool called Kiro that aims to produce code in near real time by leveraging AI agents. Kiro is designed to support both web and desktop applications with multimodal capabilities and compatibility with third-party AI agents, potentially broadening its usability in diverse coding environments. So beyond generating code, Qero can also reportedly generate technical design documents, identify potential issues in existing code, and optimize entire codebases. So Amazon already offers an AI coding assistant named Q Developer in their Amazon Q platform, which is similar to GitHub Copilot, but Qero appears to be a more advanced tool with broader functionality. So initial plans suggested a launch for Qero towards the end of June 2025, though those timelines might have shifted, reflecting the fast-evolving nature of AI tool development. The emergence of Kiro highlights the growing competition in the AI-assisted coding space, with companies like Cursor and Windsurf attracting massive investments. And Cursor also— or sorry, Windsurf also just reportedly being acquired by OpenAI for $3 $1 billion. All right, next piece of AI news, an interesting one. So in a groundbreaking use of AI, a victim impact statement was delivered in court by an AI-generated video of Chris Pelkey, a man that was killed in 2021 in a road rage shooting in Arizona. So Pelkey, who died at age 37, appeared on screen via an AI avatar addressing his killer and expressing forgiveness and regret over the tragic encounter. And this was a message created by feeding AI models with Pelkey's past video and audio. So this marked possibly the first, at least reported time, that AI was used to represent a deceased victim's likeness in a courtroom, raising new questions about how AI could be integrated into legal proceedings. So it was Pelkey's sister who helped create the AI representation, describing the process as a quote-unquote Frankenstein of love, aiming to capture what Chris might have said if alive. So the judge praised the AI statement, noting it conveyed genuine forgiveness despite the anger surrounding the case, and sentenced the killer in the case to 10.5 years in prison on manslaughter charges. All right, our last piece of AI news. There's a new king of the hill when it comes to large language models. Google has introduced a special edition of its Gemini 2.5 Pro preview called the I/O Edition. So this updated version of its Gemini 2.5 Pro model aims to improve coding performance and web app development ahead of its annual I/O developer conference here in about 2 weeks, actually. So the new model is accessible through the Gemini API, Google's Vertex API, Google AI Studio platforms, and the Gemini chatbot app on web and mobile. So Gemini 2.5 preview. This one's pretty important. So it now leads the web dev arena leaderboard. That's one of the few benchmarks that Google had previously not been number one in, although they've been number one overall in the LM arena, which kind of pits blindly two different chatbots against each other. You pick a response across all these different categories. So pretty noteworthy that now Google is essentially dominating in every category that it can compete in. And it's now 37 Elo points above the second place model in OpenAI's O3 and 40 points in front of the third place model, OpenAI's GPT-4o. So a lot of improvements so far in the new 2.5 Pro I/O version. That's surprisingly not a mouthful as some of the other names that we've seen in the past. But some of those improvements include better coding skills, especially in code transformation and editing, along with reduced errors and improved function calling trigger rates based on developer feedback. So Google highlights the model balances aesthetic web development and steerability, meaning it can follow user instructions effectively while producing attractive results. And yeah, it looks like a lot of these updates can be seen in Google's Canvas, Canvas tool. So we might have to do another update on this pretty soon. All right, let's get straight into it. IBM Think is happening right now. I'm actually, if you're watching on the, watching on the live stream in my hotel room, I'm about to go down and watch the key, the next round of keynotes here in less than an hour. So luckily I was able to partner with IBM to get a quite literal front row access to you all to see what is happening with IBM. I was able to attend the CEO keynote yesterday, you know, asking, as I often do when I get to attend conferences, you know, talking to the people that are actually building these products that we all use, you know, talking with executives. I'll probably have an interview, you know, with someone at IBM dropping on the podcast either later this week or next week. So a lot of big announcements happening at IBM Think 2025. So here is a quick highlight, at least of the things that we're going to be going over in today's episode. And we're going to be going over also just a reminder, right? And maybe if you aren't overly familiar with Watson, Watson xAI offerings, we're going to be recapping those as well. So don't worry if I'm going over some of these what's new in Watson and you're like, wait, what does that part of Watson do? We're going to be covering that. But here's at least for our audience, right? Kind of the non-technical business leaders of the world. I know a lot of y'all are technical, but, you know, it's one area where people are always saying like, hey, Jordan, love everything that you talk about on Everyday AI, but, you know, I'm the CEO of the company or I'm a, you know, CMO and, you know, I'm not always the one in there, you know, fine-tuning models and, you know, I'd love to know a little bit more of the non-technical things at the enterprise level. So here we go. Here's what's new. So we have, some new updates in Watsonx Orchestrate and the build-your-own-agent capabilities. Also inside Watsonx Orchestrate, we have pre-built domain agents. That one is exciting. Also, yeah, a lot new inside Watsonx Orchestrate. Also a new agent catalog with 150+ agents. We also have new IBM and Salesforce collaboration for AI agents. Also some big partnership announcements with IBM and Oracle in working on the new availability to have multi-agent orchestration on Oracle Cloud. We saw a preview of IBM Granite 4.0, IBM's new open source model, but the tiny version. So we'll probably see the larger versions of those released at a different time. And last but not least, the Watson x Data Intelligence for AI-ready enterprise data. There is a lot more that was announced, y'all. There's the zero-copy integration with IBM Z, the support for the MCP protocol, very big Watsonx integration with Meta's Llama stack, IBM Web Methods hybrid integration. So there was a ton, a ton that was announced. As most of these conferences, there's more than I could cover in any one show. So yeah, make sure you continue to just stay tuned newsletter. We'll be sharing a lot more about even the things we aren't going to go into too deep today. So let's start at the beginning. You know what's new now. Let's talk about what is Watsonx. If you don't know, well, it's the AI model foundation that powers business applications with intelligence inside of IBM's platform. So it provides enterprise great versions of technology similar to ChatGPT for specific use cases, all based on your data. Right. That's the biggest thing. You know, if you're using— even if you're using enterprise version of the consumer LLM tools, right? So, you know, the ChatGPTs and the Claude's, etc., right? You're not working with your dynamic data. And when your data lives in all of these different enterprise tools, right? 2 I already named, right? Salesforce, Oracle, et cetera. It's hard to work with your up-to-date data, which is extremely important, right? Those companies, those AI labs, the Anthropics and OpenAIs are working at bringing more dynamic enterprise data into their platforms. But if you need it now, right? If you need something flexible, that's what a lot of people have been using Watson X for years, right? IBM has been a leader in artificial intelligence technology for decades, right? Going back to literally Watson, right? The famous AI that defeated the world champion in Jeopardy, you know, back more than 10 years ago. So IBM has been a key player in their Watsonx platform, is really kind of the ground or the area where you can go in and connect all of your business data and use different models to change how you work, right? To rethink how you work. So this includes IBM's Granite models, which are optimized for business tasks and industry knowledge. But there's a lot of other models that you can use inside as well. So, you know, some of their most up-to-date integrations, you can work with other open models like Meta's Llama, like the Mistral models, and just a lot of now, you know, more powerful, smaller open models as well. So even Google's Gemma, GPT-3, which is an open source model, you know, Watson X has the ability to connect with that as well. So, and even talking on this whole, you know, open, you know, open model and small language model movement, right? So this is, this is something that was talked about extensively in the keynote. So IBM CEO and Chairman Arvind Krishna talked about this. I actually shared a video, you know, kind of from the keynote floor right after on LinkedIn, and I shared it in our newsletter as well. So going over 3 of my biggest takes. But one of these that I didn't really get to is really just how much Arvind was talking about this, this surgence of small language models, which if you listen to this show, you know, I've been extremely bullish on small models for a very long time, even before they were good, right? 2 years ago, you know, the gap between small language models and large language models seemed like an insurmountable gap. It seemed like they would never be on the same page. And fast forward to today, they are, right? You have even open small language models such as Meta's new Llama 4, such as some of Mistral's models. Even as an example, Google's Gemma 3, their small open model, right? They're on par. Maybe they're not in the 1A tier, you know, with the Gemini 2.5 Pros, with the, you know, OpenAI O3, but they're definitely in that 1B, right? Where before they weren't in the first tier, they probably weren't even in the second tier 2 years ago. So this is important to know that, you know, Watson X's AI platform is really set up to bring in a variety of different models and suitable for different enterprise needs. And the biggest thing is it just integrates with your existing systems, right? You know, not having to worry about, oh, is there, is there a connector for this? Is there an integration for this? Oh, if I'm building a project, you know, I have to make sure to, you know, re-upload all my documents, you know, every, every quarter or every month or every week, right? That you have to right now with some of the more consumer models that don't have that, you know, that true integration that you may need. To have your data that's from now, not from, you know, last week or last month. All right. So that's watsonx. watsonx is where it starts. Well, what about watsonx.ai? So this is the AI model foundation that powers business applications with intelligence. And watsonx.ai provides those enterprise-grade versions of the technology similar to ChatGPT, and it includes, like we talked about, IBM's Granite models. All right. Now, orchestrate. All right. We're going to be talking a lot about orchestrate because I think this is where a lot of our audience, you know, if, if your company does have access to IBM Watson X, this is probably where you're going to be excited to go look inside the orchestrate platform. So first, what is Watson x orchestrate? So this is where you can create AI assistants and agents that can actually complete tasks across your business systems, and you can connect multiple enterprise applications to provide unified workflow automation, and you can transform AI from just answering questions to performing valuable business actions. So, you know, Orchestra is where you can actually put like AI to work, right, by building AI agents and, you know, automating repetitive workflows. So if you haven't seen it, you know, we'll probably be sharing a link to a couple of pages in our newsletter where you can go see it, you know, for yourself. But this is where there's a really like it's, it's, it's low code, right? There's low code and even no code options to connect your enterprise data. So, you know, think, if you've used something like Zapier, right, where you can build these visual flows, right? Like, oh, if I get an email in Gmail and if it has this type of subject line, then it should trigger, you know, update something in a spreadsheet, then it should update that same line of data in our CRM, right? So there's, there's been marketing automation for very, like, for a very long time. right? So think of that, but for your internal data, right? And when you're using all of these different enterprise systems. So within Orchestrate, this is where you can really build, you know, visual agents in no-code and low-code using a variety of models, right? You can use different models for different tasks. And this is where you can really just unify your workflow automation without having, you know, to, you know, try to patch something together, you know, in Zapier or, you know, using multiple AI tools, right? This is where it's just, it's live. It's, it's living inside your company's dynamic data. So a couple examples, IBM said that their internal HR implementation handled 94% of company-wide HR requests. So, you know, being able to build something like that and working off of these as templates as well. So even inside orchestrate, when we talk about the agents, and we're going to get to the agent catalog here in a minute, you know, just the ability to work with templates I think is extremely important. So this was during another keynote presentation I watched yesterday. You know, these agents are described as more doers that act autonomously and orchestrate workflows across your enterprise, right? Versus a lot of the AI, you know, especially the large language models that we use right now, we don't think of them as necessarily autonomous because we, the humans, still for the most part have to go in and feed them information, right? Whereas with Watson X Orchestrate is much more agentic, right? It is creating flows that actually do the work and aren't always waiting for the human to go in and set off one of those integrations. All right, next, the build your own agent capabilities. So some new updates to this inside Watsonx Orchestrate. You can create customized AI agents in under 5 minutes with no coding required. That part to me is extremely impressive, right? Even things, right? I'm not an enterprise. Let me start that out. Like, let me start this out now. And this is probably why I haven't talked more about Watsonx, right? I'm not an enterprise. It's not always right. If you're a super small business, if you're an entrepreneur, you know, if there's 10 employees, Watsonx probably isn't something that's for you because the reason is, is you don't have a lot of that enterprise data that bigger enterprise companies would need, right? So that's why I think I haven't even talked as much about Watson X, but I wish, right? Maybe I'll have to ask, ask my friends at IBM, like, hey, hook me up with some, some, some access to Watson X here so I can talk about it a little bit more. Right? Because having all these enterprise tools, right? The Salesforce, the Oracle, right? These are all systems I don't use as an extremely small business, right? With, with, with, with a handful of people on the team. But when you have, these and the ability to think that you can create an agent in less than 5 minutes that taps into all of your enterprise data. That is not normal, right? That's not normal. Uh, you, you know, uh, a year ago to think that you could build your own agents with up-to-date enterprise data, we would've said, oh, you know, a year ago if we were talking about this, we would've said, oh, that's, that's 3 to 5 years away. And, and here we are in a year with this new build your own agent, uh, capability inside Orchestrate. It's here. Uh, so simple. Drag-and-drop interface. So, you know, seeing this demoed, uh, in different sessions on the showroom floor, in keynotes, y'all, it's, it's simple, right? You don't have to be a, a technical person to build some of these agents, and they can be tailored, uh, to your unique business processes and requirements. Uh, so, uh, IBM says that organizations can start 70% faster, uh, with Watson X orchestrate and really just going beyond and bringing this AI agent creation beyond people with technical backgrounds. Right. I think that's huge in this, you know, build your own kind of movement that we've seen not just from IBM, but from a lot of other companies bringing in, you know, no-code and low-code agent builders. Right. It's something I've been extremely pro no-code, low-code, right? I've been saying this for a long time that non-technical people are going to be building their own agents. They're going to be building their own applications, you know, without even necessarily knowing what's going on under the hood, right? And this, the new Watson X orchestrate, build your own agent capabilities is another step towards something that I've been talking about for a long time. So, you know, here, for our livestream audience, right? A quick example and something that I love is, you know, you have a familiar panel. So in the middle, when you are using this, you know, build your own agent capability, you can edit your knowledge settings and choose which, you know, data sources that you want to connect to a particular agent. Maybe you want it to connect to your, you know, your Workday, but maybe you don't want it to connect to Salesforce as an example. So you can easily pick and choose which data sources that it has access to. But the thing that I love is you can get a preview of the agent on the right-hand side. So you kind of have this dual pane where you can kind of build it, you know, connect the different settings, upload files, etc. But then you can see and preview it on the right-hand side. And something that I really like, which is, I know sounds like a detail, right? But you can use reasoning models with these agents, which is extremely important because then, you know, yeah, it's fun to see how these, you know, reasoning models, how they kind of think and how they plan. Right. And that's another thing that can't be understated. The fact that, you know, these agents have access to reasoning models, that's extremely powerful. But even looking to see, right, and look at the reasoning and look how it's accessing your data. Right. When we talk about some of the main problems with AI adoption, not even just at the enterprise level, right? Just even at the consumer level, it's always trust, transparency, data governance, etc., right? And being able— it sounds like a small thing, but being able to see the reasoning and see how a large language model is or is not accessing your data, to see how it's handling the query. So to be able to build an agent and just to see like, oh, okay, you know, In this example here I have on the screen, it says, you know, what is the active— my screenshot is actually cut off. What is the active return policy for orders, you know, of a certain amount? And then you can see, okay, it's going into a tool. It's getting the get order status. It's analyzing the return policy for an order in transit. It's looking up different items that you can set. And then it says, okay, this order is in transit and the 30-day return policy applies, right? Think of that, right? Being able to ask a question of an agent that you've built with no code and being able to see it reason, and then you could easily go in and change it if you see something is not performing how it should be based on the reasoning. Extremely powerful updates in this one, in the build your own agent. All right, next, there's prebuilt agents. Yes, this one also pretty big. So the prebuilt agents and the, you know, agent catalog, we're going to cover them both back to back here. So prebuilt domain agents, huge. So as an example, a company like IBM that works with the largest enterprise companies in the world, right, they know how different sectors are using their agents, right? Their agents aren't new. So now when they're announcing these prebuilt domain agents. These are like agents that have been fine-tuned across the enterprise, right? Across, I'm sure, millions of interactions. So, you know, these are, you know, talk about a great starting point. So as an example, an HR agent, right, that can handle employee inquiries, time-off requests, and policy questions. That's been some of the coolest demos I've actually seen from the showroom floor and from keynotes is just, you know, all these like like, hey, how, like, do I have time off? How can I submit it? Right? All these HR policies that, you know, is probably a pain for employees and maybe for HR people to deal with, to be able to see an agent, a domain-specific agent based on your company's data that goes in and says, oh yes, this employee does have enough PTO. And yes, it's submitted. Right? So it cuts off the need for back and forth. Oh, is this going to get approved? Oh, the HR person didn't get back to me. Oh, the HR person. Oh, 50 people just emailed me about Thanksgiving. Like, what's going to happen here? Right? So for that to all be automated is huge. The same thing with sales. I think HR, sales, marketing are going to be agents, prebuilt domain agents that are going to be extremely popular inside IBM Watson. So the sales agent supports prospecting, account research, and proposal development. Procurement agent automates vendor selection, purchase approvals, purchase approvals, and supply management. So this is huge. And one example that they talked about a little bit yesterday was the Better Business Bureau. Right. Most people, especially if you're here in the US, you know, the Better Business Bureau, you know, this is the organization that essentially gives the rubber stamp on certain businesses and says like, okay, this business is legit or this business is not legit. So the example from the Better Business Bureau they reported a cost savings of $1.5 million annually from implementing some of these prebuilt domain agents. So, you know, just one, right? And you wouldn't think of the Better Business Bureau as like a tech organization. I don't know, at least not me, right? I think of the Better Business Bureau, okay, that's got to be some, you know, some suits that have, you know, 50 layers of bureaucracy, right? Nothing against them. It's just that's how I think of the Better Business Bureau and a lot of government. But the fact that, you know, the, the Better Business Bureau, you know, which isn't a government agency, by the way, but the fact that they were able to just by using these prebuilt domain agents save $1.5 million already, I'd say that's pretty telling. You know, some other examples talked about, you know, HR agents have been successfully deployed at organizations like EY, where they provide significant improvements in employee service delivery. So yeah, a lot of examples and use cases for these pre-built agents. Similarly, now we have this new agent catalog. And this is where you can access these 150 ready-made agents and tools from IBM and third-party partners. Right? So this is kind of the, the app store, so to speak, for Watsonx Orchestrate agents. Right? So all these different, you know, agents that are ready to go and can tap into your enterprise data. So you can integrate the AI capabilities with popular enterprise applications like Salesforce and ServiceNow. And the cool thing that was just announced as well is there's some new agent observability features provide, you know, that kind of provide governance and performance monitoring. You know, that's always important, being able to trace agentic AI, right? And that's one of the, rightfully so, that's a huge concern of many enterprise leaders when, you know, not even talking about AI, right? Not even talking about, oh, our humans are gonna go in and use a large language model, right? There's obviously already a lot of things that you need to keep in mind with, you know, transparency, accuracy, trust, data governance, all of those things, ethics, right? But then when you go to agents, there's a whole nother layer. Because unlike, unlike agentic AI, when you're talking about a human using a large language model, it's at least a little easier to be like, hey, what happened here? Let's trace back our steps. Hey, human, what happened when we were working on this project inside a large language model? With agents, it's a little trickier, right? Because you can't necessarily, you know, have the same ability to go talk to an agent and be like, yo, what happened here? But this new agent observability feature inside Watsonx Orchestrate does kind of that, right? You can at least see, you know, the bullet points of how these agents are working, where they've done work, how they're operating, what may be working and what doesn't. So it's kind of like being able to look under the hood that we have similar applications for non-AI tools, right? This is what your IT people are always, you know, spending their time on. So the IT departments, I think, are really going to like these new observability features. Just to be able to see how these agents are working, especially when you have multi-agent orchestration. I think that, you know, traceability and being able to understand how these agents are actually working without human intervention at times is extremely important to monitor and ensure accuracy, trust, reliability, all of those things. But I do think, especially if you're already an IBM Watson, you know, ex-customer, if you're wondering like, okay, Jordan, there's probably a lot that was released, you know, during IBM Think, like, where do I begin? If I were you, this is where I would start. And again, all of these, many of these agents are templates as well. So yeah, you could go find a great one, and maybe it's not your certain CRM or ERM, right? But maybe everything else is perfect. So maybe all it is is swapping out the data that it is connected to. Maybe it's just tweaking, you know, something in the desired output, the output format, right? But there's now 150 enterprise-ready ready-made agents ready to go that, you know, if, if you're using the Watsonx platform, uh, there's, there's not going to be, you know, again, you always think things are easier said than done, but, you know, I'm, I'm out here on the floor watching live demos, right? Not like pre-recorded videos, uh, you know, and asking people, oh, can you go, hey, go into this flow, you know, change, change the data source. And I'm watching it happen in front of me. Obviously, these are by IBM, uh, employees, a lot of them who work on the product team, uh, but I'm seeing them very easily in a couple of quick clicks without coding, go in and customize some of these prebuilt agents. So it's something that if you are on the Watsonx platform, this is something you need to be paying attention to. So yeah, even for our livestream audience, just have a little screenshot of the new agent catalog. So yeah, you know, you can search, you can search for different things. You know, maybe you want something for sales, something with your CRM, whatever it may be. So, you know, it is very much like an app store for Watson x Orchestrate agents. So very cool. You can also, you know, there's toggles and filters for different, you know, if let's say you're using Asana or AWS, right? You can just go ahead and click those things and stack different filters. To see which of these prebuilt agents might be a good fit for your workflow. AI moves too fast to follow, but you're expected to keep up. Otherwise, your career or company might lag behind while AI-native competitors leap ahead. But you don't have 10 hours a day to understand it all. That's what I do for you. But after 700+ episodes of Everyday AI, the most common questions I get is, where do I start? That's why we created the Start Here series, an ongoing podcast series of more than a dozen episodes you can listen to in order. It covers the AI basics for beginners and sharpens the skills of AI champions pushing their companies forward. In the ongoing series, we explain complex trends in simple language that you can turn into action. There's 3 ways to jump in. Number 1, go scroll back to the first one in episode 691. Number 2, tap the link in your show notes at any time for the Start Here series, or you can just go to starthereseries.com, which also gives you free access to our inner circle community where you can connect with other business leaders doing the same. The Start Here series will slow down the pace of AI so you can get ahead. All right, let's talk a little bit more about this one because I think this is going to be another popular collaboration as well between IBM and Salesforce. So the Zero Copy integration connects mainframe data to Salesforce without duplication. So the new sales— or the updated sales prospecting agent— this collaboration between IBM and Salesforce pulls leads or sorry, finds leads, pulls contact info, and can even draft outreach to certain contacts in your Salesforce. And then because it integrates with a lot of your, you know, other tools, you know, there's even an employee support agent in Slack that you can collaborate with as well. So the Salesforce integration is part of IBM strategy to extend beyond just the traditional you know, enterprise systems and into other enterprise software that many, you know, large companies are relying on. So this one's, you know, this one's pretty like, I got to see this one more in action because I looked at it. I'm like, okay, this is obviously very powerful, right? But I'm wondering for those, you know, for those companies that are both, you know, using Watsonx, their AI platform, but maybe also using Salesforce's Agentforce platform because it looks like there's a lot of crossover here. Here. So yeah, if that is you, I'd love to hear from you, right? If you've been using Salesforce's agent force and now we see this new and improved collaboration effort between IBM Watson X and Salesforce and these new improved AI agents, you know, some of it's got me scratching my head like, okay, so does this mean like, let's just say some enterprise companies are only, you know, tapping into 10% of what Salesforce's agent force offers. I'm like, well, maybe that 10% is now what is covered in this new IBM Watson and Salesforce collaboration. I don't know, right? And Ty will, you know, answer that story. But, you know, this one kind of struck me as interesting. So I'm like, oh, you know, good on IBM for being able to pull off this level of agentic collaboration with Salesforce. But I'm also like, okay, is this going to potentially take away from, you know, Salesforce's agent force? I guess in the end, right, you still have to be a Salesforce customer in order to integrate it within IBM Watson X. But I was like, huh, maybe this takes a little bit away from the AgentForce platform. But like I said, I think time will tell on that one. All right, one or two quick other ones here as we start to wrap up today's show. So the other, another big announcement with, you know, when we talk about enterprise, I mean Oracle. So now Watson xOrchestrate is coming to Oracle Cloud Infrastructure for enterprise customers. AI agents can operate seamlessly across Oracle and non-Oracle business applications. Some of the initial use cases, at least right now, are focusing on HR use cases with plans to expand to other departments. So this partnership right now gives organizations way more flexibility in where they can actually deploy their AI workloads as part of IBM's hybrid cloud strategy. So this collaboration with major cloud providers supports clients who want to run AI where their data resides. And that's pretty big, right? Because, you know, a lot of companies, they have, you know, huge data investments. And so you might need this, you know, kind of official layer of collaboration between, you know, in this instance, IBM and Oracle. Let's get tiny. We've talked about some big announcements. Let's get tiny. Granite 4.0. All right. So this is a new preview of IBM's small, small version of their language model, Granite. So right now, I believe the latest full model, I think we're at the 3.3 version of Granite. So this is a preview. So we should see, as with most large language models, anytime you go up from like, you know, a 3.3, 3.4, 3.5 to a 4, you know, it's usually a pretty significant step in terms of capabilities, performance. So, you know, pretty, pretty big here. So we don't have the full thing. Unfortunately, I was, you know, hoping we might see the full Granite 4 series, but we have the 4 tiny preview. So what this is, it is a small 7 billion parameter model from IBM in their Granite series that's more efficient and performs similar, IBM says, to larger models. It also dramatically reduces computing requirements, making AI more affordable to deploy. And it's open source, right? So the open source release encourages community innovation and customization. And here's the thing that I really liked about Granite 4.0 Tiny. It can run on a consumer-grade GPU, which is funny because there's always arguments and companies always say, right, like, oh, this can run on a consumer-grade GPU, but this one actually can't, right? Because sometimes people say like consumer-grade GPUs and they're talking about like a, you know, like a $2,000 Nvidia chip. This can run off like a $350 GPU, right? So like for the most part, especially if you have a newer, you know, PC, this is one that you probably wouldn't even have to go and upgrade your GPU to go ahead and run this model. Locally. So that's extremely, extremely exciting. And like I said, this is part of the Granite 4 series that should be released later this summer. So we will be covering that when the full family, but at least right now, there is the Granite 4.0 tiny preview. You can access it in Hugging Face. IBM does say like, hey, this is a preview, right? So this is not something that they're necessarily recommending that you, you know, integrate across enterprise because it is a preview. But if your company, if your organization is using the Granite models now and you're finding a ton of value, which I know a lot of companies are, it's anytime you get a preview, even if it is one of the smaller models, I think it's always important to start experimenting with what's new, right? Running side-by-side comparisons, seeing how this may or may not change your workload, how it might make it better, some things you might have to improve upon, and honestly, just the new capabilities, right? Because presumably we're going to have a lot more capabilities in the Granite 4.0 family of models. All right, let's talk about last but not least, the Watsonx Data Intelligence. All right, so this transforms unstructured data like documents and email into valuable AI inputs. So yeah, some big improvements here on their data platform. IBM says that this improves AI response accuracy, this new update to the platform, by 40% compared to conventional methods. And content-aware storage automatically processes and indexes information for AI use. This is one thing that really also stood out to me in Arvind's keynote yesterday. So the CEO and chairman, Arvind Krishna, what stood out is, you know, he said right now 99, 99% of businesses' data isn't being used within AI systems, right? Which at first kind of shocked me because I'm like, wait, that's a lot of data that, you know, companies really need to spend more time bringing into large language models if you want to be an AI-first or an AI-native organization. you really have to work hard to bring that 99% of data, whatever that is, whether it's structured data, unstructured data, your company's knowledge, right? Uh, it's a long list of things that are not currently inside large language models. But that kind of struck me as also like shocking, but also like, okay, duh, that makes sense, right? Think of even right now, uh, I think so much of the data that we're just bringing into large language models is just, it's, it's, it's predicated just on tasks that we need to complete. right? Unless you already are using, right, something like IBM Watson X and, you know, have hooked up all of your major data sources, you know, your data warehouses, data lakes, whatever, right? But otherwise, and for, I would say even for a lot of enterprise companies, they're really just connecting their data on a need-to-have basis. So there's, I think there's so much room for growth, right? Because using AI is not a separator using AI is not a moat for your organization, bare minimum, you have to start bringing in, you know, more of that 99% to really stay competitive, I think, in 2026 and beyond. So this is pretty cool, this new Watsonx Data Intelligence and just how it's able to better bring in unstructured content and make it usable within their platform. Form, right? So think, you know, unstructured, right? So structured, unstructured, right? The easiest thing is I say structured data, right? If you're not a technical person, structured data is something that would live in a database, right? It's something that can be categorized. It's something that could be in a, you know, in a CRM, right? There's a dropdown list or a spreadsheet, right? Unstructured data is everything else, right? I think one of the biggest gold mines of unstructured data is how your company makes decisions. So decision-making process, all of this domain expertise that lives in people's heads, all your meeting transcripts, right? I know everyone, especially I think, you know, since we've had the hybrid and work from home era, it's like it seems like there's more and more meetings. Meetings are gold, right? And being able to get the transcripts and being able to get the insights from your team, from your leaders, that's huge. So this Watson x Data Intelligence, great, great advancement from IBM Watson. All right, that's a wrap for at least for our more non-technical audience, our everyday business leaders, what came out of IBM Think 2025 so far. All right, we're going to have more throughout probably a little bit of this week and next. I know there's a lot going on. I'm I'm, like I said, about to go head down right now for another keynote presentation here at the conference. So hey, I hope this was helpful. Let me know if there is anything else that you want to know. If your company is using IBM Watson X, if you have questions, if you know, because I know anytime there's new, there's new releases, there's also so many questions like, oh, this is great. How do I get access? How do we use it? What's the difference between, you know, feature A and feature B? If you have those questions, Reach out to me, right? That's the thing. You know, I'm lucky enough to have, you know, good connections at a lot of these big companies. And, you know, maybe, you know, you have a rep at some of these companies. Maybe you don't. But I love hearing from you all what questions you have or, you know, maybe what's exciting you most about what we just went over in today's show. I hope it was helpful. If so, if you're listening on the podcast, please subscribe and leave us a rating. If this was helpful, click that repost button. You probably have some coworkers, some colleagues, some people you know using IBM Watson. you know, their AI tools. So let them know about it. Thank you for tuning in. Please go to youreverydayai.com, sign up for the free daily newsletter. We'll see you back tomorrow and every day for more Everyday AI. Thanks, y'all. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit youreverydayai.com. And sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.