---
title: "Transcript: Who Owns AI Inside Your Company? The Rise of the AI Ops Role"
webinar: "Who Owns AI Inside Your Company? The Rise of the AI Ops Role"
date: 2026-11-11T15:00:00Z
duration: "1 hour"
speakers: ["Charlie Cowan"]
topics: ["The four jobs of an AI Ops role: surface, own, centralise and chase", "Why individual AI wins stay siloed, and what that costs when people leave", "Giving every business-critical AI tool a named owner", "Building a company brain: one governed home for skills, prompts and context", "How to spot, spotlight and reward grassroots AI innovation"]
recording: https://www.youtube.com/watch?v=FeS4cCR-O4A
canonical: https://kowalah.com/resources/wednesday-webinars/who-owns-ai-inside-your-company/transcript
source: Kowalah Wednesday Webinars
---

# Transcript: Who Owns AI Inside Your Company? The Rise of the AI Ops Role

**Recorded** 11 November 2026  
**Duration** 1 hour  
**Speakers** Charlie Cowan

## About this session

Your people are building AI tools, prompts and skills on their own laptops, and nobody owns them. Charlie shows what the new AI Ops role does about it, from the first tool inventory to a shared company brain, so your best AI work stops leaving with the people who built it.

## Topics covered

- The four jobs of an AI Ops role: surface, own, centralise and chase
- Why individual AI wins stay siloed, and what that costs when people leave
- Giving every business-critical AI tool a named owner
- Building a company brain: one governed home for skills, prompts and context
- How to spot, spotlight and reward grassroots AI innovation

## Recording

https://www.youtube.com/watch?v=FeS4cCR-O4A

## Transcript

**Charlie Cowan:** Good afternoon, good morning, depending on wherever you're dialling in from. Welcome to this week's edition of the Kowalah Wednesday Webinar. If it's your first Wednesday Webinar, we get together at the same time every week and spend an hour talking about what is top of mind in the world of AI, and specifically enterprise adoption of AI and enterprise use cases. There's so much going on at the moment, as we will cover today, that it gives us some time to take stock and think about how people are implementing and using this and getting value out of it.

We will go through our normal flow. I'll spend about 30 minutes on our main topic of the week, which I'll introduce in a moment: hiring an AI Ops lead. Then we'll spend maybe 20 minutes on AI in the News. Like I said, there's so much going on, and there are some great new releases that have come out this week, so we can talk a little about those and I can share how we're using them. Then we've got some time for Q&A.

We've got people live on the call, and we also know that a lot of people catch up on YouTube. If you're watching live, don't feel that you need to wait until the end to ask a question. We're in a Zoom meeting, so we have a chat box, which I shall open up on my screen, and there is also a Q&A box. I'll try to keep my eye on both, but if you've got a question as we go through, ask it. Caitlin, my colleague, is on the call as well, and she'll keep me on track if I miss any questions. If you're watching on YouTube and you've got a question, ask it. We love seeing those on YouTube, and we'll come straight back to you with the answer as soon as we see it.

So, straight into today's core topic: hiring an AI operations lead. We're going to go through what they are, where to find them and what they do. Just the title, AI Ops lead: is that the right thing? When we look across all of our clients, we've consistently seen them hiring someone who fits this persona. It tends to be someone a little bit younger, who might have come from a consulting background. They are going to be put into the role of being a bit of a doer in the AI function in the business. I'll talk a little more about what that means in a minute, but I want to clarify that when I'm thinking about an AI operations lead, I'm not talking about a Chief AI Officer, someone in an executive role. It's the person who's getting the work done.

So, do you even need one of these AI operations leads? Why would you possibly want one? This is the challenge we see popping up in organizations. The business has got a problem. By the business, we might think of a function (marketing, sales, legal), a business unit or a region. How do we drive more pipeline? How do we solve retention? How do we increase the average order size? How do we increase the number of customers coming in through referrals? Whatever it might be, it is a business problem. And we know that on the other end, there is an AI or agentic solution to that problem. But there's no one to do that work. The business can't solve it, because they're busy with their day-to-day and they don't have the AI expertise. And on the IT or engineering side, they don't have the time, they don't have the capacity, and they don't have the business knowledge to glue these things together. This is really where the role sits.

So what are four ways you might spot that this is a problem you need to go and identify? How do I know there's a gap between the business and having AI solutions that work?

The first one, which we're seeing a lot of, is a sprawl of skills across the organization. For those of you who use Claude or ChatGPT, you'll have come across skills. They're a way of doing things, and they're very handy. A pre-written prompt would be one way of describing them. Let's think about a sales team, and salespeople desperately trying to prospect and qualify their deals. What we see a lot of is each salesperson building their own prospecting skill and their own qualification skill. Now you've got this sprawl of different skills. Some are quite good, some are quite meh, some do one thing, some do another. No one owns it, and there is no single version of the truth. That's one indication that you need someone at the center of this.

The second: there is a lot going on in the world of AI at the moment. We're going to talk about AI in the News in a moment. All these new features are coming out, loads of things are getting turned on, and no one knows about it. We see this a lot in our clients, with basic things like skills. Skills first came out 12 months ago, and yet we're still explaining to people what a skill is, because no one told them. Again, a good indication that you need this AI Ops person.

The third: usage is flat, and no one is doing anything differently. This is just the killer. We bought everyone Claude licenses, we bought everyone ChatGPT, we rolled out Copilot or Gemini, and nothing has changed. They're just using a little bit of copy and paste. All I know is that people's emails and documents are worse than they were when they were thinking for themselves. So this is a problem. Who is going to fix it? It's not just a training session. You need someone who owns that role.

And the final one: increasingly, loads and loads of pilots that never make it to production. This is more in the pro-user space. Someone learned about an agent, a plugin, a connector or a dashboard, and they've built something that is 70% there, or 60% there. It's just not quite working, and so you've got more of this sprawl around the business. If you speak to the business unit leaders, they're saying, look, we've got a lot of stuff happening, and I can't tell you that we're getting any benefit from it.

I talked about the difference between an AI Ops lead, or an AI Ops person, and the Chief AI Officer. The Chief AI Officer is an executive. This is someone on your executive team, not necessarily on your board, who might sit alongside the CIO or the CTO. They're brought in to think about what this company looks like in three years' time and what our AI strategy is for our products, and so on. The AI Ops person is the one who gets stuff done. They are the doers. You might hear different terms for this. If this person came from a different company, they'd be called a forward-deployed engineer. They might be called an AI business partner. But they are someone who is building skills, building prompts, able to join team meetings, and delivering work and delivering evidence.

Our recommendation depends on your business and your core strategy, but it is to hire the person who is going to do the work first, rather than a figurehead. Not to be too dismissive of a Chief AI Officer, but they come up with slide decks and presentations, and they stand up at the all-hands and talk about how AI is critical to the success of the business. But did they build anything? Probably not. Are they able to point to a metric that has changed in the first 90 days and impacted the business? Probably not. So this is where these AI Ops teams are becoming so important.

So what does an AI Ops lead do? Everything is the short answer. It's one owner for everything to do with AI. That might be platforms. We've rolled out Claude, we've rolled out ChatGPT, so they're helping people understand and build that out and making sure it works. The big one we see at the moment is connectors. If you don't plug your Claude or your ChatGPT into every other system people work in, they are running blind. Companies are having a real challenge approving these connectors. How do I know what the HubSpot connector does? Should I allow it? What permissions should I give people? This is a key thing for AI Ops people to unblock and grease the wheels. We talked about skills: making sure we've got an agreed library of skills and that we continue to improve them. Agents are becoming even more important. If we're going to have 100 agents across our organization, what are they? Who do they work for? What access have they got? How do we improve them as the models improve? And then there's the broader people network when it comes to AI. You might have a Claude Champions program. You might be running training on all the new models and features that come out. This is all in the field of an AI Ops person. If it's to do with getting AI value in the organization and into customers, then this person has got it on their plate.

Where do they sit in the organization? Another reason I like to call them AI Ops is that Ops, operations, means this is not an IT role. In fact, they explicitly should not sit within IT. They are much more focused on the business. When I look across our clients and think about the individuals in this role, they all report up to, typically, the COO, or someone in that part of the organization. They are focused on operations rather than IT. They might peer into IT, which is absolutely fine, and they peer into the business itself, but they're very much cross-functional and focused on the process and operations side of the business. So, considering this is a how-to-hire-your-AI-Ops-lead webinar, it is most useful for someone in an Ops-type role. If you're in IT, then you probably want to bring your COO into the conversation.

What do they do day to day? We've talked about their area of control, but what do they actually do on a Monday morning? Going hunting is how I'd position it to start with. They've got to go and find the work. Across an organization of 5,000 people, there are going to be thousands of potential AI and agentic use cases, but the people closest to the work probably don't know how to turn what they do into an AI use case. So you're seeing these AI Ops leads going off into the business to spend time with the marketing team, with sales, with customer success, with finance, and asking: tell me how you work. What are the top projects you're working on? What are the biggest time sinks? What would really change the way 2027 works for you? Then they build out a backlog of hundreds or thousands of use cases that they can score and prioritize.

Having built the backlog, it's about going and building and shipping these things. That might involve a little light development in Claude Code, but often it is building skills, creating templates, showing people how to use Claude Docs or Claude Slides, and helping people turn on a connector. We can unblock a vast amount of work without anything that requires code. This type of person might also be building a few things. I'm reluctant to use the phrase vibe coding, because it has a negative connotation, that this person is just hacking together things that wouldn't pass a quality check. But it's a phrase that comes up a lot. How do I build a proof of concept? How do I build a dashboard or a quick application using voice, using Claude Code, using artifacts, without having to come from a 20-year computer science background?

The third thing this person is doing, day by day and month by month, is running the estate, the lay of the land. We've got a bit of Claude, some Gemini, maybe Notion AI, some connectors, some skills. How do we make sure it is functioning correctly, that people are aware of it, and that as new models come out we're improving and enhancing those things? That is keeping things running.

Then we've got the people side of things. Say there are 5,000 people in our organization. We need to make sure those 5,000 people understand what's in the estate, how to use it, and how to find skills, agents and templates without creating duplicates. In any organization with turnover of anywhere between 10 and 25%, there are new people coming in through the company every single day. If we ran some training in October, who has joined since October? We need to keep building that domain knowledge up. The models are changing so often that training is never done. It's certainly not a one-off thing.

And then the AI Ops person's role is about keeping on top of things. One way to do that is to join the weekly Kowalah Wednesday Webinar, but other than that, it's following X, formerly known as Twitter, subscribing to the changelogs of Claude and ChatGPT, and building things at the weekend for yourself. Your job is to be right at the front end of this, so that you can support people in the business.

Another visual I think of is the AI Ops lead looking down on the business and asking how they're going to support it. It is not about locking yourself in an IT cupboard and saying, I'm the AI guy, I'm the AI girl, I've learned how to use these things. It's about supporting the business and looking down, so I think about these different dimensions. What are the goals the business is focused on? It is not AI adoption. It is: we need to close the month faster, if we're in finance. We need more pipeline, if we're in sales or marketing. We need to respond to our customers sooner and resolve their issues, if we're in customer support. Those are the OKRs the board is focused on. What are the processes that support those goals? We've got them all listed here. These are the processes the teams, and the executives running those teams, are focused on. Then we build some AI to support that (agents, skills, templates and so on), and those plug back into all of the IT systems our colleagues use to run those processes. It's really this top-down lens that I think is so powerful for AI Ops people to have.

The final thing in what they do is the cadence of the schedule, which will depend on your organization. Every day, you're keeping on top of the estate and what is happening, making sure we haven't bankrupted the company by people using the wrong models for inane tasks. Every week, it might be looking at the backlog, checking any new releases, or joining the Kowalah Wednesday Webinar: how do I keep on top of things? Each month, it might be launching new training. What do we need to teach people? What new use cases have we launched? Have we recorded any new outcomes or ROI that we should talk about at an all-hands? On a quarterly basis, it's more about your OKRs as an AI Ops lead, reporting up to your senior leader. Cutting across all of that are the things that don't happen on a regular schedule: new connectors, new agents, new skills, new features coming. It's whack-a-mole, keeping on top of all of these as they happen, live.

So that's a bit about the role, what they are, and how they're different from the executive Chief AI Officer. Now, how do you hire one? If you're the chief operating officer and you've said, right, I need to go and get an AI Ops lead, how are you going to do it? The first thing I'd think about is what you're looking for. I've talked a little about the character of these individuals, but the ones I have in mind are builders. They've built things, either before in a different role or for themselves at the weekend, and they're happy blending between the technology side of things and the business side of things. They're not just a deep engineer who says, I know how to write code, I can do TypeScript, I can build test cases. That's not what we want. We want someone with enough of that, who can then go and spend time with the business and really understand a business problem.

This is slightly tongue-in-cheek, but not. If you want a picture of who you're looking to hire: top left is business acumen. You can go and sit with the sales team, or with the CMO, and understand their objectives, their business processes and what they're trying to do. Top right, you can build something, maybe with Claude Code, maybe by building an artifact, and bring it to life. And the third is that you don't know you're worth $350,000 a year. Slightly tongue-in-cheek, but not, because this role is being hired for left, right and center by companies that would call it a forward-deployed engineer. You go into a bank, an insurance company or a large retailer, you sit with the business and understand how it works, and then you build the solution that drives revenue or cuts costs. That is worth a ton of money. So, tongue-in-cheek, but not really. The more this role develops, the more highly sought after it's going to be.

What kind of things might you ask in an interview? This is my go-to: show me something you've built with AI that other people use. I don't want to know what you might build. I don't want to know how you think you might build it. I don't want to see some slides. What have you built? If you love this stuff as much as I love it, you have built something. It might be a dog-walking app for your three dogs. It might be that your kids play in a sports team and you've built a scheduling app. It might be that you're organizing a party and you've built a little skill to help people plan what they want to eat. But show me something. And if your answer is, I haven't built anything, then you're not right for this. We want someone who has agency, ironically, who has done something themselves. So: show me something you built.

Once that person has come into the interview process, maybe in their second interview where you're spending a bit more time, get them to build something. Come up with a use case, a scenario or a process that is real in your business. Maybe it uses some access to your data, maybe it's a hypothetical situation. I'd say: here's a scenario. Our marketing team is focused on driving new business pipeline, and they're focused on doing that through events. I want you to build out an agent that will help us plan our event series next year. We're going to spend 45 minutes now. You're going to ask me some questions, and I'm going to tell you some answers. I'll act as the customer, and you're going to build something for me. This shows you whether the person can talk business, talk process, understand the OKRs (the objectives and the key results), and translate that into the tools they're familiar with and actually build something. Again, if someone crumbles during this, they're not going to be the right person you need.

You may have your own way of measuring in hiring, a different rubric, but here are five areas I'd be thinking about. One: can they build? We talked about that. Can they actually build something in front of you, and what tools are they going to use? Two: the process. This is the business side of things. Can they talk to me about the business, the process, the team, the objective? Why does this team exist? What are we aiming for? That is going to be important. Three: technical range. We're not looking for someone with a computer science degree here, but someone who can open up Claude Code in Claude Desktop, maybe talk about Lovable or Vercel, talk about a database, and has some experience of building a managed agent, or at least can talk about it. If they don't understand what any of these terms mean, they're not going to be able to bridge the gap between the business and IT and engineers. Four: can they stand up in front of the executive team, the CMO, the chief revenue officer, the chief customer officer, the chief financial officer, and talk about the processes and how this business runs? A lot of what they're going to be building is about what this company looks like in the future. They've got to feel comfortable having that conversation, not hiding away in a bunker somewhere. And five: judgment and taste. This is the human element. This is not a computer game. You cannot just build a backlog of 100 and then fire Claude at it and deliver. We need human taste and judgment in all of our roles, and it's going to be so important for this person moving forwards.

I was listening to a podcast a couple of weeks ago, my favorite podcast, Lenny's Podcast. Shout out to Lenny Rachitsky; I definitely recommend it for anyone working in the world of AI. A couple of weeks ago, he had Josh Woodward on. Josh runs Google Labs, and Google Labs handles Gemini but also a lot of the side projects they experiment in. The question Lenny asked him was: for Google Labs, how do you think about hiring? How do you find people who would be a good fit? And Josh said that up until recently, his main question was how quickly you learn stuff. He was looking for people who could learn really, really quickly, because that was valuable. But in this world of AI, and how fast things are moving, he's now looking at how quickly people can unlearn something. If someone is going to hold dear to something they learned six weeks ago, that's no good. It's out of date. So he asks: how fast can they learn something and then walk away from it? This is something we think about a lot at Kowalah. What are we throwing away that we held really dear just a couple of weeks ago?

The final thing on how to hire: once this person has taken the role and signed their contract, how do we set them up so they can be successful? We want this person to come into a defined role they can get to work in. They've got to have a sponsor sitting above them, maybe the COO, maybe the CEO, someone who can say, I can walk you into any meeting, I can tell people they need to speak to this person, and they're going to pay attention to you. So it's got to be a senior named sponsor. Then a budget they can control. This doesn't necessarily have to be a P&L budget. This person isn't necessarily going to be going off and paying for the tokens. But they need to be able to spin up new technology if they need it, and they probably need some budget to spend on third parties, maybe other consultants they need to bring in. Maybe it's a use case budget: you can build five use cases a month, whatever it is. But they've got to have that available to them. And then I'll jump on to the fourth one here, about being focused on the business's top priorities. As they walk in the door, they need to be told: the top three things that are going to change the direction of this company in the next 12 months are A, B and C, and that is what I'm going to point you towards. If we bring in this really high-value person and point them towards, I don't know, some internal HR use case that isn't going to change the direction of the business, this person is just not going to be able to stand on a pedestal in six months' time and say, look at how we've changed the direction of the business. So, if you haven't got these things, if there's no mandate and no high-value use case, I would hold on and not hire that person quite yet.

Coming towards the end of this section of today's webinar: how do you know it's working? How do we get this person standing on a pedestal in six months' time, when everyone goes, look at John, look at Helena, look what they have done and how it is impacting the business? For that first 90 days, let's get them up and running first. Don't point them straight at the first use case and say go build, because suddenly they become really, really tactical. In the first 30 days, I'd want them to map out the business. How does this organization run, from marketing to sales to product development to customer onboarding to service, and then our internal support functions? What does this company look like? Where is value accrued? What are the leverage points that, if we get this right and unblock them, are going to dramatically change the direction of the business? Through this, they can come up with the two or three domains, process areas or customer journeys where there is high value, and say, right, that is where I'm now going to embed myself and build out a backlog of use cases in that domain. Having picked those two or three domains, let's get to work. This is not a job where we're creating slide decks and theories. Let's build agents, let's build skills, and start driving value into the organization. And then thirdly, coming towards the end of your first quarter, how do we start evidencing this? What was the metric? What was the number? What was the KPI? And how have we changed it? I would love to see this person at the all-hands, at the end of their first quarter, standing up and saying: this is what I have done in 90 days. If I can do this in 90 days, imagine what I can do in two quarters, in a year. Let's get to work.

I talk a lot about OKRs. You may not have come across them. It stands for Objective and Key Result. Let me go to one of my favorite books over here. OKRs come originally from Intel, the chip and memory manufacturer, and they were first defined in this book from 1983 by Andy Grove [High Output Management]. They were then popularized at Google, and there's a good book called Measure What Matters, which is all about OKRs. There are two parts to it. An objective: where are we going? And a key result: how will we know that we've got there? The objective is the goal and the key result is the measure. For your AI Ops lead, you may give them a couple of objectives linked to the objectives of the business units they're working in. For example, AI is part of how sales gets work done, and some key results could be the number of proposals that go out written by agents, the speed to lead for responding to an inbound, or the percentage of your deals that get auto-qualified. We want the objective to be about the business unit and the outcome, not the number of people who logged into Claude, or the number of people who attended training over the last few weeks. That does not change the direction of the business.

What might good look like? It could be the number of things you have delivered: skills, templates, agents. It could be the number of people using the things you built, and we definitely don't want to build things people aren't using. But what we really want to focus on is the key result for the actual business unit. The example here is a finance team that has gone from 9 days to close the books to 5, and therefore is able to make better decisions and so on in our organization.

That brings us to the end of our section on AI operations leads and hiring them. We've got a ton of extra content on the website, which you can use. Let me show you. On the Kowalah website, you go to Resources, then Templates, and you'll come to this page and find a lot of these things. If I go to AI Operations, you'll find a lot of templates that are useful for someone in an AI Ops role. AI Operations Lead First 90 Days, if you are that person or you're hiring that person. Each of these templates includes a skill that you can download and put into your Claude or your ChatGPT, and it will help you with that first 90 days. Or you can chat about it in your own Claude. Let me open that in Claude. It brings the whole template to life, so you can chat about it in your own Claude and plan your first 90 days as well. There you go. I've read the full page; here's how the first 90 days works. So those are super cool. If you're the one hiring an AI operations lead, let me go up here, to HR and Talent. We've got the AI Operations Lead Job Description and Hiring Guide. Exactly the same thing: here's a skill, and the template, and all of the details that you can either chuck into your own Claude, or use the skill so that you can plan and hire that role. Interviews, screening, the build exercise, everything we've covered today. So, knock yourself out on that.

So, your action today. Do you need one? We went right off at the start: do you need one of these? If you don't, then don't waste your time or the time of the person you're going to hire. Make sure you're hiring the doer, not some executive who's going to create some fancy spreadsheets for you. Give them the right responsibility and permission, so they're not a powerless person. And then use the templates I've shared with you, so you can get on and hire that person effectively.

Now, AI in the News, our recurring segment every week, and some really good stuff this week, which I can't wait to show you. The first one was a couple of days ago. It's the 11th today, and this was on the 9th: the launch of two new frontier models from Anthropic, Fable 5.2 and Mythos 5.2, bumping them up from 5.1. The one you're most likely to come across is Fable 5.2, because Mythos is locked away for only certain specific companies, based in cybersecurity, because it's too powerful for you and for I. Fable 5.2 is the one you're most likely to come across.

As I always say on a new release, benchmarks are benchmarks. They get better every time a model comes out. So there's not too much you need to worry about there, other than that they are getting better and more powerful. What does that mean? It means you can start to tackle more challenging, longer-running, more complex, more detailed, more futuristic work. For the majority of your employees, they don't need Fable 5.2. They don't even need to know it exists. I was chatting with a client just the other day. They'd uncovered one of their users using Fable 5.1 to read their emails and their inbox five times a day and tell them if there was something they'd missed. This is not a use for Fable 5.2. Fable 5.2 is for when we're thinking about the industry and the company we're going to be running in three years' time. We want to build out a super detailed strategic plan, thinking about all of the competitive pressures, and the markets, and the research and the analysis. This is where Fable 5.2 comes in.

I'm going to share this in the notes, and I'll give you the link. There's a great person to follow called Ethan Mollick. He writes well on X and on LinkedIn, all about all things AI, and he has a blog called One Useful Thing. This is one shot with Fable 5.2. His prompt was: I need you to create the most impressive thing I can share with people who want to understand how Fable 5.2 works. I'll share the link so you don't have to watch the 1 minute 30, but this is just amazing. [Video plays.] This is a single shot where Fable is creating a video to explain how amazing Fable is. And it's super cool for a single shot, when you think you might have gone to an agency to create a video like this. So: know that it exists, don't use it for everything, think of your most challenging problems, and that's what I'd be pointing Fable 5.2 at.

Now, something you really can get your hands on. Regular viewers of our webinar will know that I do all my slide work here in Claude Slides. I don't use Google Slides anymore. I'm creating all of these decks through my conversations with Claude, pulling in our call transcripts, our CRM and our website, so it's really personalized, but I'm not building slides anymore. It's just not a good use of my time. What was launched yesterday was Claude Sheets and Claude Whiteboards. Let me show you what I mean. So, here I am in Claude. On my sidebar I have Artifacts, and if I click on plus, I've got Docs, Slides, Design, and now Sheets, Data (which was called Dashboards until a couple of days ago), Whiteboard (we can have a look at that in a minute), and then Motion. If I click on Sheets, we've now got a sheet. We do not need to use Google Sheets or Excel anymore. I won't go through an example here, but I could have the chat here and say whatever I want, and we'll be able to work in Sheets here. So whenever I think now about building out a budget, a planner, a project, any kind of P&L, whatever it might be, I'm just going to be starting in Claude, because Claude has all of the access to the other information I would need, rather than me manually working from within Google Sheets.

Next up is artifacts plus Whiteboard. All right, now we can all learn together, because I've not done this myself yet. What do I want to map out? Hey Claude, we are currently on the Wednesday Webinar, and we are doing today's section on AI operations leads and hiring. You've got access to the deck because it's in another artifact. Can you create a whiteboard that brings to life everything we talked about in the AI operations hiring section? Bad question, but I would absolutely imagine that if you're having a workshop or a planning session with your team, you're going to have Claude up on the big screen and have Claude build out your thinking as you map stuff out. One thing I'm very often doing is taking pictures of the actual whiteboard where I've been doodling, or a flip chart, and copying that in, so I'd imagine doing that. There's a company called Miro, which you may have used for digital whiteboarding. I'm making up the numbers here, but I think it was previously valued at tens of billions, and it was just bought for less than $2 billion, I think, by a company called Bending Spoons. An entire market is removed by tools like this that are turning it into a feature. I'll let that carry on. I'm excited to see what it builds out. But within Claude you now have Docs, Slides and Sheets, so I can't see myself going to Google anymore, and now these Whiteboards, and the Motion videos as well.

Just conscious of our time. This has not been released yet, but it could be a potential Granola killer, or a Wispr Flow killer, where you're talking about call transcription. It has been announced but not released: Claude will be able to be invited to your Google Meet, as you would do with Granola, and that goes straight back into your Claude, so that you can start to have discussions about it. I'm interested to see this. I can see Claude and ChatGPT getting more and more into enterprise use cases.

Final thing, I think, on AI in the News: yesterday, Google launched its top-rated model, called Gemini 4 Carbon. It's not up at the level of Fable 5.2, but if you are a Google and Gemini user, then you've got a better model, and depending on what your account is with Google, you've either got this included in your enterprise license. Why wouldn't you use it? My sense, and we use Google Workspace here, is that because the application around these models is not at the level where ChatGPT and Claude are, it doesn't matter how good the model is. The model is not the limiting factor, and it hasn't been for well over a year. It's the applications. If I'm building my slides, my sheets, my demos and everything in a tool like Claude, I don't see myself going back to Gemini anytime soon.

A few things are coming up if you're an admin, and I'll point out a couple of them. Specifically, if you're on Enterprise, Chat and Cowork are being merged together. You may have turned this on already. If I come out here and go to Claude, New, we've got it turned on here, so there is no separate Chat and Cowork option. If you haven't got it turned on, it will say Chat and Cowork. By the 16th, so one week's time, that will just be the default.

Oh, hiring an AI operations lead. So this is our whiteboard. That has come up. Hiring an AI operations lead: the business has got a problem, and the AI operations lead owns it in between. Do you need one? What they do, how to hire, how to measure. So there you go. That is a whiteboard, and we could do some more fun things with that. That's nice.

Okie dokie. Anything else? I haven't got a slide for this, but let me find it now. ChatGPT business registration. They just launched this last night. As a business, you can now claim your business, as you might do on Google for a Google listing. The idea is that you can start to own your listing and update it, and it is going to help you get found in ChatGPT. Even if you're not ChatGPT users yourselves, your customers are, so get on and Google that, or ask ChatGPT, OpenAI.com, to search for Business in ChatGPT. That was last night, so it's worth taking a look at.

Alrighty. I haven't seen any other questions come through live, so no worries about that. If you're watching on YouTube, you can feel free to send them through to us, and we'll come back to you. Next week, we are kicking off a series of Wednesday Webinars really focused on the marketing team: the CMO, marketing leaders, event marketing, field marketing and pipeline generation. We're going to talk about building out your design system in Claude, a good starting point for everything like we've been looking at today with these decks. So we'll see you at the same time next week, 3pm UK time, 10am Eastern, and we'll see you then. Thanks for joining.

---

This transcript is published by Kowalah, a UK Anthropic Implementation Specialist.
Webinar page: https://kowalah.com/resources/wednesday-webinars/who-owns-ai-inside-your-company
All transcripts: https://kowalah.com/resources/wednesday-webinars/transcripts