Navigating the AI revolution with Sitecore

Note: Video transcript has been edited to provide a summary of this episode.

Sean Broderick (Sean): Hi everybody. Welcome to Digital Visionaries. My name is Sean Broderick. Today I'm delighted to be joined by Ru Barry, who's the Senior Director of Product Management at Sitecore. Ru is one of the foremost AI thinkers in Ireland today. He brings a great knowledge of product management, [and] a great sense of where the market is going. Ru, thanks for joining me today.

Ruadhán (Ru) Barry: Thanks for having me, Sean. Lovely to be here.

Sean: It's great to be here in Dublin recording this. Let's take it back a couple of months, back to Nashville and Sitecore Symposium where we launched Sitecore Stream. Tell us a little bit about that launch and the impact that's made on Sitecore.

Ru Barry: I think it's been big. We needed to have that point of view around Stream. We needed to be able to communicate very clearly, and that takes a lot of work in the background. Launching Nestle there as well was important for us, having co-developed a lot of those capabilities. It's great to have that as a turning point for us from a product perspective, a marketing perspective, and a wider business perspective. We got a real chance to put our mark out and say ‘this is where we're going. This is where we believe the market is going.’

Sean: I think you saw the splash that was made in Nashville, the kind of impact that it's had online, and how much it resonated with marketing teams. Tell us a little bit more about that vision: what was launched in November and how that evolves into 2025 and that winter release for Sitecore.

Ru Barry: Absolutely. We said on the stage that Stream is three things. Stream is brand-aware AI, it is generative co-pilots, and it is AI-enhanced workflows. What do those mean? For anyone who hasn't seen the demos or the collateral, brand-aware AI is about supporting our customers to generate content and enabling them to generate content that is aligned to their brand.

Users can create a brand kit, which is the foundation of knowledge for content creation. It provides a safe space for the user of the AI, the corporation, and our customers to be able to say, ‘you know what? People who are building AI within our business? That's fully aligned to our brand.’ Brand-aware AI is important. That'll evolve over time - right now it's very focused on brand because most of our customers are in the CMS space.

The second one is generative co-pilots. These are everywhere, all over the place in the industry. And they will continue to be. We've spent a lot of time thinking about our end users and how we can provide AI that enables them to be efficient and collaborative in the work that they're doing. The co-pilots are there interspersed across all the Sitecore products. Whether you're a Content Hub customer, an XM Cloud customer, XP, XM, CDP, Personalize, Search. There will be copilots available to your users to help them.

The first [co-pilot] that we released was the brand assistant with Nestle. As I said, they helped us co-develop that. That's an important first co-pilot for us, because it allows our users to interact with chat and really be able to ground that chat in the details of the brand. [It] helps them reference details, make sure things are in the right tone of voice, that the content created is aligned to the content that their business wants to create.

The third was AI-enhanced workflows. We had a demo for anyone who was at Symposium and wanted to get hands-on experience. The winter release will be the first general availability announcement of the capabilities. AI-enhanced workflows are about enabling marketers and technologists to be able to build, collaborate and deliver in Sitecore. The first part of that will be around creating tasks, being able to create projects, create campaigns, collaborate, comment. It'll be about understanding end-to-end content delivery and marketing plans and how you're going to manage those across products. But eventually that'll expand over time. Today, a lot of those tasks are done by humans. Over time we do see agentic AI taking over. You'll start to see humans approving agentic workflows.

An agentic is essentially an AI capability that is doing a piece of work and coming back to you and saying, ‘this is what I've done. Is that okay? Can we move on?’ It's the ability for humans to offload some of those redundant tasks to an LLM, an agent, a co-pilot. Eventually I think those will chain together and we'll have maybe workflows that are being done by agents for content creation, aligning that content to your brand make sure the tone of voice is correct, and then coming back to the human and saying, ‘Here's the version, here's three versions. Does this work? Do you like it or not?’

Stream is those three things.

Sean: Let's dive into the brand-aware AI piece, because for me that is crucial to this. As a marketer myself, you look at what Chat GPT and different LLMs and Perplexity provide you and in the main, it's getting better, but it's homogenous content. It doesn't really reflect your brand, and it's the wild west out there. It's not really that safe. You have people throwing reams of corporate data into these LLMs, and that is being used to train these models.

Talk about that security aspect where you upload your brand kits, all your brand heritage, all your content, and your knowledge into this brand-aware AI and the security that gives you as a brand?

Ru Barry: I think one of the main roadblocks to deploying AI in enterprise today is trust and trusted outcomes. Brand-aware AI is important. It grounds the models. When you're rolling out an AI into your organization, you want to be able to understand what people are doing and where the source material is.

The source material is critical. Sometimes, I'll use Chat GPT, or a copilot, and I'll ask it to give me some data points. It will give me stats, and then I'll say, ‘where did that stat come from?’ And it'll say, ‘oh, that’s a hypothetical stat. I'm sorry about that. I apologize.’ That's not acceptable for enterprise level content creation.

Sitecore will never train an LLM model on your data. Your data is your data. That's critical for all our customers. They don't want their brand guidelines, their internal confidential data, being used in the open. That's not acceptable for them. Your brand data doesn't leave Sitecore. It doesn't go anywhere else. It's not shared with any other customer. It's not available on the open internet. It’s...for your use only. I think that’s a big step forward for trust. Once customers understand that we're not there to harvest that data, we're not there to use that data for our own purposes.

Sean: What are some of those copilots that are going to be specific to the people interested in XM Cloud or Content Hub? What are those things that are going to augment those existing Sitecore products today and make them attractive to our existing base and prospects?

Ru Barry: The first release will be the initial steps. Some of those will be around content creation and content alignment. Then you’ll see code block creation, full variant creation, hypothesis and AB testing as part of that expansion to meet XM Cloud users where they are. Enable them to build components, forms, pages, sites and really start to think about the user experience. I think you had Liz here recently talking about the future of CMS. Liz is super focused on how we can enable those users.

Sean: The type of enterprises and the type of companies that Sitecore are working with in this space, the volume of data that they have is insane. They need to be able to work fast and to utilize all those existing brand assets and reference those when it comes to creating new material and new assets.

Ru Barry: I think the important part of Stream is content creation and content adherence. But the other part of that is collaboration. We’ve focused on that: how do we bring teams together to work better? Most of the time, we're on a laptop in our homes or we're dispersed. Those collaboration features built into the products critical.

Sean: Let's talk about the kind of conversations that you are having with some of our existing customers today. What kind of reaction are you getting to the new features that are being developed inside of Sitecore Stream at Sitecore?

Ru Barry: Early access has been open since Symposium. We have customers and partners jumping all over brand assistant and the brand assistant chat interface. What we showed at Symposium has evolved into this winter release. You're going to see more collaborative working spaces within the chat.

Most of the feedback has been around the ease of implementation. We've been talking to partners and customers and see the blogs they're creating now that are brand-aware, that are aligned to their brand. We have case studies being generated [that show] real value in that content creation space. When we think about making efficiency gains for those customers it's critical and they've been excited.

Sean: What are our customers doing to get ready for AI?

Ru Barry: That's been one of the interesting aspects of early access. People want to get their hands on AI, and they want to get involved in it. They come up against internal blockers around AI adoption. They haven't done the groundwork on policies. They don't have governance in place. They don't have a company policy on how they're going to approach AI. A lot of those customers are now starting to think about those things.

We’ve worked hard on making sure that the adoption of Sitecore Stream is as easy as possible.

The second thing is investment. Not investment in AI, but investment in people. A lot of people will talk about AI but not a lot of people are curious enough about it, and even more people aren't even enabled on it. Investing in people is critical. [AI]’s not going anywhere. The last thing is investment in the infrastructure and that comes once you've chosen the platform.

Sean: If we turn the camera a little bit towards the back end of 2025, what does that vision look like for Sitecore Stream in the future? How is this going to evolve over the course of the next 12 or 18 months, in your view?

Ru Barry: I think there will be a tipping point at some point this year - work will be fundamentally different than it's today. Marketers and those technologists need the power to build. We can create space for marketing teams to do the things that humans do well. Humans love human connection. They wanted crafted storytelling. They want impressive experiences. That can only be delivered right now by human thought, human creativity. To make that space, we need to remove the redundant tasks. The majority of think of marketers today, they spend probably about 80% of their time on approvals and publishing and maybe 20% on the creative part. When we talk to customers, they're looking for that fun element. They want to be on whiteboards, they want to brainstorm. They want to be talking about their brand and what it means and what the ideas are and then be able to execute much quicker.

Sean: You and I have been working in the marketing space long enough to know that the promise of personalization has always been out there and talked about for many years now. To be able to create these variances for different audiences, that's where you're going to see a massive boost in conversion.

Ru Barry: For I think for the past four years we've been talking about this concept of marketing technologists being the real linchpin of a highly efficient marketing team. And one of the things that I'm quite excited about is code creation. That really changes the game. The enablement of those marketers and those technologists to build with Sitecore is a huge opportunity and something that I'm particularly passionate about.

Sean: Ru, it's been amazing talking to you this morning. Thanks so much for joining me today.

Ru Barry: Thanks very much for having me, Sean.