In this episode of Good Morning AI, host Serena Cho interviews Noelle Russell, Director of Product Management for Search at Elastic. Noelle shares her career path from a long‑time developer to a product leader, giving her a unique view of how search technology has evolved. She explains how Elastic’s early focus on highly opinionated enterprise keyword search provided strong controllability and explainability, but the rise of generative AI and large language models (LLMs) has shifted conversations toward richer, semantic experiences. Noelle discusses the trade‑offs of introducing LLMs—loss of visibility versus added flexibility—and how Elastic is helping customers blend traditional keyword interfaces with AI‑driven chat, embeddings, and multimodal data. Real‑world customer examples illustrate challenges of redesigning workflows, A/B testing chat overlays, and responsibly labeling AI‑generated content. Looking ahead, Noelle envisions a democratized search future where voice, avatar, and multimodal interfaces are easy to add without massive re‑architecting, making powerful retrieval accessible to organizations of all sizes. The conversation closes with optimism about making search more intuitive, accessible, and universally usable.
Chapters
Hi everyone, and welcome to today's episode. We are so excited to talk to you about enterprise search and search capability. I know a lot of the companies that I am talking to are trying to figure out like, how does AI change the world of search? And so I happen to know a couple of people, uh, that understand the space really well, and one of them is my now very good friend Serena Cho. She is the Director of Product Management for Search at Elastic, and I could not be more excited to have you here chatting with us today. I thought it'd be cool for us to start a little bit about like you and, and who you are and what you do and kind of how you got into this space, but also maybe what you've noticed about like the beginning, maybe 2, 3, 5 years ago, like how has search kind of uncovered itself in your career?
Noelle Russell: Noelle, thank you so much for having me and coming to talk to you about this topic that I'm very passionate about. Just to kind of give you a sense of my background, I started my career actually as a developer a very, very long time ago, more years than I'd like to admit. And it's given me a unique kind of perspective into the changes that are happening in search today. Because if you start from where you know the technology, where it can be from a software perspective, as you kind of move into a product management role, which is my craft of choice, where I have found my most passion and enjoyment in my career today. And I especially enjoy being part of the product team here at Elastic because we are at the center of a lot of changes here. Search, to more directly answer your question, search has just been a pivotal part of so many applications and development experiences, ranging from, you know, being critical to your retail experience. As a personal shopper myself, I enjoy the benefits there. Maybe a little too much. But with the advent over the last 2 years of more interesting technologies that can bring a lot more semantic meaning, it felt like for many people out there that this was something brand new. You know, you have people who come talk to me because they know that I'm in technology, that I'm in this space, and they're like, oh my goodness, is Skynet here? Do I have to worry about the robot revolution? Do I need to build that bunker? And I get to have this moment where I get to tell them it's not a scary thing. This technology is something that can be used for good in so many places to enhance our life, from new research, from accelerating our day-to-day activities. And that's really surfacing in an interesting way at search and at Elastic, because for us, this was not actually the meteor event that was going to kill search that people posited was the case 2 years ago. And in fact, this was This was a place where Elastic was positioned to really step into this space and show what we can do, because we were already there investing in the capabilities that you need to serve the problems that are surfacing here, to build the next generation of search experiences, to change your retail, to change the way that you scavenge and try and discover insights in existing datasets, and use that in an enterprise way. This has just been a very exciting time for us, and I'm thrilled to be part of the product team. To kind of go from individual developer all the way to the product team and be able to see this new evolution and new space kind of unfold for us.
Yeah, it's incredible. It gives you such a unique perspective that I don't think a lot of people have, like starting from like, how do I build this? To now, how do, how do I build a solution with this technology? To how do I actually build the tech to provide those solutions? I think it's such an interesting journey you've been on. One of the things I'm very curious about is kind of like what you've seen change over the last year and a right? We've, we've seen this advent of generative AI and everyone's talking about generative AI. Like, how has the, the conversation around search shifted since, you know, you, you maybe started or, or maybe just since before the generative AI kind of big announcement and, and the big changes in the conversation? What are some of the use cases or maybe some of the different conversations you're having around what search is now today?
Great question. Um, 2 years ago, uh, Elastic was building highly opinionated enterprise search solutions, uh, because we believed in the opportunity to accelerate workflows for internal workplace, uh, organizations. And so we had some product thinking that was definitely part of the conversation. It lived side by side with what people assumed was our only capability, which was to really provide great capabilities for keyword search. Everyone knows what a keyword search is. We have a whole generation of people that are trained to do keyword search like black shirt or, you know, blue toes, you know, whatever that search experience is. Not sure why we went to the toes in that particular example, but here we are. The search experience conversation has evolved though, because during that time, we saw with the opinionated enterprise search experience, the keyword searches that were being built as just kind of the golden standard for how you should access and retrieve data really had the benefits of a lot of controllability and a lot of explainability. You could pretty easily take a look at that search experience and say, oh, here's my highlighted thing, it matches my keyword. The relationship is so simple and it seems so intuitive, but then the conversation, as you start to think about, wow, I really want this magical large language model to come in and simplify this and maybe make it a little less hard for me to take all of these different levers and different techniques that are available in this search space and maybe just put a little magic in there and The conversation has really centered around, it's actually not that simple. You can't throw a large language model and expect magical search experiences to come out the other end. And you will also lose some of that visibility, that controllability that you had with keyword search. So a lot of the education over the last few years and the conversations that we've been having as people take a look and take stock at, you know, the capabilities that they can have ranging from that large language model that they want to apply to, really unveiling the full buffet of machine learning capabilities out there. There's embedding models, there's natural language task types, there's all sorts of things that can be brought to bear to evolve your search experience past like previous expectations when they were just thinking in terms of a keyword search.
I, I love that because I remember, I don't know, maybe 5, 6 years ago where I would go to a company and I would completely dismiss their search capability. Like even if I was inside a company like search and I wanted to find something on their own website, rather than look inside their own website, I would go off and do a keyword search on Google. And I saw us kind of evolve past that. And I'm curious, like, do you have any specific examples of companies that kind of changed their view of their own search and how they maybe implemented that search? Like today, you know, Because I think like maybe it's become a conversation that a lot of people are like, wow, well, our enterprise search is just awful. We can't find things in our own company. That's— it was just kind of, we just, we accepted it. I'm just wondering if you've seen that tide turn a little bit and if there's any specific customers or, or case studies that you've worked with that show how a company might like unlock, right? All this information inside their own organization through a better search capability.
Oh, I love this question because there are so many customers who have done such interesting pivots to their search experience. It's funny because I love what you said about having an opinion about a search experience, because everyone, even if you don't realize it today, you have an opinion about your search experience. And it's deeply ingrained in us that if you can't find something within a short amount of time, then clearly the search experience is terrible and you will go wherever the fastest path is. And our attention span is shortening. Our threshold for the user experience in not only just accessing the data quickly, but to put it in a way that's intuitive and human, human readable to any user has increased dramatically, especially with something that people get excited about, like a chat experience. This starts to become the experience that more companies are wanting to incorporate. But the challenge that we see with some customers who try and evolve that is that it's not a perfect fit for the ergonomics of their website. They have to do a fair amount of redesign for the workflows in their products or search bar or even their analytics engine if they're using that. And we have some interesting, like, customers who have like just your everyday run-of-the-mill, probably traditional search experience, which still has a very valuable place because can you imagine a world where you're trying to shop in a retail environment and you're only presented with one choice? The people who are trying to sell and upsell, you know, side things that are related to that would scream at this type of user experience. So there is a place for traditional constructs and designs of search and we have, For instance, one customer who has a very interesting product catalog experience— it doesn't have to be e-commerce, but just generally a catalog of things. It could be a catalog of knowledge, a catalog of classes, whatever that catalog of things is. They have a standard for how they want to be able to maintain the preciseness of the search. They have a current user base that's trained to access data very quickly through that existing one, and they want to uplevel that without damaging that high standard that they already have for quick access and the filtering without gigantic leaps of changes in the user experience. So the challenge that they've been incorporating there is to work with their design team, maybe pull in a chat experience on the side, maybe present like a chat experience almost as an A/B test side by side with that experience. I've seen some customers choose to anchor like AI-generated content and also be very explicit about showing that this is AI-generated, to kind of COVID their bases in responsible use there. And the choices that they're making as they evolve these projects range from working on a product, project catalog to unveiling new types of use cases, because maybe they started with logs with us, maybe they realize that they can bring in more and do more with the current architecture. And so they're able to bring in new sources of data, new types of unstructured content, image, audio, and start to evolve the user to incorporate that in a really interesting way.
Yes, it's so fascinating. It actually leads me to kind of think about, I mean, there are always these clients and customers, right, that are thinking about kind of the future now, and they're pushing the edge of what we think about with search. I know like in my house we have hundreds of voice-enabled devices, AI devices. It, it helps, like I have a son with special needs. I have a, my dad who's aging and lives with me. So we try to create as much accessibility as possible, but what, I'm just thinking about the future, right, of search, both from an inside a company. I mean, I definitely have the dream of one day going into my own company and being able to find what I need when I need it. That's like very, maybe, I don't know, it seems like it was a hard problem to solve just a few years ago. But I also think about kind of customer experiences like you're talking about, like what, what a better experience to offer to just be able to ask for what you want as opposed to know what keyword to say, know how to say it, right? What if I could just say, you know, hey, I would like to find a watch. And what about if I could just say, hey, I wanna find a watch that looks like this, right? How specific? So I'm just curious, like, you know, as you think about search and enterprise search and even customer experiences, like what are some of the forward-looking things that you are either seeing customers lean into or that you're just dreaming about from an experience perspective? How does search change in the next 3 to 5 years?
Oh, there's— you're kind of talking to a PM and saying, here's a blank check, like, what can you— and I could spend hours and days to just give you the list here. We have a lot of really interesting targets and goals that we have in our hearts and minds here at Elastic. And the predominant one is that I think we want it to be easy. For people to be able to expand out, add accessibility. It's important to democratize access here, not just from an accessibility point of view in the user experience, but also just the fact that there's, there's definitely a gap in access between like folks who have the capital, the research, the staffing, the opportunity to leverage some of the research that's coming out there from these larger players. But there's also a really wonderful market of vendors who are coming up with some just astounding innovations in information and retrieval. And so that intuitive access and that, that cause of democratizing this is in service of saying you shouldn't have to re-architect everything, reinvent the wheel again and again. You should have a purpose-built foundational pillar that unlocks your ability to add not just a text traditional search bar, but to be able to build an experience that can be voice activated, that can be leveraged through an avatar, and really just promote equal unified access to all of the capabilities in a way that's not going to break the bank.
Yeah, I think it's so— I mean, it's so fun to kind of think about what giving the world better access to finding what they need at the right time and in the right context. And this is just the beginning, like understanding— I love the conversation. Thank you so much for being here. I think it's so interesting to just think about search and maybe even revisit search for a lot of us that have kind of given up on having a good search experience either inside our companies or even for our users or our customers. But We of course kind of took a high-level view of search. I think it's so important that we actually talk about kind of the more practical and maybe even some of the more difficult challenges that come with implementing a good search experience. And that's what we're gonna do next time. But I am so excited that you were here. Thank you, Serena, for spending some time with us, and I cannot wait till our next session.
