Welcome. I'm Diane Burley, senior director of content here at Coveo. Today, we're going to not only give a sneak peek into our twenty twenty three tech survey on search and AI that we just conducted on enterprise search. But we're gonna hear from noted expert Isaac Sikolic on a couple of our findings and have him share how, hopefully, you can rethink AI search for digital transformation. Today, we'll be going through, the following topics. So I think it's gonna be, a really interesting, broadcast. A little background. Annually, we commission Dimension Research to survey more than six hundred tech professionals like yourself. We qualify our participants by making sure they really are an enterprise search and not just some dude in the garage with his dog. Like last year, we got some very strong affirmations that search is still really hard and that there are a number of things that are keeping you up at night, data integration from multiple silos, dealing with different data formats, including unstructured security issues, and management not always understanding what you're up against. In fact, ninety nine percent of you said you're still manually tuning. Problem with commodity search is it's easy to do some things, but more difficult to do others. So I am really excited here to, introduce, longtime friend, search expert, author, and the founder of Scar CIO, Isaac Sacolick. Welcome, Isaac. How are you doing? I'm doing great, Diane. It's great to be here. One of the fascinating results from our survey, Isaac, was that eighty one percent of, BEMC search as critical to driving digital transformation. Can you share why that might be the case? And I know that sort of differed from some research you did with senior executives in the spring. Yeah. No. It's a good question. I mean, we think about digital transformation, three of the areas that are most front and center around customer experiences and improving journeys and building digital products. We think about data and analytics in terms of making sure the organization is becoming more data driven. We also think about how we are changing toward the future of work and improving employee experiences, and it's really in that bucket, and in the customer experience bucket and the data bucket that search plays in. So you think about, you know, how I'm gonna enable, customers to find the information that they want. And when they're having an issue or they have a question about a product and service that coming online, they're doing a search for information about the product that they just bought from you. When I think about, employees, servicing the product, creating materials around the product, marketing the product, they're also doing search. So you you look at search, it's really everywhere. But when you ask this question two or three, four years ago, it wasn't front and center. We needed to do some things before we got to how we improved search and how we were working with unstructured data. You know, you mentioned the research I did. A lot of the CIOs I spoke to working with data and analytics, they were still focusing a lot on structured data. Right? How do we bring CRM data with our ERP data together, bring analytics together, do some machine learning around what we're doing, from a profitability perspective and understanding, where where companies are making investments. And now you look at it and say, well, I have more people working remotely. I wanna connect an ecosystem going from sales to marketing to supporting a product or service. I wanna make sure that as people are working together collaboratively using digital products, they could find the information they need. And so now we're knee needing to connect information sources across product lines from sales all the way through support. Right. And I need to be able to do this across multiple businesses, for large enterprises. And that's why you're seeing this number creep up, year over year. Well, and you and I both come out of, publishing, you, with, BusinessWeek and McGraw Hill later. You know, this idea that they're still struggling with unstructured data is really a little intriguing to me, a little bewildering too. Why is it why is it such a challenge so many years later? Well, if you look at, you know, what we did in the early days of doing search, we would throw documents and folders of documents, into a search index and then or maybe we would connect our content management system to a search index. And what the search indexes did really well back then was search for keywords or search for categories that I had to do a lot of work upfront to set those categories. Or maybe I did some basic entity extraction upfront. But I was largely working with maybe a few content sources, and I was basically doing a mix of keyword and category based searching. Even though I had facets and even though I had the ability to do a lot more of it, all the more interesting things that you were doing with search ten, fifteen years ago required a development team. First to bring in all the data sources into the search index, then to deal with the taxonomies to make sure that they were accurate, then working with stakeholders. I mean, that was probably the hardest thing that we had to do back then was go back to stakeholders. And we talked about ranking results or what document fell into what category, what taxonomy. There wasn't always agreement. And so teams had to iterate multiple times to be able to get the right res you know, right documents showing up in the right results in the right order. And that took a lot of time and energy. So we were doing some of this stuff ten, fifteen years ago. Diane, do you remember I had a full development team doing customer facing search and employee facing search using one index? But it was expensive and slow to be able to do this, a bunch of years ago. So you're saying really that was commodity search too? The fact that was that commodity search then, or was it commodity search if we evolved into it, or or have we just not changed our definition of what search is because that's always a perennial problem? Diane, I think many companies fell into the well of commodity search because anything more elaborate and more customer facing and more expansive was really hard to implement. Right? I could con again, I can connect a a drive. I can connect a single data source to an index and do that pretty easily. I can do keyword searching on it pretty easily. But as you pointed out, you know, to be able to do relevancy tuning, to make this easier for, for people to find their information they were looking for, that really took a development team. So either you were doing commodity search or you were doing, more elaborate, more user friendly search, but I had to have a development team against it. The, which this this statistic sort of jumped out at me that nineteen percent no longer see the value of AI. And yet what you're describing is moving on to something that includes AI. So how do we why is this happening? Is this, because they've tried and failed, or what what's the what's the problem here? Well, again, if you're doing AI on your own, right, most people when you say I'm doing machine learning, I'm doing artificial intelligence, it means I'm getting access to cloud. I have a PhD data scientist. I'm using one of the libraries to throw a lot of data at something to come up with a model. I'm trying to see if the model is producing results. If it is, I'm gonna try to get it into production to be able to continue to manage it. We know as much as forty to sixty percent of machine learning models never make it into production, and then there's a whole cost around this to support it. So that's what when you when you see, you know, statistics around AI, a lot of it is around do it yourself. I'm building my own algorithms. What's interesting when we get into the search space is some of the algorithms that I was using fifteen years ago at BusinessWeek and then later on at a couple I mean, they've been maturing for that long period of time. The ability to do recommendations, whether you're an ecommerce or a media site, we've been doing recommendation engines with code for a long period of time. We've been doing natural language querying with code for a very long time. With code and business rules. Lots of business rules. Absolutely. Right? You go back to what I said earlier. You know, if I'm going to build a search experience in any industry, in any domain, the first thing I need to do is talk to multiple stakeholders, multiple customer personas. I'm gonna get a lot of conflicting information around this. And what does a development team do? They try to code that conflicting information into rules. And what ends up happening is a query comes in, it produces results, and you're gonna make some of those people happy around this and some people unhappy around what the results look like. There's a important context that's missing when we designed search engines with rules and with a little bit of machine learning. We were missing the context of user, who the user is, what are typical personas. Can I my search engine start picking up the user behavior and feeding that back into the algorithms so that I can search more accurately? So Isaac is going to do a search. Isaac is an expert on search. You know? You know, Isaac is a CIO. When he does a technology search, maybe that's gonna give me slightly different results when I look at a search around Kubernetes and how to scale orchestration systems and scale to the cloud, then maybe when Diane is doing that search. Really important also in employee search, you think about some of the things that go wrong around employee search without the AI behind it. Right? What department is this person? What level of experience do they have working in the company? You know, maybe the results I wanna give in an employee search for somebody who's only been at the company for a year. Maybe those results have to be a very different than an executive who's been with the company for ten to fifteen years. Maybe I also wanna analyze. You talk you talk about search analytics. They will wanna analyze what is the, what is the the the time? What is the duration or the, the the, the age of the content that is being surfaced back to my end users? Right? When I do a an employee search and I'm only getting results from content that hasn't been touched in two or three years, is that still valuable and accurate when I return it back to users? So now, you know, that intersection between being able to do AI without out without coding algorithms, being able to embed search algorithms that have been around for a long time and plugging them in directly and then being able to do the analytics to provide feedback into the results. That's what we're talking about when we talk about AI search and search analytics and saying, you know what? I don't need to be doing this by hand. I can use a a platform to do a lot of the heavy lifting for me. Well, I and the other thing that's interesting about that research that you did, back in the spring is that the executives acknowledge that digital experiences are really, important for digital transformation or digital transformation was really important in order to ensure great digital experiences, but I don't think they really got that search was the key component. That was the backbone to making sure this all happened. Yeah. I you know, if you look at what people go to IT leaders about in particular, and what they're asking for, you know, the IT leader has to translate those requirements into, you know, here are the solutions that are applicable for doing this. So when we talk about employee experience, right, we're gonna be talking about, you know, when I log in to my system working at home, do I have good enough performance for me to do my job? And when I say customer experience, it can mean a lot of different things to different people. Right? So there's, you know, you look at the first couple of years that you're investing in digital transformation, you're going to be investing in those areas. When you look at, well, how do I get to the next stage? How do I make this a differentiating experience for customers? How do I make this easy for employees? How do I reconcile with the fact that we're having higher turnover on the employee side and harder to retain existing customers and grow customers on the customer facing side? And I look at this wide body of knowledge, whether it's product information, whether it's documentation to understand how something works, whether it's a tailored answer to a particular question that I'm asking. We're all getting used to this. Right? We're getting used to being able to ask our devices questions in human tongue, ask, you know, Alexa or another virtual assistant a question and get a reputable answer, that's the table stakes that I think, CIOs are starting to see when they look two, three years down the road and what they have to build to. And that all comes down to being able to work with unstructured information, being able to take an application that does search in one domain, make that content available to multiple domains, connecting customer and in a employee experiences. That's what's driving that eighty one percent are saying search is really important in digital transformation. So you are the author of a book, Digital Digital Trail Blazers, which I read and it's really good book. Thank you. One of the things that you mentioned in there is that if you're going to start thinking strategically, you have to get your hands un messy. You've got to stop with, with doing the day to day stuff that you've always done so that you can start thinking differently and tackling the problems differently. How are you going to get rid of, ninety nine percent said they are doing manual tune. They're still manually tuning. How are they gonna get rid of some of these tasks in order to spend some time thinking of how they're going to change the situation at work? Yeah. It's a good question, and I'm gonna answer it two different levels. If you're watching this and you're a leader, right, you're you got a senior title or VP in your title, a lot of what you're doing or need to be doing in transformation work is working with customers and understanding markets, understanding the drivers of your business leaders in terms of, you know, what KPIs or OKRs are really driving their areas of business. I'm spending most of my time as an IT or a data leader really understanding how I'm gonna deliver value, back to customers and back to employees. That means that I'm relying on my technology team to do things very different than they were in the past. They need to be a lot more hands on, a lot more experimental, a lot more learning in their ability to do things faster. And, you know, if you did this ten years ago, you know, the stories I tell in that book, Diane, are about me trying to figure out how to bring software development technologies into the enterprise. And it was very hard to do back then from a mindset perspective. It was hard to do from a talent perspective. But I was always looking for ways to cut corners, to cheat, to find ways to deliver incredible value without having to get into the technical weeds to do this. And I would do this with low code platforms. I would do this with platforms that had built in machine learning built into it. If I needed to integrate lots of different data sources, I didn't wanna have an army of developers trying to connect in Salesforce, then connect in my ERP data, then connect in Google Drive and OneDrive and bring all these data sources, piece by piece by hand. I wanted to be able to say, look. Connect to these seven data sources. Bring it in. Let's look at the results. Let's do some experimenting in terms of, you know, what the experience should look like. And then let's use the feedback in terms of what the analytics are captured to tell us how do we make this better. Right? And so I don't wanna do this with code anymore. I wanna do this with low code. I wanna do this with built in AI, and I wanna be able to experiment delivering value back to customers much more faster. So one of the attention getting devices you use back then to, convince some executives was that you asked them to print out the code and which I thought was like a marvelously Machiavellian idea is to say, all right, go, go get me all your code, bring it to me. And they're like, looking at you like you were nuts. Right. But, they did. They did. They printed out not more than one ream of paper, I think you said, because you didn't wanna kill an entire forest. But, what's today's metaphor for an attention getting device to sort of, like, say, look, we gotta shake things up and do it differently. Yeah. Well, it depends on your audience. I mean, that that you know, when I think about, you know, talking to the IT group for a second, okay, I show them somebody else's code and ask them to interpret it for me and tell me what this looks like. And I I did this a couple years ago. It did make it into the book. But, you know, I had a developer, you know, doing some pretty advanced things in Barca. And looking at this code, I couldn't figure out how it was working. And I just said, look. Do you guys wanna support and enhance this? This is what our charter is, is to continually we talk about MVPs and faster market, but we need to continually improve what we're doing. And that team said, no. We don't wanna be able to go back and do this. And, you know, I go back to business groups, and, you know, it's a little bit of a different story there. You know, with business groups, I say, let's talk about who our target customers are. Let's talk about what the value proposition is, what we're trying to accomplish here, and let's get you some results as fast as we can. Let's look at a prototype. Let's look at what a search might look like across four or five data sources. Let's not get into all the nuances about how to rank, algorithms or what I wanna search on. Let's put some things up on a pallet and get people actually using this. And by making that available sooner and faster, I can use their actual interactions to tell me what to do rather than just looking at subject matter experts and saying you know, asking them that question and taking a lot of their time to do that. And I think it's terribly important in search. And the reason is because our subject matter experts tend to design search experiences based on their own expertise, not necessarily on the lack of expertise that our employees or our customers may have. And so we end up building something that's too complicated. And I did tell that story, in in Digital Trailblazer. I had to go back to our audience, our executives, and show them a completely different paradigm for searching for content, that was simpler and easier to use because our experts were driving us in down a, a path that was just gonna be too complicated for the end users and too complicated for support. So I like bring you know, if you think about the intersection, I'm gonna bring low code in. Right? So I'm not gonna have all this tech debt to manage it. I'm bringing in algorithms that are built in the box. I don't need all that expertise to figure out how to do the recommendation engines, how to do natural language processing. I'm gonna bring that together, and I'm gonna go back to the business and saying, look. I just need some basic visions, a basic understanding of what we're trying to accomplish, and then let the teams go use a platform to go iterate over this, show you results, and then you tell me how can we improve this and let users tell us how to improve this. One of the things that came out of the data is that there's still so many different of search initiatives going across the enterprise, most of which are not aligned with each other. So there's a lot of ad hoc ad hocery going on. And, you know, we we talk a lot, you and I, about developing this search center of excellence. So I'll let you read the definition or come up with the definition because there there's different ones out there, but I think this is a this is a good one. We put it up there. So if you guys are are cutting and pasting to do a business plan of some sort, you got it. Perhaps you can explain. Yeah. So remember ten, fifteen years ago, the joke around in media was we had ten, fifteen, twenty, as many as business units as we had, as many as websites that we had, we had different content management systems. And we went through the process of, you know, look, content is content management systems or management systems. How do we consolidate all this? That issue propagated into our search experiences as well. And we ended up with a couple of different flavors here. We had some that were very engineering focused. We went out and got an open source platform. We built up an index. We put it on the cloud. And then for years, we were doing that manual tuning because we didn't have the subject matter expertise partnering with our technologists to get these indices producing results that made sense. And on the flip side, you know, we had some of our business groups, they didn't wanna work with IT, so they went out and found a SaaS solution. It went and web scraped the website. It did a a reasonable commodity search, probably cheaper than what was happening with our technologists. But, again, we ended up with a basic search experience, maybe with some facets, maybe with some keyword searching. But as soon as I needed to plug in four or five, six different data sources into it and create some context, as soon as I needed to plug in analytics from our web analytics off of what was happening there, much more harder for our business groups to do. Right? So the first part of center of excellence addresses the issue of let's start really being agile and creating a multidisciplinary team. And we're not just talking about, you know, an integrator with a customer experience with a UX person. I'm talking with actual subject matter experts, people who understand the data, being part of that agile team, being able to take on a role in terms of how things are categorized, looking at the results, understanding personas, and working directly with the teams in terms of the implementation. So we're gonna get into a multidisciplinary team. Maybe I'm gonna apply that for a search for a particular department. You know, maybe this is searching, content that our marketing team is creating. I'm creating one that's around content Barca. Or maybe I'm at five or six different business lines of business that have their own search engines. Now I have multiple teams doing this, and I wanna continue to allow that to happen. Right? I want the work to happen as close to the end user as possible because that's where all that expertise lies to get the the engines and the development happening in the right starting vantage point. But we also know that we need to build up expertise around how to do things. Okay? And that's where, the second level of center of excellence really plays out. It means that we can go to a few people who are gonna be part of that program, that center of excellence. It's distributed. But when we have a question, we have we wanna create a standard. We wanna come up, create a common way of doing things. We can call into that center of excellence and say, how should we think about implementing this? What are the boundaries about how this works or where it doesn't work? Do I need more data? How should we create one integration with a particular platform that we already use instead of having our own touch points, connecting to a particular data source? So it's two levels. Level one, multidisciplinary teams, and number two, the ability to call up experts and get assistance, get training, get a best practice in terms of how to implement something. What are what are the characteristics of a person who owns this initiative internally? Is it is there a single owner? It's multidisciplinary, but there seems to be you need to have a driver. Right? And do they come out of the architecture side? Do they come out of the line of business side? How does this work? Well, you know, I'm a very strong Agilist. I believe in the notion of a product manager, who's gonna first and foremost understand the end user and the customer. Who are we building this for? Why are we building a search experience? What problems are we trying to solve for? What's the value proposition for the user, and what's the strategic intent, of investing in a platform or a capability to improve search for that group. That's what the Is that a corporate role now? I believe more companies are investing in product management disciplines. I have my own work, coming out around this. But, yes, I believe if you go back, you know, product management disciplines have been around in technology companies and SaaS companies for a very long period of time. I'm talking to health care companies, government, manufacturing that are now bringing that capability into their IT groups, sometimes directly reporting into IT, sometimes reporting into an area of business. The product manager is supposed to take all those inputs in and start setting a road map, setting a vision, providing guidelines so that every time we do a release, we could continue to improve our customer and employee experiences. It implies the search as well. Right? We might find that we need another data source plugged in to improve search. We might find out that there's some data that we wanna move out of our, out of, we wanna archive some of our data around this. We might find that there's a new persona that we wanna build an experience for or some new analytics that we wanna bring in. We're gonna get into some new areas over the next few years, AR and VR experience, metaverse type of experiences. When does that fold into what our notion of an experience looks like? So there's a lot to do here. This is gonna be something that hopefully we're not spending our time, you know, tweaking up and down relevancy ranking. We're gonna use AI search to be able to do that, but we're gonna think about what our customers and employees need next to be competitive and to bring value to them. Well, nine percent have established a search center of excellence. I was actually surprised that number was that high. And sixty seven percent are very interested in it. That, again, I was surprised. Do these numbers surprise you, or is it just me? I found it a little surprising, but, you know, here's how I contrast it. When someone says to me, Isaac, our search center of excellence, we have a team that manages our x search index, and they do all the implementations. Right? That to me is a shared service. Okay? And, you know, if you have two or three customers, a shared service might work. But for organizations that are truly trying to make search a core competency, work across multiple departments and businesses, really focus on end user experience, I don't think you can centralize that. I think we went down a centralization path because the technology was hard and the skills were hard to attain. So we brought that group together, and we sort of time sliced what product line or what department they were servicing. And so, you know, whoever screamed the loudest got the most work done in their area. I wanna democratize this. I wanna bring it out to business departments. I wanna make it easier with low code and AI enabled search. But I also know there's a level of expertise that builds up every, you know, second or third implementation you do. That's my version of a center of excellence. And I'm I'm ecstatic to see that sixty seven percent are interested in doing this because, you know, you go back to eighty one percent are saying it needs to be part of their digital transformation program. And now sixty seven percent are starting to say, you know what? We could find easier ways of doing this with platforms. Now let's cultivate a way of doing this from a process and leadership perspective so we could be successful in our implementation short and long term. What's interesting? If you take the sixty seven and the nine percent, you're really close to that eighty one. You're but I think it's gonna be if you're gonna believe in digital transformation, that search is the key pillar to that, I think you're gonna see that search center of excellence. Those two numbers should almost be parallel. Well, otherwise, you're not doing something. Right? And Something small. No. I'm slowing down. Look. I've been talking about that a lot. You know, there's, you're going into twenty twenty three. We have some trade winds that suggest we have to get more efficient. There's other trade winds that say, you know, don't slow down the gas pedal. We continually need to transform. And I look at IT leaders, digital leaders, CEOs, and say, you gotta do both. Okay. It requires some balancing. I've done a lot of writing around this around looking for digital transformation force multipliers. And what that means is I'm gonna bring people across the company who can work with a platform in our practice. It's gonna impact both customer and employee experiences and not one, both of those things. I'm gonna be looking for things that drive efficiencies. So make it easier and faster to find information, consolidate platforms, bring integrations in across multiple platforms without a lot of code, but I'm also going to drive impact in terms of growth. Right? Can I drive more customers? Can I market to them more efficiently? Can I retain employees a little bit easier because I'm building experiences that work? I think I wrote that white paper for you guys. I think you have too. Speaking of which, we have the full report available for download. So, at some in the next couple of days, I hope you to hope to invite all of you to go and download this report. And, Isaac, I wanna thank you for coming and, joining us today. Will you come back again and go through some other aspects of this? Absolutely. You know, you know, search has been in my DNA since I started working. It's always fascinated to me. I think it's gonna get very interesting over the next few years, so absolutely would love to come back, Diane. Alright. Perfect. Thank you. And, again, download our full report, to learn more about, our twenty twenty three enterprise tech report on search and AI and see if you see yourself in there. I'm Diane Burley. Thank you for joining.