Sponsored by: Glean | How Enterprises Turn Databricks Data Into Everyday Value with Glean

Summary

Glean connects governed Databricks data with enterprise context—documents, conversations, and unstructured knowledge—to make data actionable for the fifty percent of employees who never directly interact with company data systems.
Zillow's HR agent, built on Glean and Databricks, answers workforce questions in Slack using trusted company context and people data, achieving over eighty percent active user adoption.
A live demo shows Glean querying Databricks tables with full citations, illustrating how the integration enables self-service data access without requiring technical skills from end users.

Chapters

FAQs

How does Glean connect to the Databricks Data and AI platform?

Glean integrates with Databricks to bring together governed structured data with unstructured enterprise context such as documents, Slack conversations, and emails. This combination allows Glean to answer questions that require both queried data and broader organizational knowledge, as demonstrated in the live demo querying Databricks tables with full citations.

What did Zillow build with Glean and Databricks?

Zillow built an employee-facing HR agent that answers workforce questions in Slack using trusted company context and people data sourced from Databricks. The agent achieved over eighty percent active user adoption, demonstrating successful enterprise AI deployment at scale with governed data.

Why do most employees not use company data directly?

Research cited in this video shows that fifty percent of employees never touch company data, primarily because accessing it requires technical skills or going through data team bottlenecks. Glean addresses this by surfacing governed Databricks data through natural language interfaces in tools like Slack that employees already use daily.

How does an automotive manufacturer use Glean with Databricks?

A leading automotive manufacturer uses Glean AI to support engineering and product workflows with easier access to structured data and company knowledge. The use case focuses on manufacturing SKU and part-number lookup, reducing bottlenecks for technical teams that previously had to query databases directly.

Full transcript

[00:08] Let's get into why we're all here today. Let's talk a little bit about how we can turn enterprise data into everyday value with Glean. Now, my name is Jess Urick. I am a partner SE here with Glean. I work really closely with our data bricks
[00:20] partnership team as well as our customers to better understand how we can drive value and where we're seeing these partnerships really come together. So, as we get into it, let's talk a little bit about AI transformation in
[00:33] the business. Now, most businesses are getting an influx of information, an influx of AI jammed at them, and they're starting to get two different statements. either I'm going too slow,
[00:47] you're not adopting fast enough, we're not seeing enough movement, we're falling behind the competition, or we're going too fast, we're not standardizing, we're not governing, we don't actually see the transformation
[00:59] happening. We're starting to see this information inundate our leaders and they're starting to really ask this important question of how do I drive AI transformation with cont confidence? How
[01:12] do I really make an impact with my business? What are the right tools? And how do I speak to this with confidence? Now, everyone is adopting AI, right? There's AI in almost every tool we do
[01:26] today. So, what makes Glean different? Well, what we're seeing across the industry is that 88% of professionals are using AI at work. They're using different tools across the board. And
[01:39] really what they're noticing is that only 15% of them see improved performance. And of those, only 38 of them are using corporate approved tools.
[01:51] So think about that for a second. Most of our organization is either not seeing enhancements or maybe they're not even using the tools we're supposed to be using. And a study shown by Gartner says
[02:03] that by 2027 90% of businesses are going to have AI deployed broadly. But of those 90% only 12% are actually AI ready with data
[02:15] quality and the architecture to support that. Now this is really where Glean stands out. What we see with Glean and what we'll highlight here today is that with Glean, our users are 50% more likely to
[02:28] say they see improved outcomes with AI. Their performance improves. They're able to move faster in their jobs. And 17% of our users are more confident in the AI
[02:40] strategy that their business holds. And a lot of that has to do with that 50% of users seeing that improvement on outcomes. So, how does this all come together? What makes Glean different?
[02:52] It's the context. Context runs deep. And in order to really understand and make an impact in your business, we have to understand what types of data your organization uses, the tools and
[03:05] processes that are most important to your business. What about the projects and teams that you have, the different departments? We need to be able to put this all together in a way that makes sense and drives the business forward.
[03:19] Now we do this in two ways. The first is that deep understanding of the business and the second is a deep understanding of your users. So that this understanding of the enterprise is used to power everything that glean does.
[03:34] So let's take a look at what Glean actually is and how it comes together before we get into how we come together with data bricks and our customers that we see using it. So Glean is a horizontal enterprise
[03:47] context layer. We can work throughout every department in your business. And we do this by connecting into the key enterprise tools that your organization is using today. Think of all of the different SAS products that you're
[04:01] probably logging into and accessing on a daily basis. Your email, your calendar, your messaging tools. We have ticketing and wiki pages. All of these different knowledge sources in an organization.
[04:13] Glean connects into all of those natively. And what's great is we're governed by the native sources. So you're we're inheriting those permissions and using it to build out the glean enterprise context. Now the
[04:26] glean enterprise context as I mentioned is two parts. We have on the right hand side over here the enterprise graph. Now the enterprise graph is that understanding of the business. So what
[04:38] types of tools are your teams using? What are the most important to your business? What types of data? What customers do you have? what metrics are we tracking and what drives your business forward.
[04:50] But we also have a personal graph. The personal graph gives every single user an individualized experience. So now we're taking all of this enterprise knowledge and we're translating it into
[05:02] your specific work habits. what types of data you interact with, the teammates you have, and what metrics are important to you, the way you like to work, and your writing style.
[05:15] And what we do from there is use that to help prioritize and provide context to the Glean search. So now you're going to find those documents you're using on a regular basis much faster. But even more
[05:27] than that, we're providing personalized answers and responses for every single agent and assistant query a user runs. So that means every time I ask a specific question, knowing I'm a partner
[05:40] I see, it's going to use my enterprise graph context and the personal knowledge about how this question relates to my role and what's important to me and it's going to translate that into an answer.
[05:53] Now, the reason this is so important is because now I'm getting not only an understanding of the question I asked, but it's also understanding why it's important to me. Why might I have asked
[06:05] that question and what data should it be providing me back with to give me that full and complete response? And from there, we have our Glean agents and assistant. You can use those right in our Glean UI. So, everyone in your
[06:19] organization can have their own personal work assistant. that they can ask all those questions to all the time and help them build out their work content or we can go out and build agents. I can build my own agents. My company can help
[06:32] support and build agents. All of you as users can go out and build these agents. And now we're seeing work progress faster because these agents are those multi-step workflows that take those tasks that we're doing day in and day
[06:44] out and make them a little bit easier. So instead of me having to go into 16 different documents and 12 different apps to reply to an RFP, I can actually do it right through an agent.
[06:56] And the great thing about this is from there we can take this and move it to anywhere you work. So whether you're working in our Glean UI or in data bricks genie for example or Genie code we can plug right into those with our
[07:09] MCP server or even better we can embed our Glean agents right into the UIs of where your team works. Think call center UIs and other places where now you don't even have to leave where you work to get
[07:23] the enterprise context and the deep understanding from Glean. So now we know how glean brings the enterprise context. So how do we work with data bricks to bring the analytics
[07:36] and the enterprise context together? Glean works together with data bricks in a few different ways to ensure that we're getting the best responses. So Glean is going to do all the enterprise
[07:49] data. We're going to get that deep understanding of your business, the emails that you sent, what information you're working on, your day-to-day, and we're going to use that as our system of
[08:01] context. And then we're also going to use data bricks. Now, data bricks we're going to plug into data bricks genie for example, and we're going to use that and bring in the structured data. Now, what we're not
[08:14] doing is we're not moving your data bricks data into glean. We are using the analysis done by data bicks and taking that response and adding in our enterprise context. So if you're doing
[08:27] that inside of data bicks that's great. We'll layer in the context on top. If you're doing that inside of glean which is what I'm going to highlight here today we are going to make that call out to data bicks show you where all that information came from and provide you a
[08:41] response and layer it into that enterprise context as well. And the great thing is is there's a few ways we can do this. We have four different integrations, all of which are meant to meet all of our users where
[08:54] they're at. Whether that be using the Glean agents, calling to datab bricks agents or using SQL. It can be our assistant just calling out
[09:06] to SQL and asking data bricks questions which I'll be highlighting here today or for all of my primary data bricks users it can be done right inside of data bricks genie using our MCP. So there's a
[09:18] few different places we can leverage that as well. So the idea here is we're bringing these two systems of context together to democratize the insights, bringing faster data analytics to users
[09:30] who aren't our data analytics teams. We're unifying that context. So now we're not only understanding the data, but why that data specifically matters to my enterprise and why it matters to
[09:43] me as a user and why I'm asking that question. And then the consistent governance that we apply. Remember we are inheriting the native sources. So if I don't have access to it, I can't see it. Same goes
[09:56] for data bricks. If I don't have access to a table in data bricks or let's say I don't even have access to data bricks at all, I'm not going to be able to call that data bricks information into glean. So that consistent governance allows for
[10:09] our customers to feel more secure in how they're using these AI products today. So before I get into the customer story, I want to highlight the specific use case that we're really seeing our
[10:22] customers drive and the two stories I'm going to highlight here today. In this particular case, we're talking about SQL in Glean Assistant and Glean agents. Now, this is going to be great for our technical users or business users or
[10:35] anyone with some expertise in uh inside of Glean. And there's two ways we can do this. Our newest beta is right inside of our Glean assistant. What this means is you can type your SQLs or natural
[10:47] language queries right into our Glean assistant and it's going to go ahead and call out to your databicks genie. Now, this is really great because think of all your business users who have questions about data that's stored in a
[11:00] data warehouse that they might not regularly go and access or even know where that data lives. Well, now they can access this right through the assistant. They could be experts in SQL and write it out themselves. They could
[11:12] take legacy queries that they know work and paste them into Glean or they can ask a natural language question. And then of course we have our Glean agents. Now the Glean agents are these
[11:24] multi-step workflows. They allow your organization to create stepbystep workflows that move teams faster. or think about your monthly reportings or regular workflows that go and look at
[11:36] data and data bricks but also need to understand why that's relevant to what you're doing today. We can build out these agents that walk through these workflows step by step. Whether it's using a direct SQL query,
[11:50] right, using those pieces that are still important and our business is relying on today or moving into a more advanced function with a datab bricks genie action. We're able to interact and use
[12:02] these for the best repeatable workflows. So, we've now seen how Glean and Data Bricks come together, bringing the enterprise context and our structured data into one unified response. Let's
[12:16] take a look at a couple customer use cases and where they're really seeing value today. So, here we can see we have a couple different use cases. Now, I'm going to go ahead and start with the Zillow use case. And as I talk through it, I'm also
[12:29] going to make sure my demo is ready to go here. So the Zillow use case is for their HR team. Now they are one of our great customers. They they allow us to share this story here today. But what
[12:41] they went ahead and did is they built a glean agent. So that second use case I shared that helps with some of their HR use cases. What they're doing is they went ahead and built this agent that takes some of this context from the
[12:54] daily questions that their users are asking, but also their legacy HR data. So people data, demographic data, things that they need to be able to tell over time, but is not something they're probably storing in a quickly accessible
[13:08] Excel doc for example, and they went ahead and they built out a agent. Now this agent is actually not plugged into just Glean. It's actually embedded into their Slack. So as people are going into Slack and asking these specific
[13:20] questions, if they have access to the data at the native source, of course, we're Glean goes ahead and provides them with this response. Now what they're seeing is they're seeing a better response in a the answers that they're
[13:33] getting because they're getting them faster. Teams are moving faster because of this and they're able to scale their support with their HR much faster. Now what's great about this is they've
[13:45] even shared that over 80% of their users are natively and actively using Glean every single day. So think about that for one second. over 80% of their company is already using Glee natively
[13:57] and I think that quote was a little bit ago so I would be surprised if it's not even higher now and they have over 3,000 agents built all of which are driving their business impact forward. Now in this particular
[14:11] case, this uses that legacy data and data bricks. The the the Slack agent embeds them together with the workflow created by Glean and they're seeing scale.
[14:24] Now on the other side here, we have a manufacturer, right? An automotive manufacturer. They use us in their engineering and productivity stages with the trusted data. So they have all this
[14:36] legacy data in their data bricks tables of part numbers, manufacturing SKUs, how long it takes something to manufacture, the product defects for that and they want to better understand the impact on
[14:48] their engineering. So what they're doing is they're creating a workflow that allows them to reach out to the data bricks data and better understand when they're going to build a new part on a
[15:00] new car or add an existing part to a new car, what that impact would be. Is it actually going to be more costefficient? Is it going to produce more product defects? and they're using that to build
[15:13] and manufacture more efficiently. Instead of having to go and reach out to their data analytics teams with these lists of parts, all their questions, they're getting these responses significantly faster because they're
[15:27] able to track across all the systems that they're using and leverage the enterprise context. So with that, I'm going to go ahead and quickly switch into a Glean demo here,
[15:39] and we will see Glean in action. So let's see. There we go. So you can see I already have it prompted up here. So from data bicks, find the
[15:52] distribution of the industry sector among the businesses in a B2B addressable market. This is a very specific question, but what I really wanted to do is go to datab bricks and find the industry and look at the different markets and
[16:05] provide me a response. Now, what you're going to see is Glean is going to go ahead and start working on my response. And as it goes through this response, step by step, it's going to show me what it's doing, where it's calling from, and
[16:19] the information I need to be aware of. So here we can see it's starting to use the data data analytics from data bricks skill. We went ahead and added a skill. This is an additional piece that you can add to make sure your performance is
[16:32] consistent or you're getting a consistent output. And we can see it start to run. Now I'm going to go ahead and I'm going to jump over to my run of this from just a few minutes ago. I went ahead and pre-ran it
[16:44] for us. And I'm specifically gonna pop into the step here using search with data bicks genie. Here we can see the SQL it generated directly from my question where it's going and what it's
[16:57] going to go look at. There is markers along the way to show you where it's pulling the data from why that data is important and every single step along the way. So there's no questions of where did this come from or why did it
[17:10] come from this specific table. You can see it every single step of the way. And here we can see it produced the results. And we can see it shows me
[17:23] exactly where it came from. It gives me a list of the businesses, the shares and the total businesses returned. The one thing we want to be aware of and the question I really asked is at the
[17:36] bottom here. And we want to be aware of that our software development industry is nearly half of the businesses returned. So it went out, it searched through it, created my SQL and returned a result. So here we can see how data
[17:51] bricks and glean comes together. Glean understood why I wanted to ask the question, the context behind it, and datab provided me with the tabular response, and the data to back it up as
[18:04] to why and what is going on, allowing me to move faster in my organization, allowing my teams to get better access to data, and just like we're seeing with those two examples I share, higher adoption on our AI investments.
[18:20] So, thank you all so much for joining our session here today. I'm happy to answer any questions. Feel free to stop by our booth as well. We are here all day. So, thank you all.

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.