Your AI Strategy is Only as Good as Your People Strategy

Summary

This video explores why 60% of organizations still fail to generate value from AI investments while market leaders—those who have upskilled more than 50% of their workforce—are already redesigning workflows around AI.
Robin Sutara and Melanie Botha introduce a practical roadmap using the Databricks data and AI maturity framework to assess organizational readiness, close skills gaps at scale through structured learning programs, and integrate AI agents in ways that elevate human talent.
The enablement model follows two tracks—Foundations and Fluency—and emphasizes building cross-functional AI communities, running hackathons, and using Unity Catalog to make agents discoverable across the organization.

Your AI Strategy is Only as Good as Your People Strategy

As AI moves from experimentation to enterprise impact, success depends on workforce readiness, not just technology. Market leaders have upskilled over 50% of their employees and are redesigning workflows around AI, while 60% of organizations still struggle to generate value from their investments. In this video, Robin Sutara and Melanie Botha explore how to build an AI-enabled workforce by aligning culture, people, process and enablement with your data strategy. You'll learn how to assess your data and AI maturity, close critical skills gaps at scale through structured learning programs, and integrate AI agents into teams in ways that elevate human talent. Drawing on customer proof points and the Databricks data and AI maturity framework, this video provides a practical roadmap for treating upskilling as a core investment—so every project becomes a data and AI project and every employee contributes to AI-driven outcomes.
Talk By: Melanie Botha, Director, Learning & Enablement APJ, Databricks ; Robin Sutara, Americas Data + AI Enablement Leader, Databricks ;

Chapters

FAQs

Why do 60% of organizations fail to generate business value from AI investments?

According to this video, the gap between AI experimentation and enterprise impact is primarily a people and process problem rather than a technology one. Organizations that struggle have not aligned culture, workforce skills, and operational workflows with their data and AI strategy.

What does the Databricks data and AI maturity framework help organizations assess?

The framework provides a structured way to evaluate where an organization stands on its data and AI journey, covering dimensions such as data governance, skills readiness, workflow integration, and use-case prioritization. It serves as the diagnostic foundation before designing an upskilling or transformation program.

How should organizations structure AI upskilling programs at scale?

The speakers recommend a two-track approach covering Foundations for broad organizational fluency and Fluency for deeper practitioner skills. Structured learning programs, hackathons, and cross-functional AI communities are cited as practical mechanisms for closing critical skills gaps systematically.

What role does Unity Catalog play in AI workforce enablement?

Unity Catalog is used to make AI agents discoverable across the organization, allowing different teams to find, reuse, and govern agents without duplicating effort. This discoverability layer is presented as an enabler for scaling AI adoption beyond individual teams.

Full transcript

[00:08] Hi everyone. Thanks so much for joining us this afternoon for our next session as part of the data strategy track. So my name is Bria Clevens. I'm an account executive here at data bricks based in Australia if you haven't already picked up on the accent. So for the session
[00:20] this afternoon um we've got Robin Sutara and Milani Botha who are going to be talking about how your AI strategy is only as good as your people strategy. So a little bit about our speakers for today. So Robin Sutara is the America's
[00:35] AI enablement leader at data bricks where she helps customers partners and field teams turn AI strategy into measur measurable business outcomes. With more than 25 years in data and AI, Robin brings a rare end-to-end perspective
[00:48] shaped by roles as field chief data officer at datab bricks, chief data officer for Microsoft UK and COO for Azure data engineering. Her path from the US Army to law school to the seauite
[01:01] gives her a practical no jargon style that resonates equally in boardrooms and with practitioners. And then a little bit about Milani both who's also speaking today. Milani leads capability engineering at data bicks drawing on 25
[01:15] years plus in the tech industry across roles including CEO, CMO and COO spanning Europe, Middle East, Africa and APJ at AWS in A&Z. She designed industry
[01:27] first programs that skilled hundreds of thousands of people across the organization. So for the session today, uh a little bit about what you're going to be hearing about. So as a as AI moves from experimentation to real impact, success
[01:41] depends on workforce readiness as much as technology, Robin and Melanie will explore how to build an AI enabled workforce by aligning culture, people, process, and enablement with your data strategy. Drawing on the data bricks
[01:54] data and AI maturity framework, they'll cover how to assess maturity, close skills gaps at scale, and bring AI agents into teams in ways that elevate the people who are already there. So, please um give a warm welcome to Robin
[02:07] and Milani as they um join us on stage.
[02:19] So, I guess before we get started, I have to do the little disclaimers. I have no idea what this slide says because they threw it in when we weren't looking, but essentially I think it does say that anything that we cover here is covered uh or that we won't give inside information, etc., etc. We are not doing
[02:32] any technical product talks. So I seriously doubt that anything you learn here about your people or process is going to impact in this capacity. Um and I do want to make a call out as we finish the session if you could please
[02:44] make sure to fill out your survey. Uh the teams do use it as they think about what tracks to include next year on whether it was helpful for you as you think about your data and AI strategy. So I think we were just going to take a little bit of time to set some context
[02:56] for you. I think we have been talking about AI for years at this point. I think many organizations continue to feel like they're either really far behind uh the rest of the world or that they haven't quite figured out how to
[03:08] get over the the um hoop of being able to do it at scale in production. And so what Mel and I are going to share with you today are some real world of examples of customers and organizations
[03:20] that we have worked with globally and how they are actually overcoming from getting from just testing and and iterating on those early lowhanging fruit use cases and how they actually start to think about it using at scale.
[03:33] And because we both have the benefit of doing the field teams inside of data bicks, we're also going to give you a little sneak peek behind the scenes on some of the things that have worked and not worked as data bicks has tried to uh create an AI enabled workforce inside
[03:46] the company. >> And I think Robin uh what's hello is so interesting is someone from a military background and then an African you're gonna get the real deal.
[04:12] is for most organizations you they are either early still in the process where they're just starting to see productivity gains. So about 35% of organizations are still just in that early phase of getting low uh
[04:25] productivity gains out of their workforce. And about 60% of organizations started on that journey and now they're just sort of stuck trying to figure out how are we going to get outside of just gaining those incremental gains and how do we really
[04:37] think about transforming our organization. The problem is as we see more and more agentic systems enter into the ecosystem that that gap is not closing. It's actually getting wider because we're not thinking about the people and the evolution that we have to
[04:50] take the workforce on to be able to leverage those agentic capabilities in the right way. So I I know several of you joined me last year when I sort of talked about data strategy. I continue to have the belief that only 10% of this problem is
[05:03] really about technology. The other 90% of your AI strategy success lies with your people. It lies with your processes. It lies with your culture, your change management. It's all of the other pieces that um we we don't talk
[05:16] about quite as much at data bricks, but we are absolutely here to support you on that. And so we will talk a lot about how do we help you and partner with you as we think about helping you cover that 90%. Because the reality is as much as
[05:28] everybody wants to talk about models, as much as we read about models and the new model uh that that's coming out this week or next week, everybody has access to that same model, right? That is going to become a commoditization. Everybody
[05:41] has access to the data bricks platform. The thing that makes your organization unique, your true competitive advantage is your workforce. It is your people and is the knowledge that they have in that domain expertise to actually do
[05:53] transformation >> 100%. And if I look back at last year, Robin, when we also spoke about data and our culture, how the message we had then what we had
[06:07] today versus oh even even two weeks ago was different and how you need to think about skills and how you brought that up, talent. So, everyone's struggling with it or or coming to terms with what
[06:19] this now looks like. And it almost feels like we're repeating the world when we were thinking about how you think about cloud. It's now about how you really make sense of of your workforce in this
[06:31] new new new environment. >> I know uh I mean Mel, you do a lot of work with our Asia-Pacific customers and I know you've done some work with NAB who's actually figured out like how do you approach the workforce? Could you maybe share an example?
[06:44] >> Right. So they've really an incredible customer of ours and I've enabled >> Oh, thank you for being very transparent on
[06:57] that one. So this do we do we need help? Is that better?
[07:12] >> Is that better? Now, as I said, I'm African. I can shout. So, that's good. Okay. So, NAP has um trained over 500 people um in through an academy and always focused with all with
[07:25] everything they do really on uh fluency within the organization. But through that work that they've done, they've seen a decrease of 50% on uh ingestion
[07:38] costs >> and a 30% increase in ML model performance. So um where it really becomes interesting is when you really see the results behind uh what you're
[07:51] doing and how you're enabling. >> Absolutely. So prior to coming back into the Americas, I was in the Europe region and we've done a lot of work with Shell. Very similar. How do they set up centers of excellence? How do they really start to think about enabling their workforce
[08:03] so they can get things like code now from two hours down to minutes? And so really focusing on the people and taking them through that enablement capability and building their capabilities has really made a difference for multiple of our customers. So what we wanted to
[08:16] share with you was maybe a little bit of where do we see organizations when we think about their AI in uh capability building and sort of the journey that they go that maturity journey that they go through. The majority of
[08:28] organizations are in this first pillar, sort of the tactical approach, what we call the scalers. This essentially means that you have done no organizational changes. For the most part, your teams and structures have stayed in place.
[08:41] Your workflows have stayed relatively consistent. This is really focusing on some of those quick productivity wins, right? This is where you're thinking about how do we automate some of these repetitive processes? How do we focus on
[08:54] the tactical things that our workforce is doing today? And how do we leverage AI to take some of that tactical work off of the workforce and leverage AI and agents to be able to do that capability? We would probably say about 60% of
[09:07] organizations do fall within that tactical. They are still working through some of these uh initial use cases to be able to gain those productivity gains. The interesting part is when we start to see organizations take the next step on
[09:19] that maturity journey to what we call reinventors. This is where you're now actually starting to think about the workforce in a new way to say, do we have the right rules? Should we rethink how people are doing their jobs? Should
[09:31] we rethink processes? Um, Mel and I have done a lot of work and we'll some give some great customer examples, but even internally, we've thought about how do we make sure our teams are rethinking existing legacy processes and workflows
[09:44] and if we could recreate them with the technology that we have today, what would we do differently? So I would say even data breaks probably falls in this middle in that uh right we are definitely thinking reinvention. How do
[09:56] we make sure that we're creating net new capabilities of the workforce and shifting workflows and making sure that we're aligning the organization. And then we have the relative minority which is the native operators. These are the
[10:09] ones that are completely rethinking their organizational structure. Many of them have uh agentic employees that are actually essentially on the payroll being performance managed just like everybody else. Uh but they are really
[10:22] thinking about net new revenue streams, new uh capabilities, new roles that they will create as a result of being able to do that type of work. But I think Mel the biggest thing would be I think if we could share some real world examples. I
[10:35] think so. >> And maybe just to add to that as well find that um many organizations u might also have different of these phases in the same organization. So I find that
[10:48] you can have >> absolutely >> one area and you know a scaler in another part of it. So it is very interesting to see uh where everyone's at uh when we speak with them. So that
[11:02] that's quite interesting. But uh in uh we we've had a government uh agency and this is more on the scalar side. So on the productivity piece where um they
[11:15] enabled their teams on um their data analysts on just um Genie and and the natural language pieces and uh they their dashboard creation has reduced
[11:28] from 90 days to 30 days and they've saved 8,000 hours in uh productivity uh in perom with that. So that's very
[11:40] interesting example just on the basic you know um another we just stick to government New Zealand government agency they had these three pipelines that were
[11:53] typically traditionally um created uh manually uh just by the subject matter experts and the architects started to look at the capability the uh
[12:07] development capability of Genie code and got the raw data, built these pipelines and they saw that they had a fully end to-end pipeline development in uh two weeks
[12:21] where they usually have it a quarter. Yeah. But also what was really interesting is they've managed to um consolidate the SQL server processes,
[12:33] the uh spreadsheets uh spreadsheet visualization as well as the SAS orchestration. So that was really interesting that where they just started off on the money, they
[12:46] started to see more and more benefit as they go down that route. So you can get deep scaling uh focus there. I find it's even super fascinating that lots of organizations even as you think about scalar are leveraging it for traditional
[12:58] AI right so I have a lot of manufacturing customers for example in the Americas that are really saying are this process takes 75 steps how do I use traditional AI or ML to be able to optimize and increase productivity or
[13:11] gain I think a lot of times we talk about workforce and we talk about we hear about workforce being displaced by AI but I think for most organizations that are in this early sort of journey they're not displacing what they're
[13:23] trying to do is get more out of the existing workforce. Um right because you can't hire enough people to get the work done. And so how are you really thinking about those productivity gains so that you can get um right increase in
[13:36] revenue, more productivity, more outcomes with the same number of people that you have on your existing team. So I think it's it's really interesting that it's not just AI, but it's revisiting traditional ML for many
[13:48] organizations and they're using the AI storyline with their board or their CEOs to gain the support to be able to drive some of these tactical scalar type of activities. >> And Robin, do you have an example of a reinventtor
[14:01] >> uh outside of um data bricks since I just shared that we were probably reinventors? Think uh one of the big ones that I've seen in in Europe and Middle East uh Europe was really uh Adidas uh right so they are a big data
[14:13] bricks customer they were really thinking about how to uh innovate and net new create pro products and workflows and rethink so they essentially have serviced up early on agent bricks they were one of our early
[14:26] organizations that were leveraging that and they thought about it not just from the perspective of their technical team they actually gave agent bricks up into their executive team so that they could start to make CFO, CMO, COO type of
[14:39] decisions um so that they were actually then able to revisit the existing work force uh workflows. So for example, the CEO would send a question typically would have taken an analyst X number of
[14:51] hours to come up with that answer or that response. Now the CEO literally had it at their fingertips and they were able to recreate that process so that via Genie agent bricks at the time that CEO could self-service that data and
[15:04] that information directly and so they were able to u recreate the entire workflow or process to get those types of answers. Um I know you have an example of where it didn't quite uh work as well. Actually I have an example of
[15:18] between the two stages of reinventtor and um >> ah okay >> the native so uh that one is with 7-Eleven most of you know that 85,000
[15:31] stores and uh 19 countries so the complexity is very real and what I love about that is in the CMO world and coming from TV and radio before and um
[15:43] marketing uh let some of businesses. The interesting part for me there is more and more and I'm so excited to see the announcement of customer lake. >> Um I think what will be really
[15:56] interesting is to see how the uh these business teams and the tech teams have to collaborate more and more together and I still find that a lot of customers
[16:08] are asking us please just don't let the zombies out. um don't give too much enablement to the business users. Others say let's just train all the business users. But what what happened with
[16:21] 7-Eleven was um they had um they had to reimagine their whole marketing process. They were in particularly wanting to have creative concepts, fresh creative concepts, uh
[16:35] copywriting, um uh have have more multi- channelannel copywriting and a proper brand alignment. And they found with the current chatbots that they had that that
[16:47] brand voice just didn't come through. So um they started off first as a basic um process where they just had basic rag and they had um a document based wall
[17:01] that helped for proper queries but it didn't really help in uh the heavy image content uh carrying that load. So um they set out they they brought the two
[17:14] teams together and they totally reinvented um looking at creating a few agents and how they will then orchestrate those agents which meant one
[17:26] was a creative uh um content uh designer uh one was a copywriter and for multi- channels is really
[17:38] challenging for those of you who come from that space. you have to have so many different operations of of your copyrightiting and then one was a um a supervisor and one just a general
[17:51] agent. So they really thought through very carefully what the type of agents are that they will have um in this reimagination of the role. So that that one I I really um am excited about what
[18:04] they achieved that where they just started off with doing just a scalar activity ended up in being a fullon >> um reimagination. >> I love that because it is iterative. I
[18:17] think everybody sort of looks at this and says how do we solve for it end to end. I was at a CMO event here in San Francisco probably four weeks ago and it really was how do we think about these roles and functions where we can think about identifying some components of it.
[18:30] Do they understand the market the way I would ask the questions? Do they understand the way I would create the content, the personas that we're trying to target? And so I think a lot of times when organizations look at these steps, they sort of think it's an all or nothing, but for many organizations it
[18:43] really is attack um attacking it. That sounds really negative. It's really focusing on those specifics uh and how to sort of build it out. Maybe as a an example as well, I think workday is another one of the data bricks customers
[18:56] I think that has really done this agents as a workforce. I mean they have completely re um relooked at their rev ops. They have completely looked at entire end toend processes and then really thought about uh how they were
[19:09] going to think about the workforce to be able to deliver it. In fact, today I was meeting with an organization who was saying, "We have an aging workforce who probably only has another three to five years. How do we think about identifying their current capabilities because I
[19:21] won't be able to hire as fast as I will lose my workforce over time?" And I think two years ago that was not a reality. Organizations could not think that way. It was very much how do we get as many interns as possible so that we can uplevel them before we lose the
[19:33] workforce that has all the tribal knowledge. Now it's a reality to really think about how do you completely AI natively rethink the workflow flow from left to right and where you can actually build out now the technology to capture
[19:46] that before your workforce leaves. Um, and I think for many it's not again it's not displacing uh workforce. As much as we uh love to read the headlines and we've all seen them, I think most of us would see it's really a reduction in
[19:59] force to cover the costs of AI. It is not AI displacing the workforce. There is more need for more capable people within your workforce um to be able to deliver against the innovation and the transformations that you want to do in
[20:11] your work. >> Um Robin, we um I love learning from our customers. So, I'd love to see a raise of hands on who um already have agents
[20:24] in place like what what shall we say? One or two agents. >> Yeah. How many have one or two agents in production in your environment today? >> Quite a lot. Uh more than five.
[20:37] More than 10. >> Oh, >> okay. Uh more than what? 20 >> 20 >> 30 Okay, interesting. So, we definitely uh
[20:50] love to learn uh from you all as well as much as we are are working through this and um we'll we'll also talk obviously a little bit more about some of our lessons learned. Yeah. As we go through
[21:02] it ourselves inside data bricks, but very keen to hear more from you afterwards as well. Yeah. on on that. >> As you can see, there are many that are doing some early uh use cases with agents. Without a doubt, what we know is
[21:15] that a majority of organizations actually expect to have um agents deployed in production by the end of year 2026, which is six six months away. So, um the problem is for most
[21:29] organizations, they say the limitation the limiting factor is not the technology, it's not the models, it's not the platform. It still continues to be their workforce. Right? This isn't something that you're going to be able to throw money at. We've been throwing money at the technology. So, how do we
[21:42] really think about that transformation of the workforce? So, what Mel and I thought we would do is share a little bit about how we've done it inside of data bricks. What we have been talking about for the last year to 18 months is really a builder mindset. How do we
[21:55] ensure that every person in the organization regardless of their role, regardless of their background, is a builder? Now the problem is we said that 18 months ago and then we didn't define what is a builder right and so we've
[22:07] definitely done a lot of work over the last year um to working with our our founders and our executives to really help the organization understand what do we think that that means and I think
[22:19] there's a few fundamental components to it um the first one is not reinventing what already exists it's really thinking about like we talked about that end toend workflow the end toend process uh
[22:31] exactly Can you take ownership from identifying a problem and then actually build something that's going to solve for that that problem on the outcome. Um the other thing is we found we have a
[22:43] super technical field that we sort of service inside of data bricks as you can imagine just listen to all of the amazing announcements super technical workforce. Um it was a lot of how do we help them show not tell this is not
[22:55] slideware. It is an AI strategy should not just be something that lives in a document or in a deck. How do we empower the workforce to actually feel empowered to be able to to build it from end to end so that they could actually
[23:07] demonstrate the value and not just articulate what they thought the value would be as a as a result what we see there I mean we've seen it yourself you now find that well what we
[23:20] found because both of us part of our team manage the internal uh field enablement is you have to relook that skill days of >> um yes they're a builder or from the
[23:33] very technical teams where they can demo but now bringing back to the business realization and value and how they become more industry aware um is
[23:45] something that you now start to look at those other verical skills >> that are becoming more and more critical. So balancing out the skills I found in the areas where I'm at uh was
[23:57] quite an interesting one. And I think it's not just balancing the skills but also building the skills. Right? So I mean initially when we started it was very much can we use agents to be able to help us with a specific skill or a specific task. Now we're really saying
[24:10] like look it's a it's a teammate. It is part of your organization. How do you use it end to end? How do we really think about some of these compound systems where it's not just doing one part of the task but you're thinking about leveraging it as you would uh
[24:24] somebody else on the team to be able to help you solve it from end to end. And then the last part was really stop asking for permission. Like just go do something. Go build something. Take the ownership. Find the problem. Do the
[24:36] solutioning and build it out. Uh and and so then how do we sort of do it? All of that sounds really positive, but I think we would be remiss if Mel and I didn't share with you sort of the good, the bad, the ugly of going through the journey. So of course we've got to start
[24:48] with the good because you have to start with positive, do you not? And being with Mel, you always have to start with the positive. She one of the most positive people I know. So Mel, I know you have done a lot around in APJ around builder days. Could you maybe share what
[25:00] what that is? >> Um, as you can as you can imagine, a engineer in datab bricks um would profess that they're quite deeply technical already. So you got to
[25:12] consistently update them um have them totally um enabled. And so now um and obviously across so many countries >> uh and languages. So uh this year for
[25:26] our tech technical summit where we kick everyone off, we did a two-day virtual enablement which is not ideal never but then we brought everyone together in um
[25:38] areas and they just built for the day. I was amazed at some of the uh apps that were created and the rule was it had to be free use case for the customer. We
[25:50] then took those use cases to some of the customers who have adopted that. But that was one day sitting together and building together. What we find more and more it's about communities. So our subject matter
[26:04] expert communities are critical in our business. So the face of enablement is changing rapidly. Yeah. And we have industry communities and subject matter expert communities that we have in our
[26:16] teams to really learn from each other and have some structured mechanisms um to to to submit submit requests that
[26:29] they are struggling with with customers and then they quickly go and build together. So that's we're finding is working well for us. >> And I think with with any personas we're trying to create that demonstration like having each of personas or champions
[26:43] within those communities share. So how we're doing it for example for our account executives which are our sellers is how I AI right so the teams are definitely doing very small groups focusing with them they're letting them share best practices but they're also
[26:56] sort of leading them to say if you don't know how to get started pretend I I let you hire a college intern. So it's intern as a service. Your intern is actually Genie, right? And so how do you actually what would you tell an intern
[27:09] to do? What would you ask them to do as they were learning the business as they were trying to get the context around what we do at the company? Same as you would do for an intern, you're also going to double check their work and how do you give them feedback on how that is working. So for a non-technical audience
[27:23] like our account executives, that just seemed to totally open up the door and opportunity for them to realize they were also builders. this was not just the technical teams that were going to build for them, but they could figure out how to build their own solutions to
[27:35] be able to solve for that. >> So, um, then there's a a brainchild from our VP and the the business that came to life literally we were working backwards
[27:47] at as a leadership team um as an enablement leadership team November and said we have to focus
[28:00] And uh that's where we want to go. So changing face completely. We um the team built this this app and because you can imagine skills data is everywhere as
[28:13] well. So there is it's scattered spreadsheets assessments if you think you're doing an assessment of your team uh in a week's time that will be outdated. So no one really knows where
[28:25] their teams are at and where they should be going. So for our field engineering team, we did that and we're using AI to infer their skills and um also all their
[28:38] proficiencies. So that means that tool is so powerful you now um you now ingest uh enablement based on where they are at because of what they've been doing in
[28:51] their work today. What I find fascinating is, you know, someone that have already done a lot of work on um uh demos and they've been with a customer so that that will already show up that
[29:04] they are quite proficient. So you don't have to go and brush stroke and have everyone um shove things down their throat on everyone have to do this. >> Yeah. >> Um there's a there's a time and a place
[29:17] for that though. But uh we we're using AI now to drive hyperpersonalized enablement for our own teams in data bricks. >> And I love the fact that it's not just
[29:29] what they do inside of data bricks right because we're also now thinking about how do we take into account external thought leadership blogs that they publish etc. Can we use that to help assess where their current capabilities or competencies are and what their
[29:41] maturity is so that again we're not rolling out the same enablement or capabilities to everybody across the workforce. How do we make sure it's really helping them grow to the next level of capability? And so I think super >> So I spoke with uh someone from a telco
[29:55] as we walked here this morning and we spoke about this topic because they're working on some pieces there as well. And he said the only thing that he would be cautious about is that managers would
[30:08] want to use that as a performance management tool. So you got to have the you know cross the fine line of um ensuring that it's really a skills focus um when you think of how you reinvent
[30:21] your enablement internally. >> Absolutely. >> The other thing we wanted to make sure that we covered was a little bit on where we didn't do maybe quite as well um uh and one of my team members Vun is in here who helped drive us across the
[30:33] organization. He did a fabulous job. But essentially we had um an agentic system that we essentially called scaled prospecting. If you think about your early business development uh resources and sort of how do we leverage them more
[30:46] effectively. A lot of that we wanted the actual BDRs to focus a lot more on relationships with actual customers as opposed to doing things that could be automated. Outgoing outbound emails uh
[30:58] responding to incoming emails that were relatively straightforward answers based on public websites etc. Uh and so the team did a great job of testing uh essentially with one of our digital natives teams uh and sort of how we
[31:11] thought this uh agentic system that would cover some of the scaled prospecting role or work would would sort of handle. Um the the problem is we didn't do a great job on thinking through what would be the actual change
[31:23] management to roll it out across the entire organization. How would we establish the operating model? what roles and responsibilities did each group have as we rolled this change out across the company? How did we get how could we get executive buyin uh sort of
[31:37] for each of the different industries that we were going to roll it out against? Uh and so I think we learned a lot of lessons on how do we get more effective about that? How do we think about net new solutions that we come up with that we really drive that clarity
[31:50] at the beginning so that we can set it up for success as opposed to taking a great pilot that delivered great results and thinking that we could scale it across the entire organization uh as is. We found um on my side of the ocean
[32:19] we're going to start with a few pilots which is great um and a few areas where you want to change portfolios. However, you got to paint that vision of where you're going and have the leadership teams on board. And what we found in
[32:32] this case for instance is um there wasn't a single threaded owner >> or a type of group that owned it together. >> So that is something that we see where some projects fall. >> Yeah. That
[32:44] >> the other thing I'm sure all of you how many of you have AI sprawl where there's so much AI across the organization it's just super nobody. We're the only ones who struggle with that. I think uh that is definitely something I mean I have we
[32:56] have emails from the CEO that says stop the slop right like stop creating things for AI just to say that you're using AI how do we really think about the articulated end value that you're going to create and how do we make it easier for the teams to actually have
[33:09] discoverability of what's been built and so we've really sort of focused on leveraging Unity catalog and what are some of the capabilities how do we make sure that we're building skills that help the team uh be able to search for those for those things so that we can
[33:22] start rewarding ing building on the work of others as opposed to recreating the same tool, the same skill, the same uh uh system, product, etc. that someone else has already created across the organization. >> And and what I encourage people as well
[33:36] is collaborate with people you typically not collaborate with like um your communications team or your marketing team in business to actually you you literally have to come up with a way you
[33:49] you rephrase. For instance, in just very quickly on that AI one, uh the the scaled AI very passionate about that one that um was very painful. Um they um those sales
[34:04] teams uh when they heard they will have to do marketing qualified leads, they will have to adopt that. It was like an allergic reaction because they're used to a a business development rep. That's
[34:17] their job. I'm not going to do I'm not going to follow up on those leads. So if you if you if you position it that you're actually having an agent that's going to support you in this, you have to rethink how you're going to do it.
[34:29] But if you use old terminology, I found that that's also an interesting one that that's not always working. >> Yeah. >> So I Mel already referred to it. We're definitely thinking about now how do we leverage some of this insights that
[34:42] we've gained, the best practices issues that we've run into in the past. How do we think about now that hyperpersonalized? How do we make sure that we're doing more push versus pull or sorry pull versus push, right? We don't want to push the same training to
[34:54] everybody. How do we make sure that it's in workflow? We're serving up net new capabilities. Whether that means we can help one of our teams understand the product better based on looking at all the use cases they currently have open and that are not making the progress
[35:06] that they would want to. We're thinking about it not just from technology, but also some of the power skills. How did they do executive presentations? How do they do demos etc. And so we're really trying to think about how do we grow the capabilities of our workforce in a way
[35:19] that's already in the natural workflow that they have. So I want to make sure that we get to the what can you do now as you think about walking out of here. I think transformation is not new for any organization. I think we've all lived through digital transformation,
[35:32] data transformation. Hell, the internet came out and we all had to transform right into some capability. And so there are some fundamental components of any transformation. How do you make sure that you're thinking about like we just talked about that change management? How
[35:45] do you make sure that you have ownership at every level that you make sure that you're building that appetite for change? And if it's not an appetite, at least make people comfortable with the fact that things are changing. And so, how do you partner with them to take
[35:57] them along on sort of that journey? I think the other thing is starting to rethink the roles and the functions and how they exist. Um, and how do you make sure that you're getting really specific on what is the current maturity and
[36:09] capability of that function or that role and when that what's the aspirational um capability that you want that role to have ultimately across the business? Where do you think about automation or building out by default and then making
[36:21] sure that you have the right foundations and and platforms in place? >> 100%. And then on uh fluency, there's mixed views from customers on from from what we see here from many of
[36:35] you uh about AI fluency. Some want it across the board and others find it very limited because they don't have the environments ready for for them. Um but
[36:47] where I've seen bold leadership and going allin um we've enabled uh some big conglomerates where we train three 3,000
[36:59] business users in one go doing hackathons and business impact accelerator and and now um obviously you have to make it fun. Um you will see when you are going to our enabler
[37:13] section there right in Moscone West >> um we have the data detective game >> uh where we're helping business users to just find um the solutions in a different way. Um
[37:27] so making it fun and relevant to the different personas is critical. a a finance uh person, a uh you know, and and and what I find is you can actually
[37:39] not separate it anymore because you need those people with those insights when you build these agents. That that is so critical. So, um you can almost not not have those.
[37:52] >> Well, I think that's just it, right? You could have each of those components and you can do them very very well. But if you're not thinking about the people at the center, each one of those are not going to actually help you drive the transformation that you want to be able to drive. I think the other thing is we
[38:04] wanted to make sure that we left you with a little bit of call to action or what you could actually do. So I think how we would suggest that you approach this is uh in the next 90 days really think about assessing whether it's a
[38:17] self assessment, whether it's you think about the existing content that you already have, what could you actually pull together? What is the current capabilities of your organization? And like Mel mentioned earlier, each business domain or each persona within
[38:29] your organization may not be of the same maturity or the same capability today, but really getting a good handle on where do we think we are as an organization across these different business domains? And then aspirationally where do you want to be?
[38:42] What is where do you have you defined as an organization that ideal looks like or at least ideal progress looks like? Right? You may not know what it looks like three to five years from now, but where would you say that we have been successful in helping the workforce
[38:55] adopt and transform with AI and what does that look like? And then what are two or three priority workflows that you could focus on that are going to help you build the capabilities of of your workforce to be able to achieve that aspirational maturity that they want.
[39:09] >> Um, and then on I I think we're almost >> I think we're almost at time >> time. So just on the next one, Robin, what I >> uh would encourage I uh
[39:21] >> one area that I'm extremely passionate about is the uh capability focus now we've just launched the uh context engineer exam and uh I walked uh I I
[39:36] specifically went there to find out what they were experiencing and the one person came out and he said and he's one of our trainers >> and he was he said ma I've done ML all
[39:50] these search you can have >> and this was brutal because it was it really helped you understand the context of which you have to manage within this complexity yeah
[40:02] >> so that that was brutal in a good way not a bad way um but it was just fascinating so I really would encourage that um get a data and AI maturity assessment um that that's where we also
[40:16] can help you to look at what's the type of skills required. Um let's help you build those capabilities um be it together with us or how how you think of doing it and um every project is a data
[40:29] and AI project. >> So hopefully this was helpful gives you a little bit of a framework that you can start with gives you a little bit of comparison about what you see across the market and just know that data bicks is here to partner and support you as you're going on your workforce journey.
[40:42] So, thank you very much.

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.