bret-tushaus-audio-edit-raw
===

Bret: [00:00:00] I see things like this replacing traditional reporting. If I can just have a conversation with my ERP and say, Hey, tell me the projects that have the highest amount of accounts receivable, or what's the current profit percentage or effective multiplier on this project? Goes a long way to get data in front of people in a much easier way, and if it makes a project manager more effective, that's gonna pay dividends as well.

Welcome to launch. The show from Log Rocket, where we sit down with top product and digital leaders. Today we're talking with Brett Tess, house, VP of Product at Deltech. In this episode, we discuss how Deltech built Della an AI agent framework, powering everything from smart summaries to autonomous accounting, why AI is rewriting the role of product management and what leaders needs know now, and how Brett's team leveled up their AI skills fast with short, high impact sprints focused on real tools and real problems.

So here's our episode with Brett Tess House.

Jeff: Hey Brett, welcome to the show, man. Thanks for joining. How you doing today?

Bret: I am doing great, Jeff. Thanks for having me. I'm excited for our conversation today.

Jeff: It'd be fun. , I'm stoked we get to talk about some AI stuff, some some product. One of the fun things

you start out as an [00:01:00] architect the actual like. Building big buildings and stuff. But now you find yourself in product. You've been at Dell Tech for 15 years. We're gonna talk specifically about how you guys built some really cool AI functionality there to move the team forward , and how that has fit into product in such a vertical focused area.

But first maybe can you just give us the quick, spiel on how did you go from being an architect to running product at , a company like Del Tech, and what does Del Tech do exactly. So people have that kinda context

Bret: As you mentioned, I'm trained as an architect, worked as an architect for about five, six years. And then sort of transitioned to the technology side of an architecture firm where I was a Dell Tech customer. During that time, got to know Deltech well, struck up some relationships there and, and actually ended up, you know, joining Deltech as a. As a product manager 15 years ago, . And the reason that they sort of relate is Dell Tech focuses on ERP and complementary solutions for project focus businesses and architecture firms fits squarely into the project focus business area as well as like engineering, government contracting, consulting agency. So that's sort of my [00:02:00] history, you know, what Del Tech does and, and how I ended up at Del Tech.

Jeff: You went from being the customer to knowing the customer so well that you got, that you were really, really well suited to help them build. Great solutions for the customer , and seems like it's been a successful decade and a half so far, which seems rare in tech, but it's always cool to see 'cause like you built up a lot of experience in this area.

One thing that really drew me to this when we were talking early about maybe having you on was the idea that like, there are a lot of companies like this and I, I think at times there's a, you know, everyone knows about like. The hot, sexy in notion open ai, all like the very producty companies, but like you guys are doing cool stuff and we would be remiss to miss that.

So you guys built de ai which has been a very powerful way to kind of pull a lot of . What the company does into kinda like the next stage of, of efficiency and moving quick.

And, you know, what, what is it we always say? , using AI to move faster, be stronger, be smarter so , let's just dive in. Where was the star and kinda how did that come to be , and what did that [00:03:00] development look like over the past?

A little bit of time.

Bret: Yeah, it's been pretty exciting and I, I agree. We're doing some pretty cool stuff. Um, and, and sort of The evolution is interesting because, prior to CHATT and the onslaught of generative AI and, and all that, we had more traditional narrow AI capabilities in our solutions.

But, they were very, specific task oriented and, very focused on specific functionality. And, when generative AI hit the scene. , just like other enterprise software companies, we started dabbling in various ways to leverage the technology and it was kind of all over the place as the technology sort of took hold and, as we sort of honed in on the, the types of capabilities we felt we could introduce to provide value. We felt we needed this sort of overarching brand. And looking around, at software , and the various sort of AI companions that were out there. We wanted to brand it similar to how other enterprise software and consumer software were doing. And that's really , the conception , of D as our. Overarching AI brand, , the Digital Business Companion , as you mentioned, and it's not just one feature, , [00:04:00] it's a whole host of different features of functionality that all sort of fall under that that Della umbrella. So that's how we got into the Della world and. Like you said, we introduced , some pretty cool , and interesting functionality under that de umbrella across the entire product portfolio.

Jeff: A lot of companies at the, even up to current time have taken kinda the route of slap a chat agent into something, say it's ai, and like now, now you're AI first, or something like that. And sometimes , is the right thing to do and sometimes it's not the right thing to do.

Like, you guys took a very. Focused look across the whole business of, , four very distinct pillars of Dell's capabilities. And with that, like took a very nuanced look at what do customers really need?

Where you go, where can you really add value with ai? Not just, oh, we're gonna throw AI on, on a name and like say where ai, maybe first step here is let's go through like what those pillars are and, and how did you and the team kinda determine, hey, these are the areas that we're gonna get value 

Bret: The one thing I would say, just foundationally off the bat, it, it sort of relates to one of the [00:05:00] comments you made in terms of slapping AI , on something from a. Marketing perspective , and putting it out there. One of the things , that Delta has already been focused on even before, generative AI , took such , a mainstay was what we call purposeful innovation.

Anytime

we talked about innovation at Del Tech, we always prefixed it with purposeful innovation. 'cause let's face it, you can. Introduce tech for the sake of tech , and it not necessarily providing value for people. And we were very deliberate about that. We didn't wanna bring tech to market just , for the sake of technology.

And , we

approached AI in the same way. So , the pillars you mentioned, three of which are, well established, the fourth of which is becoming established as a little bit emerging in some ways. But, we tried to identify those , four pillars in a way that really was purposeful. One of the first ones not surprising in the generative AI age , is generating , and automating what we call smart content. The great thing about , the Dell Tech product portfolio is we have a massive amount of data in our solutions that are sort of the byproduct of our customers delivering projects. And we [00:06:00] can leverage that content to generate proposals to, distill a large amount of information in a natural language way. So it, it really was the right place to start , when generative AI came on the scene. In fact, we have this concept called smart summaries. We know that a lot of people in our customers organizations don't necessarily know what report to run or where to find information.

So if we can just take information using generative AI and present it in a natural language way,

that's gonna save users time, it's gonna save the step of someone going to someone else asking for information. So that's one of the ways that that

pillar has, has come to life. 

Jeff: , that pillar, I feel like people often oversimplify because , it seems like, oh, just summarize something or just pull a couple little notes or something. And it, it sounds easy and I think likelihood is to get it, 50% good is probably really easy. To get it, but like, especially government contracts or on our end, we, we found, you know, value in summarizing sessions, for instance, and pointing out to people the key areas within a session.

'cause, you capture so many of 'em, you can't view 'em all. How do you speed that up? It seems like a key thing here [00:07:00] has been quality though. How do you, not hallucinate, but also how do you accurately pull out like what you really need to summarize and to get, like I said, like 50% , to really good.

You can probably just slap a open AI integration in, call it a day, work the prompt slightly and you're good to go. But it seems like that's, you know what you're talking about , these things are probably pieces where you want high fidelity. You want people to trust it 'cause you show one or two wrong things or false positives and they're gonna walk away forever.

Bret: Yeah, I think you hit it spot on when you said walk away forever.

Because that is a risk, you know, somebody uses one of these tools and it, it hallucinates . You know, it doesn't provide a lot of value , and chances are they're never gonna come back. As cliche as it may sound, that first impression is so important.

So,

We have spent, a decent amount of time 'cause we have the smart summary feature in, in several of our products. And in each of each of the products we've spent a lot of time , refining the prompts, testing with different models because. As new models come out and things like that, , the responses change. So on one side of it, we are spending a lot of time refining prompts and testing. But then also obviously the data you serve up to a mechanism like this needs to be the [00:08:00] right data. It needs to be delivered in the right way. So it's taken , a lot of fine tuning to, I think we're well beyond 50%, 

you know, top 20, top 10% with our

responses. And I think, even the part that maybe falls in that 10%, it's not that it's wrong, it's just not, might be as valuable as some of the other stuff. So

It could have been easy, but, we wanted to make sure we were making a good first impression so that people

continued to rely on, on this type of mechanism.

Jeff: Okay, so that's one. And then, but it gets, it gets cooler from here even. So like, I think there's something around like project prediction , or success prediction or something like this.

Bret: As I mentioned, Dell Tech serves businesses that deliver projects so. Anything that we can do to help predict ensure that our customers are being successful on their projects ahead of time, that's very, very valuable for our customers. So this is a little bit more on the machine learning and sort of algorithmic side of ai, but it definitely weaves in some generative AI capabilities as well. The whole idea here is, you know, anticipate future outcomes, predictive, that type of thing. There's several examples of this, as you might imagine. We have a lot of project manager [00:09:00] customers and a lot of project management capabilities. So being able to have that, , line of sight out in terms of, , what's my estimate to complete profit?

What are my projections from an effective multiplier perspective. All of those are. Lines of sight into whether a project might be going in the wrong direction and being able to course correct. But then most recently we've introduced things like what's called a smart fit score.

So, one of the solutions we have in the market is called Government iq, and it helps companies that do work for the federal, state and local governments. Identify opportunities they wanna pursue. And this SmartFi score can look at their capabilities, compare it to what's in those opportunities and whatnot, and say, Hey, you know, you're, you're 80% fit for this particular opportunity.

It might be worth pursuing. Or, for these reasons, your capabilities don't , align well with this one. So, you might wanna look somewhere else. So that obviously saves our customers time just in looking around, but also in pursuing . An opportunity that they might not be well suited for.

Another one we also have our, our Dell Tech talent solution, which is, which is a human capital

management product, and that has applicant tracking and recruiting and [00:10:00] things like that. And. Like a lot of HCM solutions are doing these days with generative ai, we can do sort of a candidate relevancy score.

So on job description based on, those types of things, you can say, Hey, , these top 10 candidates are probably the ones you wanna take a closer look at. those are some of the ways that this particular pillar, pillar, predicting success , has come to life.

Jeff: Looking at these two things, one is basically take data, summarize and help people access data.

The other is kinda the other way around. Ingest and tell you if it's a good fit , and make sure you kind of focus in the right direction. what problems did you guys run into here like back to the idea of it seems very easy to get 50% good at these things and 50% good basically just causes your product to fail.

What were the pieces that really unlocked, like, this, really got some 50 to 85% faster or were there no big aha moments? It was just constant iteration and training , and just getting a little bit better every day.

Bret: I don't know that there were big aha moments. I sort of covered one of the challenges , that we ran

into and, the, like the smart summaries capability, you know, we did that. Two and a half years ago I think is when we

started that. So you can imagine, some of this technology [00:11:00] was a little bit nascent at the time and it was sort of not really knowing what to expect and quite frankly, not knowing what to adjust.

I remember I said before we were well below 50%, , and I remember early on in smart summaries, the models would, if the data wasn't there, it would just make stuff up. , it was

creating projects that didn't exist in the system and, things like that.

So, not necessarily aha moment, but realizing that, you really , gotta adjust the settings as you're using these models. You gotta test, you gotta refine et cetera. , I think we underestimated the the level to which we were gonna have

to do that to make this stuff work.

Jeff: , there's almost a level of training. Some of these models, it was just like constant reinforcement learning for it. So I wonder if it was, did similar process there of just like you had just whack, whack 'em all through that.

Bret: We've learned that we had to, if there was any you know, even maybe a little aha moment,

we learned that we needed to be far more prescriptive

than I think we we expected to. So that's sort of similar to your comment on the reinforcement. 

Jeff: I found that getting these functions to work, sometimes there are these like magic. Light bulb moments, and sometimes it's just [00:12:00] diligent refinement. And usually it's both or usually, usually you don't have the aha moment without some very diligent refinement as well.

But like we built what we call Ask Galileo, which is in our session we play analytics and kind of end-to-end product feedback. Connection at this point. It is a way to, to query all this kind of non-structured data.

And what it does is, has the, has an agent we built actually watch all these sessions and talk to you about findings from , what would happen if you could watch 20 or four hours of sessions, but you could do it in like 30 seconds. And the interesting thing was, same, same as you talked about early on when we were testing internally, it made stuff up.

It, you know, told me. Metrics existed that didn't and, and other data points didn't exist when I was like, I am looking at it right now and telling you what what it is. And the biggest unlock we had was we realized we had to train it like you were onboarding an an actual pm.

So we, we call it now like an IPM, but the unlock was, we trained it like you would a pm. 

Bret: Yeah, it's kind of, the prescriptiveness comment reinforced. But even, I think back , to your last comment [00:13:00] there, it's gone from entertainment to fascination. It was entertaining to see, it was making stuff up. It's not what we wanted, but it was entertaining.

Jeff: Comedic. Yeah.

Bret: And then the fascination today of, okay, now that we've, been more prescriptive and really fine tuned and whatnot.

Yeah,

it's fascinating what it's capable of. Absolutely.

Jeff: So, it seems like the idea of kind of like intelligent exploration or I, I don't know what you call this function actually.

Do you actually call it ask? I think,

Bret: We do call it.

Jeff: yeah. So we call it Ask Galileo. Yeah, we talk about this. You guys call it Ask, ask. We laughed about that earlier. . So how does this piece kind of work for the team or for users?

Bret: As you mentioned, the pillar is intelligent exploration. That's , what we call it. But , this sort of gets back to one of the comments I made earlier too. Let's face it, people that use enterprise software, there's a whole range , of user skills and

levels, and certainly not everybody knows how to run a report, how to configure a report, where to find data. Often people go to ask, their account, their control or whatever the case may be when they , are looking to find some information. So this intelligent exploration , is really intended to. Help facilitate that better , and [00:14:00] make it so that people can find what they need through a natural language conversation without having to know how to run a report or go ask someone else that's really what Ask De is all about. In fact, I don't think this is a, a novel idea for me, certainly, but I see things like this just replacing traditional reporting. Because if I can just ask a question, have a conversation, if you will, with my ERP and say, Hey, tell me the projects that have the highest amount of accounts receivable or, you know, what's the current profit percentage or effective multiplier on this project? Boy, that goes a long way to get data in front of people in a much, much easier way, and. If it makes a project manager more effective or someone more informed in an organization, that's gonna pay dividends as well. So it

sort of pays dividends on, on multiple levels which I think is, is huge.

And that's, yeah, that's really what our Ella component

Jeff: Yeah. , the chat is interesting 'cause we, we, as an analytics company, we have had a lot of conversations internally about chat interface analytics and as. , someone on my end , who I've run, operations and marketing teams , and been deeply involved in analytics for, 20 years now.

I always kind of go back and forth on, on this piece [00:15:00] specifically cause like you described, right, if you can just ask for the metric , or the graph or the chart or , the figure that's obviously a lot easier , and it opens it up to people who would not necessarily be able to do it before, without understanding how to use the UI to do that.

What's your thought on. I have a theory that like power users are always, or for longer probably than not going to continue to want to use some form of UI because. There's always some weird, , every, I don't know, every company I've ever been at has some weird little data problem , or cleanliness issue that, you know, like, oh, we don't use this version of that field, we use this version of that field.

Or even in a ui, right? Trying to like filter between break it down by this, now break it down by this, like, you can probably do some of that work faster in a UI potentially, if you're going really deep in like a power user. But as a. casual user. The, the conversational interface seems like 10 times better.

Bret: I, yeah, I would agree a hundred percent, but we're still gonna have those power users that.

Probably don't want the natural language conversation. They wanna see things in a more traditional way. So it's necessity to sort of meet both ends. 

Definitely I think as the technology evolves I [00:16:00] see being able to fulfill both needs probably through the same experience, if you will. Right now maybe it's a little bit less towards the power user and more towards the casual user, but as these things improve and the capabilities and output capabilities improve, like I said, I think it can fulfill both needs pretty effectively.

Jeff: One kind of general way we've looked at this specific piece is.

Can we a, provide people the ability to do things that you couldn't do manual wi with this kind of chat interface. So like, can you instruct, for instance, , the AI agent to watch, , a hundred hours of sessions real quick, but there's also then can you like build a chart?

It's also every kinda novel question that's not going deeper on something. But , is someone asking something new? Is like, was this a chance for us to. Surface something we should have proactively, like should we look at this as a learning moment where we could have surfaced this and known to surface this proactively.

Bret: The way I like to look at it, and I often talk to customers about this when we talk about this, this concept of intelligent exploration is, yeah, it's artificial intelligence, but it's also the intelligence of everybody that's come before you.

So, these large [00:17:00] language models, have massive amounts of data around how project managers operate and how project managers communicate. , how you communicate with customers, what are good numbers, what are bad numbers, those types of things. So it isn't just about necessarily getting , the hard and fast answer based on the data. It's also, the intelligence that we can add around it. And in other words, hey, this, this is a simple example, but hey, this customer's. Aging is, over $250,000 and we're at 82 days. Boy, that's, that's probably a red flag. Why aren't they paying versus a high AR amount? That type of thing. So, it's as much about AI as it is about the, the intelligence of, those that have come before us.

Jeff: Those are the three. Kinda more mature aspects of what de AI is delivering. But let's get into , the frontier stuff. It sounds like you guys have done some work kinda on the agent front as well, and maybe that's not as, as production ready , or as mature, but I'm curious to see , , what are the use cases there?

How does that look in a company that's as vertically focused , as you guys are?

Bret: , we've done some, what I would call sort of more narrow minded agents. Things like filling in time sheets and expense reports and things like that. So [00:18:00] very, like I said, very focused task orient agents. , we're sort of expanding to, or, or evolving to a bit more autonomous agents and a bit more you know, I love the, the recent term that's come on the scene.

Ambient agents , and a type of agent that reacts to events and even engages with the users to make decisions or take actions along the way. I think a, a great one that we're very focused on and will lead certainly to other things is the concept of a month end close agent. Defines the activities that are involved in it detects anomalies does some reconciliation, analyzes, some subledgers.

All of those types of tasks that typically go into, say, clothing and accounting period or , a month. building that into an ambient agent that does certain things, analyzes certain things, asks you for help when it needs it asks you to finalize things. As you might expect , in a lot of the software that we deal with, human in the loop is, is absolutely critical.

You know, when I talk about an agent that does month end close, and there's accountants in the audience, their heads wanna explode so,

Having a human in a loop with something like an ambient agent is absolutely critical. [00:19:00] So, yeah, this is like the fourth pillar that, yeah, not quite as ma mature, but , we've introduced some things and are, really looking to build out , that agent framework. And those more intelligent agents across the, the product stack as we speak.

Jeff: We had someone from a med tech company on recently who talked about the ai agents and assistance. They've built that basically. It with the doctor in like, in a, a patient appointment and take notes and, , take action items from the conversations the doctor can be present or summarizes medical notes leading up to, so they don't have to read three, four, or five pages before going in.

They can just kinda get the synthesis and, and walk in prepped much faster. And she brought up similarly, like the biggest thing was there's just so many things that an AI cannot action. It has to ask for permission. And that's been really key There is like, it seems, maybe a, even a little more consequential than getting accounting wrong is like medical things, but they're both high severity. You want a human in there who's responsible ultimately , for making the call on something. So, but it can make their life still a lot faster and easier.

Bret: And it's audit trail too, right? It's not

just human in the loop making, but also explaining, hey, the a as the [00:20:00] agent, this is why the agent came to that conclusion, or is recommending this action. So that it's easy to, give that approval or, or take that action for the human that is in the loop as well.

So,

Jeff: Exactly. 

Bret: part of it.

Jeff: So going through that, what led or, or how did you guys come, how did Del Tech come to, these are the areas that are going to add, significant value for customers were there examples of things you looked at? Go like, this could be an area? No, actually this seems like it would just be slapping a, on it to, to say ai.

Bret: Yeah, I, I, I sort of would, would reflect back on the, the purposeful comment I made earlier. We tried to look at the workflows that our customers go through, within our tools that we felt were maybe the most laborious, but at the same time the easiest for a non-human to take on. That was sort of the, the lens that we looked through as we figured out, okay, where are, where are our opportunities? Where are we ripe to sort of leverage this type of technology? So it was a little bit less of. You know, that's not the greatest idea in terms of, of, of slapping on ai. It was more, let's find those very high value, very resonant pieces [00:21:00] that we can focus on.

So, that's really what's been driving a lot of our decisions. And when you think of, intelligent exploration and content generation, and then of course some of the autonomous stuff to as cliche as it may sound, free up our customers, our end users to, to do more value added things.

That's, that's really been our focus.

Jeff: One thing we talked about earlier before, before we were on the, on here recording, was given the history of Del Tech and the number of projects that you have had access to , the data from there is this like giant data moat. Of so much insight you have from, from being able to train on all these projects, tens of thousands of projects over years and years, that it's also kinda looking for where is that kinda wisdom going to be best applied , and where can you make a real difference in workflow that others can't copy really, because you have this level , of data to train off of that, that no one else has. .

Bret: Yeah, you hit it spot on. So many decades worth , of project related information from our customers. And of course, we do respect data privacy and anonymity and all of that kind of stuff when we think [00:22:00] about this. But we do have a unique opportunity. Again, like a little bit of cliche, data's the new oil, but T Tech's got a lot of data from from their customers and we can do some pretty cool things for our customers as it relates to, what we can learn from that data and insights we can pull from that data.

Jeff: So I, I do wanna switch gears because , I could continue to ask about all the nuances , of agent tech that you guys have done stuff like that for, for quite a while.

But I think one question that you probably have pretty good insight on is there's a lot of people who are doing and who are quite advanced, like, the team over at Dell Tech.

But there are equal, if not more people who have boards or CEO saying like, just do AI now you need more ai or who want their team to start using it. . But can't figure out how to either frame where would AI work or how did they even get started? So maybe let's start on, on the actual just functional teams themselves.

How can teams just even start to understand that stuff like this is something that's going to transform not just verticals, but product management itself quite likely.

Bret: Oh yeah, absolutely. I mean, it's, it's scary, it's [00:23:00] exciting, it's uncomfortable. It's, opportunistic all at the same time, which is, it's an interesting time We're in. And by no means do I think we or me or Deltech has it all figured out. We still have a long ways to go, but, you know, I feel like we're doing some good things to prepare our teams and to, , put us in a good spot for the transformation.

So, you know, just a, a few things that I, I think are relevant. , you made the comment about boards just saying, Hey, we need more ai. Our leadership is certainly saying that for good reason. I think there can often be, and this isn't just related to product teams, I think this can be pervasive elsewhere, but there can often be sort of this paralysis in terms of, okay, we know we need ai. But there's so many tools out there, there's so many techniques. , and it's moving at a pace that it's, it's unreal, where do I start? And, and people often think, well, I need to figure all this out before I can jump in. And, you know, one of the, one of the messages I'm constantly carrying to my team is just do it.

As cliche, again, as that may sound find a tool that you think might be interesting and see what happens. Because if you wait till you figure it out, or until you [00:24:00] master it, guess what? You're, you're never gonna make any progress. So. I think that's probably one of the biggest things that we're putting in front of our teams today is, if you got a thing that you can have AI help you with, give it a shot, find a tool, talk about it with our team and see if it's something we wanna make. More sort of standard across the team. So , that's one of the key things we're doing. to be a little bit more prescriptive, because, I understand, you know, some people need examples and use cases to solve is I've been doing these sort of AI call to actions , with my team where through my own exploration and whatnot, come across tools, come across techniques, and I'll put them in front of my team and I say, guess what?

I want everybody in the next, I don't know, 30 days to go out there and use each one of these tools.

Come up with a use case, just try it , and see, , what it can do for you and share your stories, your experiences with the broader team so that we can disseminate this to the

rest of the team.

Those are sort of looser ways to approach this. I will say that, you know, at Deltech we do have a, a more formal look at how we're looking at product management, ux product marketing and development are gonna work together in this new AI augmented world.

And, , we're looking at that in a little bit more [00:25:00] formal way. It's definitely evolving and changing as we look at it, but. You know, we have this concept, and I didn't coin this term called fusion teams is what we're calling it now. So, you know, the, the mix of engineering, product management and UX primarily, how they're gonna work together in a more fusion oriented way moving forward. Think of it sort of the next generation of agile, if you will. Fusion is the AI augmented version of Agile. We're trying to figure out, the techniques, the processes, the tools to support that and whatnot. So that's been a big push that we've been trying to focus on and progress with as well. 

Jeff: I love the, the idea of here are tools that seem interesting, go use them, go try them. And, and, go make something happen because,, at some level I think it's, it's tough where you can have, there's, more tools every day than you could possibly ever touch.

The thing I've tried to talk to my teams about is. When we're doing something new or, or even, you know, when renewals come up, maybe for tools that we use, think, , do we need this? , or could we do this better, faster, more efficiently, kinda without this?

Bret: I think you've hit on something that, know that I thought of as explicitly until you were saying it. One of the reasons I do the [00:26:00] AI call to actions is that it hits everyone on our team. And I

think, you know, we're in a, we're in a spot of some democratization of, of this technology that some of our best ideas on some of this stuff might come from places that are not expected or not, sort of the

usual suspects, if you will. So, liberating people or giving people permission, if you will, to just go out and work on this stuff. Test the stuff, find tools and whatnot. In a very. Liberated way I think is a potential for big dividends. Because like I said, that next big idea might come from a place that, that we're, we don't usually expect it to come from.

So I think that's that's key there as well.

Jeff: I get to travel in the country a lot and talk to product leads and it's been interesting seeing, I think there's a commonality of the companies that really have got their teams really moving is, is at some level you can, you can reduce it down to just. Just be like, Nike, just do it.

As trite as that sounds like, just try. , like, the worst that's gonna happen is you're gonna fail and you learn and then you try again and you get better. Before we go, any good parting wisdom here of how people should be looking at this stuff.

Bret: I think you captured a little bit there with sort of your comments that you just made around sort of the, [00:27:00] just do it approach. But , I guess what I would say, . There's all this talk of, AI replacing jobs and that sort of thing, and I think that's a real thing that we're gonna start to see come to life.

What I say to my teams and what I would say to anyone at this point, boy, if you're not using ai. You should really be jumping in and figuring some of this stuff out. Because if you wanna end up on the winning side, , I think you need to be aware, you need to be equipped. You need to understand sort of the, are the possible. Maybe that's a common sort of perennial recommendation these days, but I think it's more important than ever. And then, , the other comment I would make is. We talked about a crossroads. We can talk, we can call it an inflection point. we can call it whatever you want, but product teams, engineering teams, product marketing, ux, it's gonna be dramatically different for those teams in the months ahead. And I, I purposely use the word months because of how quickly this stuff is moving. And like I said, it's a little scary, but I think it's also very exciting and I think those of us that. Really work to figure out what that new model looks like are gonna be in a very good, spot, sort of the pull position in that [00:28:00] situation. So I think people just need to recognize that, hey, from product teams, things are gonna transform , and you need to be ahead of that and proactive so that you can be in the right spot, months down the road , as some of these changes start to take shape.

Jeff: Brett it's been a pleasure, man. This has been so interesting, so fun to hear about both what you guys are doing with Dell ai, as well as kind of like the general move of how products should be looking at this. If people wanna follow up, ask more questions chat with you more about AI or other stuff I assume LinkedIn is probably a good place , to reach out.

Bret: LinkedIn's probably the best way to get ahold of me. Definitely.

Jeff: otherwise if you're listening to the episode and you liked it and you thought this was useful we'd love for you to give us a follow, subscribe to the show.

Give us a, like, write a review if you're on Apple Podcast or if you're on Spotify. It really helps us get the word out. That's how people discover us and how this kind of thing grows. Help us keep doing this , and help us get the word out , on great product leaders.

Doing awesome, cool product things. Brett, it's been a pleasure . Thank you for coming on and before to stay in touch.

Bret: for having me, Jeff. Really appreciate the opportunity.