LaunchPod - Eric Anderson
===

[00:00:00] Hi, Eric. Welcome to the show, man. Good to see you again. How you doing? Great, Jeff. Thank you so much for having me. Thanks for coming on. I'm excited for this one 'cause I think it's been interesting where we've seen this major change as AI has kinda changed how we worked, more and more people getting back to practitioner level and some of that changing, and you have great insights around this, and I think have experienced it as much as anyone, but a great story to tell around it, so wanna dive into that.

But just to set context and introduce you real quick, how'd you get from starting in tech up to SVP of product over at DataSite? Yeah, it's interesting how everything old becomes new again in terms of being a leader and being in IC and how that all flows around. But to answer your question, how did I get here?

I think, like so many people, my path into product was serendipitous. I really never had any aspirations to get into product until I learned about what it was. It's not like I woke up one [00:01:00] morning and said, "Someday I'm gonna become a product manager." I happened to be in an account management job, basically the other side of the sales coin, and, uh, had a client in town here in Denver and, uh, was in their building every other week doing demos.

And at one point, their SVP of product, again, everything comes full circle, asked me if he could take me out to lunch, and I assumed he was gonna wanna renegotiate our hourly rate, but instead, he explained his business model for this product that we'd been working on, so I was fairly familiar, and he said- I really need a leader to take over a senior product manager role for this product, and I think you'd be really good at it.

And I told him I was honored, and I didn't really know what that meant, but I'd like to learn more, and the universe just delivered it, and I've worked in product ever since. It's always so funny to see that because I feel like everyone I talk to says like, "Oh, I have a non-traditional path into product."

But I think in all honesty, the person who I meet one day who said like, "Oh, from the beginning, twenty-five years I've been a product manager," that would be the non-traditional path into product 'cause everyone has [00:02:00] kind of fell into it. Because when we both started, there wasn't really a product role. It didn't exist back then, or if it did, it was a shell of what it is now.

But I do think that said, of all the stories I've now heard, going from sales into product probably is one of the more unique ones. Yeah, that might be true. It's interesting, though. The team that I was so thrilled to take over in that job, to your point, nobody came from product because it wasn't a thing.

They may have been in sort of embedded analysts. That was probably the closest thing to a product manager. But most folks came from customer service or design. I digress, but onto, you know, you started out in sales, you moved up and got this kind of serendipitous move into product and continued through education and, you know, multiple different industries until you find yourself now at DataSite, which is, for people who don't know, basically virtual data rooms, right?

So very high privacy, very high trust. But the interesting thing is you can go from that industry to e-commerce and almost anything in between, and I think this experience [00:03:00] of how AI is changing how we work is true. So maybe we just dive into that. Let's talk a little bit about what you guys have experienced, 'cause I think this view of it used to be, you know, you come in, you're strategic, and you have a bunch of people doing work, and that has changed.

Like, how have you guys experienced it over there? What's that look like at DataSite right now? Yeah, it's been as transformative and disruptive as it probably has for everybody else. You know, we talked previously, and I talk all the time when I'm interviewing folks or coaching folks about climbing up and down the ladder, and I don't mean the corporate ladder in terms of role.

I-- You could think of it as the very bottom of the tech stack up to where the customer meets your product. You could talk about it in any sort of technical architecture sense. You could talk about it more pragmatically from going from customer service to sales to design. But as a product person, you're moving around the org all the time.

I often say I go from rolling up my sleeves and getting my hands dirty with whiteboard ink all the way to, "Oh, I need to go wash my hands and throw my sports coat on 'cause I'm gonna present to the board." And that's just a [00:04:00] very common sort of whiplash-y type of day in product, and that remains true. But what's changed over the last, we'll say year or so, and especially the last six months Is climbing up and down the ladder, as we said, is more like riding a high-speed elevator.

Yeah. And it's really collapsed the org chart. I'm sure your listeners and viewers know this, but there's a lot of LinkedIn and news article activity and literature right now about the flattening organization- Mm-hmm ... and how the flattening organization is better suited for what's happening with AI, and I'm finding that to be true, and it doesn't necessarily mean job loss or minimizing people's role.

I think it's quite the contrary. It means that everybody, increasingly, if they were a leader who did no IC work, is going to have to do a lot of IC work to be successful and to navigate this. And similarly, if you were an IC and you felt like you were doing a lot of busy work, the expectation now is, well, the busy work gets done for you if you've got the right ideas.

And so there's a real leveling of that playing field and a real leveling of [00:05:00] the expectation of how you spend your precious time every day. Yeah. I remember many years ago now when I had my first big company director job, and there was one or two of us who were kind of new to big team leadership, and we all would get together.

And I know a lot is getting done, and I know I help the team move and tackle the most important things, but I'm not doing the work. I just spend my day talking about work. And it's interesting because I always kind of chafed under that and pushed where I was getting involved in something, and always had, like, spikes I was gonna work in and tell the team, "It's not you, it's that I can't possibly be my best if I'm not close to something here to keep understanding, keep sharp."

But I definitely saw some people, peers move the other direction where they embraced, "Oh, I don't do it, I just plan it." And I see those people, you know, when I keep in touch with them, struggling a lot more because, like you said, you've always kind of gone up and down. If you didn't do that, though, you might find yourself in a rough spot right now.

The pure talkers. I don't know. No, you're absolutely right. I think that'll play out across different functions. But if you think about your general archetype of a designer or an [00:06:00] engineer, they have that job because they like to do stuff. But people in product, you could sort of pick that path. I think you're exactly right.

When I flash back to that job that I got nine years later, when I decided to leave, my wife and I were talking about it, and I said, "You know, I don't get to build stuff and deliver stuff anymore. I'm in a sport coat. I'm on an airplane. I sort of stand up, and I do a quarterly big dog and pony show," and it's really important.

I don't mean to minimize what I was doing. It really helped connect the work to the business and the planning and the investor market. But I am somebody who likes to get my hands dirty. And so as an example, when I first started paying for my own Claude over a year ago, I immediately started thinking about small businesses I might create, not to really operationalize them, just to see if I could come up with something using these tools, and it really re-inspired me.

It's really fun. We tell people, "If nothing else, just dive in. Just start using these tools." We've got tremendous access and materials to self-train, [00:07:00] and by the way, if you ask Claude or OpenAI to train you on using them, they will. So it's just been an incredibly inspiring time as somebody who likes to make stuff, to just dig in and throw your title aside and just start to build stuff.

You're diving in. You're doing more within DataSight. What is this kind of change of how people operate? What does that look like across the product team and across how they work with you, and how has that changed how people do stuff, and have you found those gains of, you know, harder, better, faster, stronger with AI yet?

Or is the team kind of getting there still? Yeah, we have found gains, and something we'll come back to in a moment is we've also figured out where the next set of fail points or pressure points are, which I think is natural. But what I like to do to start the framing of this when we're talking with teams, you know, we're doing a lot of piloting of new ways of working but also new feature builds and new imaginations of feature functionality not yet in production.

Mm-hmm. So almost like a full stack experiment [00:08:00] as a pilot, and so then as we broaden out and we talk to other teams, "Okay, here's what we're learning," one of the things that almost has become kind of the calling card or the headline of everything is the contract has changed, and what's so cool is that didn't come from leadership.

That came from one of my senior principal PMs. We were just a couple of weeks into a pilot, and he said those words, "Eric, the contract has changed." And I said, "Man, tell me more. What do you mean?" And what we're finding to be true is rather than things being process and training driven, and here's how we do things around here, it's almost like the product person's dream, where the catalytic moment is from the idea, whether the idea is around a problem that we've had trouble solving in the past or a solution that somebody's been dreaming of building but can't.

Mm-hmm. And that becomes sort of the icebreaker, the thing that goes first, and then the ways of working just fall into line around that center of gravity. And I think if you ask people like us who got into product for a [00:09:00] reason how we'd like to work, that would be it. No entrepreneur in the fabled garage when they come up with their idea says, "Man, I can't wait to come up with a cool process to deliver this thing."

They're passionate about delivering the thing. I can't wait to sit a awesome quarterly planning meeting- ... and really get into the meat of what we're gonna build our KPIs around and OKRs for the next four quarters. Yeah. That's right. I can't wait to wrangle my PRD into this new PowerPoint template in 8 point font Right?

But what we get excited about is, I think I've got this idea, and the ability to move from idea to building a foundation to execution in minutes or hours is almost exactly why we like this job at some point so much in the first place. So to answer your question directly, what has changed is the PRD now, as an example, we formulate in the first meeting, call it an integration zero meeting perhaps, and the product person or the designer says, "Okay, here's what we're thinking.

We want to solve problem X," [00:10:00] or, "We've got this idea that we've been batting around that we think would be really cool. Let's call it solution A." And in that moment, if you're using perhaps Whisper or even Claude's audio or OpenAI's audio feature on desktop or mobile, we've now got it where it authors your PRD on the fly while it's listening to you.

The lines between engineering, UX, products, and other folks that might be involved get increasingly blurred, and everyone just kinda jams on an idea and decides what they wanna do coming out of that. And the process just increasingly falls in line. Now, that's not to say it's perfect, and it's not to say that we've got it exactly right yet.

Mm-hmm. But we're almost to the point where you can come out of that iteration zero-type meeting and you've got your PRD, you've got the elements for engineers to come back and ask the harder questions, for them to set up an environment and to start to do everything that takes the prototype directly into code and, even if necessary, pulls from real live data so that we're not mocking things up and building [00:11:00] vaporware.

And if I think back to when I started working in products, I mean, that would've been the dream. But instead, we built a prototype and hooked it up to dummy data and then perhaps linked it into a PowerPoint. Like, if you were lucky, you built a prototype, right, you know? Right. It's funny to go through, like, where people are and just see...

And I started listening back to old episodes because we haven't been doing this that long, right? Like we started this podcast, I think I'm gonna guess and say late '24. You know, AI has been a theme throughout the entire thing, and it's been interesting to kinda go back and listen to a couple old ones and just every couple quarters pick one, because what was revolutionary 12 months ago is now blasé.

You're like, "Oh, I mean, we would've used to like to make, uh, just a prototype, but now we're doing this thing." But it wasn't even 12 months ago where people were mostly ecstatic to be able to even build a real prototype and just get out of flat files in Figma. And that's how fa- like, you know, the contract had changed.

I think it was one of the undersells of the century. It's really accurate, but it's also very easy to take something that [00:12:00] pithy and concise and miss the magnitude of what has changed. Yes, absolutely. And I gotta ask 'cause this is always one of my favorite topics. If you look at LinkedIn or any of maybe those channels where you get all the influencers and stuff like that, I am assuming you have not replaced your entire team of 40 product managers with an AI agent.

What is the difference actually for these people that are on the team day to day? Yeah, it's a great question, and I think especially given the audience of your podcasts, this is critical. It's-- Again, if we move away from LinkedIn or, you know, people always talk about Instagram as where people live their fake lives, right?

The theater of AI. The theater, right, and LinkedIn is, too. So let's call it what it is. People are afraid for their job security and their relevance, whether that means something as drastic as a layoff, or does it mean that I sort of become a BA foreign agent or something, right? Like, any number of those fears are there.

What we're seeing is, especially for folks who are willing to just dive into the pool and take that chance and get out of their comfort [00:13:00] zone is, you know, that climbing up and down the ladder we talked about, the removal of the scaffolding that we talked about- Mm-hmm ... is really empowering. People get to do the thing that makes them like their job and why they got hired, which is to be a subject matter expert that knows how to build and utilize relationships internally, that has good judgment, and that has good taste.

It's tremendous. It levels the playing field in that way, and so that's what we're seeing it. We're making progress on features or feature sets that even a year ago I thought would take a year to make meaningful improvements on given capacity and discovery timelines and all of the things that slow products down.

Yeah. We're making progress on them in two-week sprints and finding where the other bottlenecks are in the process to get us to production, and that's the problem you wanna have. I oftentimes think about product very much in terms of, like Both classic design thinking, but also, like, manufacturing metaphors or analogies.

Like, I think about a manufacturing floor and raw materials moving through the [00:14:00] process to become a constructed thing. And as far back as the Industrial Revolution, you look at those types of processes to find the blockages, the slowdowns, the fail points. And a lot of what we're seeing is that with more people getting to use their great human judgment, their subject matter expertise to put forward great ideas, we're just finding where the other slowdown points are, which also makes our delivery all that better.

But gone are the days where the slowdown point is we have a great idea on, what, June 11th, and we can't even get a look at it until July 11th, and it's certainly not gonna go live until August 11th. Right. Maybe we'll talk about it for the next quarterly epic- That's right ... where it could make into a sprint halfway through two quarters out kinda thing.

I do love this kind of renaissance of people talking about it like a factory again, because for the longest time, feature factory was a pejorative. It was one of the worst things you could say about a company, but now people talk about it and drive this kind of software factory or solution factory. But I think it was always important to keep in [00:15:00] mind Ferraris are made in a factory, too.

But I think the metaphor holds a bit where you look at the difference was you were creating, you know, maybe in big industrial factories where you owned maybe one widget being installed 18,000 times, and now it's just, but the Ferrari factory has a very few people working on a car throughout the whole thing or a lot further through, and there's a lot more crossover, a lot more expertise, and that's a beautiful work of art produced in a highly skilled way.

It's still a factory. And that seems to be more the model of what this is moving to, where you're seeing these kind of ideas of people expanding and the idea, you know, this guy, Oji Udezue, who speaks with us a lot as we travel and kinda do these dinners- This concept of the hyper creator where maybe product folks are picking up development skills and design skills, or designers able to go further forward, or engineers are going further back, but a more diverse, complete skill set.

And in the end, that's probably going to be what's gonna persist, is gonna be being able to do more and do it better and bring that taste and execution. But that's far more [00:16:00] interesting in my mind than just owning, you know, I drill this widget in and I'm the best widget driller in the world. I totally agree.

Two fun thoughts o- on this, just top of my head. One is, as the lines between the roles blur and people are hyper creators, which I love that term, I see an empowerment of that group of people where they're all really focused on their outcome or the problem that they want to solve. Yeah. And if the product person decides at 11:00 PM that they get a wild hair and they wanna create the next version of the app overnight so that the engineering team can see it in the morning, just in the last few months, there's less of the engineering vibe going like, "Well, why did you do that?

That's my job." And similarly, if an engineer goes, "You know, I was talking to sales guys at this conference, and they thought we should tackle it more this way," everybody just jumps in and collaborates immediately, and that's been really empowering. The other thing that you made me think of with the factory I'm sure we've all either done or at least had ideas of a net new product, perhaps a consumer product or a physical [00:17:00] product.

And if you do that, one of the things that you do is you basically interview factories- Mm-hmm ... to make sure that you trust that they're gonna deliver something of quality on time and on budget for you. And I think if we talk again in six months on the podcast, we might be in a position where we're saying we are seeing product people, creators, hyper creators with a little more power in the job market as well, because they're able to go to a company and say, "I know how to work optimally now.

I wanna make sure I work at a place whose operations are set up so that we can really work like this." And I think that's a really powerful thing for product people to think about too. It can feel a little disempowering and a little scary right now. The LinkedIn headlines can be scary, but I think there's a real empowerment for product and creative people- Yeah

coming with knowing what good looks like in terms of how the delivery and the factory actually work. It seems like the fastest things that's kinda solved, sped up, one is code is far faster to produce now, and we can make the argument that review [00:18:00] processes and some of these are not quite sped up as much, but it's definitely faster produced code than ever before, and, and good code, especially as every model progresses.

The thing is, most of us sell to people still at some level, and they're not going to be able to intake faster based a lot of this. Like, as a consumer, I'm still only gonna need the features I need, or I'm only gonna be able to process whatever three features that my brain can process from X vendor every month or every week.

You know, when Anthropic went through that miracle spring they had where it's like seventy-eight features in sixty days or something, I think I used three of them. So what it comes down to is you're either gonna burn your audience out or, yeah, you can work on a lot of things, but you have to find a way to quality gate them and figure out which ones do your entire audience need to see or do your entire customer base need to see?

What's gonna be the most valuable? And in the end, that's still gonna come back to how do you have taste about what they need? How do you solve problems well, and how do you kind of figure that process out? And like you said, I think the people who are gonna be able to do [00:19:00] that are gonna be increasingly valuable.

You know, we might see a blip right now where people are figuring it out, but I don't know. The companies growing the most are not shedding people. They can't hire fast enough, and if I look at that as a model, I generally think anyone in any other case is probably like, "All right. You're catching up or you're figuring stuff out," which is fine.

I think you're right. It's almost a trope now, but countless people will say, myself included, "The more I use these tools and the more I use them with real confidence and with intention-" Yeah ... not just experimenting, but knowing where I wanna get from using the tool, the more work stacks up for me. Yeah I'm busier than I've ever been.

That's right. I mean, careful what you wish for. Even since we talked last time, I have created two massive strategy documents. I don't mean massive in page number, but huge volumes of work. Try to keep them at 10 or 12 pages with a solid one pager as the intro. That's my rule for myself. But you can pack a lot of content into 10 or 12 pages.

Yeah. And I do them exclusively on Friday afternoons and on Sundays, and these things [00:20:00] each would have probably taken me six months all, not all that long ago. Right. We had Laura Burkhauser on recently. She's the CEO over at Descript, incredibly fascinating human. She talked about one of the things that I think is one of my favorite.

I love a good four box. I don't know, maybe it speaks, I went to the business school in me, but I love a good four box. It's one of my favorite diagrams you can make, and she brought forth one of the best ones I've ever heard to apply to right now, which is basically categorize the stuff you love and don't love, and can AI really help it or not?

And basically, the idea was we're not looking to have people just be sitting there and just prompting and be basically the human harness for AI. We're not looking to shed good people either, because AI can amplify greatness. It can also amplify people who are less effective, let's say, can be less effective at a lot more things.

But this concept of like, look at and understand where you can take the things that would be great to do with AI and you don't love doing, and that's the stuff that you can really remove from a lot of what you're doing on [00:21:00] the day to day, and they get done still. And then look at the stuff that you love doing and needs a human to do, and that's the special stuff.

And I think that's the areas we're spending more time in right now, because all this other work that we're doing faster enables that. At least that's what I've seen on my end, is I'm busier than ever, but I love it. It's really interesting work that I get to do. Yeah. If we select right, you can just... What I'm not doing is writing seventy-five thousand tweets a day that are just brainless rejiggerings of stuff on the internet.

Yeah, I think that's true. There's that great book, Deep Work- Mm-hmm ... by an author, his first name is Cal, I forget his last name. But if done right, a lot of this allows us to focus on the deep work. Exactly what you said, also tethers very much to the data site flavor of all of this. You know, you think about the vast majority of our users, a lot of their work is incredibly high stakes.

It is high pressure, it is serious business. I have actually lived through it. I was on a divestiture team for six months. Mm-hmm. And every single day had moments that felt like make or break. It's really stressful. Everything's due at eleven [00:22:00] fifty-nine PM, and you always feel like you could be the fail point.

It may not be true, but it really does feel like that. But what's weird is the tasks are simple and mundane and repetitive. Mm-hmm. And these are precisely what you're talking about. Yeah. I think of those as first mile tasks. It's a great analog to why legal tech was one of the first to really hit their stride in this AI revolution.

A lot of the work is Time consuming, repetitive, and especially as your mind fatigues, it's where you can make mistakes, but it's not rocket science. But the attorney, their magic is their- Mm-hmm ... subject matter expertise, their deep training. And so to get that first mile of work done for an attorney or a junior banker on an M&A deal in a way where it might save you six or eight hours a day, it doesn't mean that then you go play golf.

It means that you use your human brain while it's sharp and not fatigued to do the really hard decision-making, the judgment calls, and to make sure that [00:23:00] the internal nuances of your organization or of that deal team are taken care of. And oftentimes, in my experience, that deep work happened late at night, and I was really tired, right?

Yeah. It's funny you brought this up because when you and I first met, and you were telling me about DataSite, and I remember thinking just how possibly are you fitting? 'Cause it seems very similar to me as, like, medical and things like that, where people would be really worried about all the very sensitive data that's in there for a lot of these companies.

But I remember two things. One, like I said, I went to business school, and so I had a lot of friends that went into junior banking roles where they were doing the scut work on paper processing and all that kind of stuff to clear those deals. There's use cases for how the AI revolution has played out because- Yeah

the way we think of ourselves is the virtual data room. Yeah. So not too long ago, a data room was actually a room. So if two companies were engaged in a merger or a acquisition, they would oftentimes use a conference room in one of their offices or even rent a room at a hotel, so it'd be sort of a neutral site, [00:24:00] like a conference room, and they would fill the room with bankers boxes full of actual documents, and they would allow people in and out.

There would be security. You'd have to, like, leave your cellphone, your laptop in a locker and go in, and attorneys and accountants and deal teams would go in and actually review documents for days at a time. And so the reason they call it a virtual deal room is because in that world, people understood, oh, this is a virtual version, finally.

Of this physical room. And so you can imagine a human mind or half a dozen human minds having to process 15,000 documents, hundreds of thousands of pages of documents. Right. It's mostly theater. Nobody's processing all of that, right? You're mostly scanning and looking for red flags and just doing your best with, you guessed it, judgment to make sure that things seem like they're checking out.

But these processes take multiple months, and they are high stakes, and they can be risky. So that's sort of the nature of the virtual data room. That's why DataSight exists, is to automate that. But then the [00:25:00] next step where the things we're talking about come together, well, the tasks that people have to do as part of a deal team are, like we said, repetitive.

Mm-hmm. They're not super highly skilled, but they require a lot of training and good judgment, but they just take a really long time. And so exactly to your point on the legal side, if we can massively expedite the understanding and synthesis of lots of information so that the subject matter experts can make their decisions and use their judgment and speed up that process, all the better.

But you're not removing the human as the source of, like, judgment. Nobody with their finger on the trigger of a billion-dollar deal is gonna say, "I don't know. Let the computer decide." Claude told me it was good, so we're just gonna YOLO it and go for it, right? That's right. That's right. That's right. And then one of the things we talked about previously, and I'm sure listeners and so forth think about, is, well, where's the line if you're using these- Yeah

tools to understand all this stuff? And just like with legal tech and medical tech, that's where permissioning comes into play. [00:26:00] Mm-hmm. That's where the idea that if I'm adminis- the administrator on the deal room and I'm deciding which groups can see which content, perhaps in what format or for how long, if I can redact certain things before anybody sees it, and I get to control the flow of permissions- I can use all of that powerful AI technology for purposes that it's very well-suited in a way that does not inject risk to me as the one who needs to make sure that this information and the data is sacred.

And so this sort of governance architecture that AI allows us through permissioning in and out of data rooms and permissioning of particular content becomes this real sweet spot where you don't disintermediate people and their expertise and their judgment, but you drastically accelerate the kind of menial work that it takes to get the room ready.

I guess from a product standpoint, and I realize this is kind of an active product development thing, so maybe there's only so deep you can go, but how do you determine... [00:27:00] You know, there are some things that it just needs to be dictionary truth. It needs to be accurate, and probably the answer ends up being, for the most part, a human really just needs to read something and see it or have heavy citation and know where to look.

But there's this other general piece of this which is understanding a huge amount of content and just kind of generally understanding, like, is there something of risk here? Do I need to dive deeper? How do you actually break those two down where at some point is it actually product experience differences?

You kind of talk about permissions being different and stuff, but what does that look like to the user and how are you ensuring I as a seller am not getting in trouble? 'Cause I'm like, "Oh, let's use this tool," and you provide this, and then all of a sudden it's like the tool you chose gave me bad insights kind of thing.

Yeah, it's a great question, and it's something that's changing and evolving all the time. Yeah. But you already nailed it. A lot of it is the kind of disclaimer language that we all see in, especially in the chat bots when we use these tools all the time. Mm-hmm. So you see two big paradigms. One is citation.

So when you're really deeply [00:28:00] interrogating content, and perhaps you're comparing what's called change of control clauses- Mm-hmm ... within an M&A deal, or you're looking at the entire footprint of a company's leases, perhaps they have physical locations all around the globe, and so your leases internationally become a really complicated fabric, right?

And you are just doing some compare and contrast between, say, two or 200 documents. In that kind of a workflow, citations are everything because you're not going out to the public internet. The tool isn't going and saying, "I don't really understand. Let me add semantically some additional context from the web."

Let me see what Google thinks. Yeah, right. You're very much in a walled garden, but what it's doing is saying, "Well, we found that the leases in Portugal cited here- Mm-hmm ... tend to have this, you know, common thread that you don't see in Spain. Looks like something you should probably check out." And even the choice of language there is important.

Rather than saying, "You should check this out," oftentimes what we seek to do is say, "Deals that successfully close [00:29:00] tend to remedy differences like these." Yeah. Correlation, not imperative. That's right. It's not, "The tool told me to," it's saying, "We've got a vast amount of data from 10 plus years of running a business like this that tells us that people that have kind of conflicting information in their leases, they do better if they remedy that before they share it out with potential buyers."

Right. And then secondly, in a chat, when you're involved in more of a natural language chat that's more free-flowing, but you're still within the walled garden of the data room, just like all of us experience every day now, you say, you know, there's always the disclaimer where it says something like This is generated by an AI tool.

They can make mistakes. Right. Check your work. A paradigm that seems to be really replicating itself, because we had a guy on the show called Robert Hankhouse, he works at a company called Invarys. Literally, if you watch the show, uh, Landman, he grew up in that. It's a company that helps big oil companies kind of parse data around where they should drill and they're giving insights for multi-million, if [00:30:00] not hundred-million-dollar investments.

And the key there is citing everything, showing truth and building trust that these insights are worth it. But on our end at LogRocket, it is, you know, a slightly less risky thing. I don't think anyone's wagering a hundred million dollars on a single website experience, but it watches experiences for you and tells you where in your digital experience people are running a problem or where the friction is.

It's been interesting showing this around and getting feedback from product folks and engineering and everything, that one of the biggest things is it doesn't just tell you something broadly or give you some blanket insight. It cites, "Here are the sessions to watch if you want backup of this to really see it in proof.

Here's the correlational data. Here's how to reproduce it." And it's all the context and understanding, and, like, that kind of same way that we've seen a lot of deals like this not go well because of this. You know, ones that remedy this do better. Yeah. It's this vast amount of proprietary data you've built before- Yeah

that you're able to form opinions off of later that that is driving a lot of what is [00:31:00] really the valuable insights and not just AI slop that we see in too many places. Exactly right. I mean, one of the natural questions for a customer, I always try to use these storytelling devices. Like, imagine you're talking about your job with a neighbor who's a really smart person, but your neighbor's a dentist or somebody who doesn't think about software and AI all day like we do.

And they might say, "Yeah, but I mean, couldn't you just go do that on ChatGPT?" And exactly what you're talking about becomes kind of the interesting conversation with them is like, you absolutely could, but what we're querying against is a vast pool of knowledge based on historical trends. We never reveal names or deal names or anything specific, no PI, anything like that.

Yeah. But we've got context that the public internet just simply wouldn't. Mm-hmm. Or at least that the general user wouldn't be able to sort of ascertain through prompting and skills on the public internet, and it becomes a very powerful story in that way. Yeah. No, it, it's interesting to [00:32:00] see how this evolves where it does seem to be accumulated information.

Taste has become a bigger and bigger thing, and just marrying the ability and expertise developed over years, or just not, just na- naturally inherent in someone, against what AI can do is giving these huge gains. So it's interesting to see how the world's developed. I'll be honest, I love stories, and AI, you know, like you said, not everyone's like this where they think about it all the time, but it is something that crosses my mind quite a bit.

I could go on and on, and I think my team would tell you that's true. You have some pretty innovative stuff over at DataSite that probably needs to keep going. On my end, I'm getting looks already that it's time to wrap this up, and it's been a blast. Time flew here, Eric. It was great to have you on. I appreciate you taking the time.

Like you said, we'll have to come on six months again and see how these things have developed even further. But thank you so much for coming on the show, for talking us through both how the team is using it as well as how you kinda look at it from a product standpoint. Yeah, it was a thrill. Thank you so much for having me.

Yeah, thank [00:33:00] you.