feedback-just-the-start-how-acting-on-it-increased-zipcar-nps-50-nishaat-vasi-launchpod-logrocket
===

[00:00:00] All right. Hey, Nishant. What's up, man? Good to have you on the show. Thanks for coming today. Hey, Geoff. How's it going? Thank you for having me. It's been a while since I've seen you, but really excited to do this. I mean, we talked about it before, but I'm a long time old school Zipcar fan, so really, really stoked to get into this.

Before we kinda jump in, I, I think we got a great story here about taking a company that has been a long time success story and redoubling down in this new age on how do you focus on the user and really member experience and taking it to an exponential new level. Before we do that, I mean, your career has spanned some of the most notable Boston companies I feel like over the past 20 years.

Can you give us the TLDR on who you are and how you got here? So I started at MathWorks, right? So believe it or not, my first job was taking calls for end customers and really helping them out designing their systems. Pure B2B space. Worked my way into product [00:01:00] management over there. Really, the focus was how do we ensure a control system is set up correctly?

I'm selling to the NASA's of the world, the Toyotas, the Fords of the world, to ensure that their design systems are set up accurately. So I go from a B2B product manager working in the platform space, transition into TripAdvisor was my next gig. Pure B2C, all about optimizing the e-commerce experience and, you know, content is king, but really how do you optimize that content?

So that was my stint for a couple of years at TripAdvisor, and then I'd say Zipcar is somewhere in between, right? Where I've been for the last eight plus years designing and defining what the product experience looks like. In my current role, I head product experience, which covers anything and everything to do with the end user experience, our internal tools, our operational platform, and it's really about how do we optimize within a fixed asset, not only for the end consumer, but absolutely for profit.

So it's been a nice journey, B2B, B2C, [00:02:00] and I'd say Zipcar's somewhere in between, but I've enjoyed every minute of it, so I've been lucky so far. Yeah. You talk about kinda B2B and B2C on Zipcar. If you ask anyone who doesn't work at Zipcar, almost everyone's experience with it has been more on the B2C side, but there are other pieces to it too.

And you had a little interjection there at Flexcar, which I think it was spun off from Zipcar, and you went off, and then kinda came back and stepped into the CPO role at Zipcar. How did that kind of more startup-y end of it influence how you're looking at it now over in your second tenure in Zipcar?

Yeah. Thanks, Jeff, for reminding me on my own story. I, I keep forgetting that one and a half year stint over there, I still count with Zipcar. But here's the thing, right? During COVID, we incubated a new product. Think of it as a monthly leasing option. It did really well, so the board decided let's invest in it and really take it out from Zipcar, and it got incubated into its own P&L, its own product line, and I went with that.

The goal over there was like being one of the founding [00:03:00] PMs, right? Really, how do you set up the platform, ensure the stuff behind the scenes work, the pricing works, you know, billing systems work. Thinking really zero to one, getting your data science and the analytics function built out. That was a really interesting experience.

Startup-y, very high growth, crazy schedules, but was definitely fun. And that lasted for a year and a half before this role opened up where I had the opportunity to really come back as a boomerang to Zipcar and, uh, run product Just when you thought you were out, they pulled you back in again, right? Yeah, it was a bit too late at that point.

Yeah, we got him back, so that was interesting. It's interesting because, and I'm kind of saying this assuming everyone knows what Zipcar is, but maybe for those who don't, Zipcar I feel like is really one of the early pioneers in what kind of became the, I don't know what's called the sharing economy.

Yeah. But this revolutionary idea that you didn't have to own a car, but you could still utilize a car, and it wasn't going to rent it at an agency or something. You would have a membership and the experience was just great. You know, book it, [00:04:00] walk up, you had a card or, or some way of getting it in depending on what, you know, year and decade it was, and you walk right in, but the car was clean, it had gas.

Now we're talking many, many years forward to when you stepped in as a CPO roller coming back from Flexcar. Let's set the context that you walked in, what drove you to say, "Hey, everyone, we need to look back and really drive this concept of member obsession again and, and really double down on this idea that we're going to infuse customer feedback into our DNA"?

I'll set context a bit. So Zipcar, when I joined initially before my CPO role, we were going through a massive platform overhaul. So this is, you know, you're changing the, the engine on a plane mid-flight because no one's gonna let go of revenue. We still need to meet our targets. And so a long time was spent, multiple years was spent on this overhaul.

I got back into Zipcar, right? One of the things which had hadn't happened at the end of this migration was we'd lost sight on the end [00:05:00] consumer. We were in so much of a, oh, we need to get this done right, you know, and it's taking too long, budgets are tight. And really the customer experience was treated as a second grade citizen.

Mm-hmm. And it's just no fault of anyone, it's just the pressures of the system. Right. And so one of the initial goals was how do we get us back to be a little more customer obsessed? Member obsession is part of our core pillar, so how do we really rally the troops back around that? Step one, we were still capturing NPS, so, you know, Net Promoter Score at the end of a trip in the old-fashioned way.

Believe it or not, it was an email form. We're not really leveraging the feedback which we were getting, and we weren't getting much in terms of feedback. Like, maybe on a good month, 3%, 5% of people would respond, and you know, we just got this brand new app set up, so this was the perfect time to say, "Hey, how do we move from post-ride to in-context feedback?"

And there's multiple goals to this. So one was are we [00:06:00] capture feedback in moment from a perspective of the customers with you? They're actually more inclined to give you, "Hey, this ain't working. You know, my check-in experience failed," so on and so forth. But at the same time, like it's really about feeding back that knowledge and the insights which we are deriving- Back into the different teams, and I'm not talking product teams and- Mm-hmm

UX teams only. I'm talking, like, even our operations teams. How do you get the actual change to happen? As you mentioned, Zipcar's all about car sharing, so Jeff may be using a car and he comes back without filling gas, right? It's gonna impact Nishat, who's the next person, and who's gonna be stuck first trying to go find a gas station in the start of their ride.

Not the best experience to have. How do we close that loop quicker or ensure cars are cleaned by the right person at the right time, and really get back to being a little more customer-centric, right? So again, at a high level, you know, how do we get in context feedback? Then what do we [00:07:00] do with that in context feedback?

Outside of just getting feedback and leveraging it, it was also about how do we develop that DNA back into our product and UX teams. So we initiated, um, like I- it was first actually a mandate, but every week we'd get together and watch sessions of the user journey, map out pain points. We then organize like voice of customers from a customer service perspective and force people in, again, mandate first, but then this just started growing by itself.

'Cause the true PMs, you know, product folk and even engineering, they love listening and solving that problem, and they can see it, they can feel it, and they hear the agony or see the frustration clicks happening. So it was really about how do we instrument that. Then that takes a life of its own. Oh, we should include marketing folks.

Oh, we should include, you know, operations folks in this. And yeah, you know, you solve this, I'll solve this. It's actually fascinating to see the engine starting to hum. It takes a little bit to kickstart it, [00:08:00] but it's been transformative. For example, we moved our NPS from like thirty-six to fifty-six at this point, and that just doesn't happen overnight.

It happens through constantly figuring out what your user wants, and we deliver value when they need it and really doubling down on that experience. I love how you kinda look at it from the qualitative, what are people doing, how are they running into problems, but also how do you have quantified behavioral patterns or quantified problems or issues and looking at that.

But also, you brought up they're much likely to give you the feedback if you are looking for it and capturing it and hitting them for it during the time they're in that moment. If you're twenty-four hours later, a customer service rep is giving me a call saying, "Hey, you gave it a five. Why? What happened?"

Yeah. You know? Even like my Uber now, when it pings me for a tip twelve hours later, I forget who the person was, which rider it was. But when it's right after, I'm like, "Oh, that guy was great. We had a great conversation," you know, this and that. You're much more likely. But I guess did you find there was an [00:09:00] element at all of correlation where I feel like coming into a great experience, you come in, right, the car's clean, it smells nice.

You got full tank of gas, whatever. It's easy to key in, no problems. It's almost a network effect. I'm probably misusing the word, but you come in and you're more likely to leave the car in great shape because you came in and it was so pristine, you don't wanna mess it up. You come in in like a gas half full and, you know, maybe someone left a little clutter around or someone left a couple wrappers on the floor or whatever it might be and you're like, "Well, you know, what's another wrapper gonna do?"

You're absolutely right, right? It's just human psyche. It's like you go into a rental and if it is maintained well, right, you know, when you're renting out You feel like that's the way you're going to leave it versus if it was crap, you know, you're really gonna leave it as crap. And that comes back to, like, how you manage a community of users with a shared service.

Mm-hmm. And that mindset of how do you raise the bar is truly accurate, right? And there's two things. Actually, you mentioned the Uber stuff. Look, it's a given, [00:10:00] right? Everyone says, "Oh my God, you need in-context feedback." There's a bunch of tools out there, but the key is, I think that's the easy part.

Instrumenting your app and instrumenting the feedback loops, great, right? But it is what you do with it which really matters, right? And how do you operationalize that, and how do you bake that back into... You, you talked about not only the qualitative but quantitative, right? That moving a specific, I call it NPS one click down, right?

There's car condition related items. We made it a team's goal as a metric and an OKR for them to move it in a given quarter. It's part of your scorecards. It's part of if you want a bonus or not. So it's really about how do you set the right goals for the right set of teams, right? One may care about something else.

One may care about damages. One may care about maintenance related issues, and to leverage, you know, push them on that side. And then really, right, by also giving them [00:11:00] a platform. I think that's key on like, "Hey, see what win looks like? That's awesome. I'm gonna pull you to my all hands company meeting and talk about the effort you did in your market with the product engineering folks in the room."

And it's not just product, right? Right, and this is now a company effort to get to a higher bar. We made it a North Star. For a couple of years till we fixed it. You know what I love about this story, Nishat, is, like I've said over and over and over again across dinners that we host and I think on the show, software is not the goal, right?

Software has been for a long time, you know, a couple decades now, often been the best way to solve a certain set of problems, but we can get really tied into forgetting that, that we're not making software. Yes, we are as a byproduct in many cases, but we're solving a problem. And this is a great example because Zipcar at heart, it's a physical product.

Like this could have existed before software. It would've been a little harder to probably administer a couple things if you didn't have, like, you [00:12:00] know, the internet and a couple shared services like that. But it'd be doable because the core thing was people need to get from A to B, and they need a way to do it, and they want a nice experience.

And so many of the things you're touching on go beyond, "Oh, we did event capture. We did in-app monitor. We pinged them live for feedback." It's at heart comes down to you wanted to increase NPS. And how do you do that? You get stakeholders in the physical world, and yes, because software is part of it, the software world too, and you put them together.

But then, and I wanna dig in this piece A huge thing there is, just because it's a big company, you're not probably limited on if it's gonna improve the experience, it's probably gonna improve revenue. You can fix a lot of things. It's gonna come down to how do you fix the things that are really gonna move the needle?

You went from 36 to 56 on NPS, that's more than a 50% gain in a not super long period of time. You're not doing that by randomly whack-a-moleing problems. So when you looked at all the feedback you were getting in all the areas of interaction points and data you were getting, how do you pick across all those teams, [00:13:00] like, this is the two things this team needs to work on that's really gonna move the needle.

The rest are nice to have, but if we can do these two things, that's gonna be, you know, 70% Or did you find it was just a, a kind of wide, wide spray of problems, and it was just let's fix a lot of things? No, that's an excellent point, right? Like, it comes back to you gotta prioritize the, the hell out of everything.

Yeah. And so we picked the top three areas, right? Our top three areas ended up being cleanliness, damage, and certain maintenance related issues which would get clubbed with damage. And so what we found, there's actually more than just a pure member benefit, right? So the key is, okay, great, we could clean more often.

That's a costly proposition, right? But how do you nudge people to clean up after themselves, right? That's one, the community effect. Mm. The second one, right, on the damages front, that's a costly proposition. I have an asset. It's depreciating in value. If I ding and scratch it up, it depreciates much faster.

[00:14:00] Guess what? I make less gain on sale when I sell this in a physical unit. And so then figuring out the P&L connection, this is saying customer problem is let's have a better car available. I feel good about it. Our business side problem, right, our P&L problem is I'm losing money if this doesn't get fixed.

So where possible, connecting it back to for every 100 such events, here is our hypothesis which, again, all backed by data. We model this out. We look at different conditions. I've put a bunch of thresholds around it, but here's how many million, I'm talking about millions we're going to lose. And this is why, in fact, it's not only damages, it actually leads to how people treat your cars, which you said, which could be accidents, right?

Mm-hmm. It's not about dings and scratches. This is about people actually driving rashly. And so we started changing the behavior of people with direct reduction. You know, it's double-digit [00:15:00] percentage points is the way I would put the savings. There's actual cost savings for the business. So having that connection was really the key.

And then again, as I said, right, like really saying it's not everyone's problem to solve cleanliness. There's only four teams involved. One's operations, one product, one's, you know, UX and ensuring we're solving it in the right way. You know, operations folks are actually piloting a bunch of smoking and smoke reducing products and detecting devices, which of course, you know, comes through as an event and stuff like that.

So it's really about figuring that out, then focusing on specific teams and empowering them. Literally, the key is empowering them to be like, "Here's the goal, now go get it," right? They're way smarter than I am. Better ideas than I can ever come up with I like to think of it as maybe not smarter, just people have seen different things and have had different experiences and can pull from that, and we're gonna work best on these things when you can draw from a lot of smart people who have a lot of different [00:16:00] experiences, and you can fully synthesize the problem because at some level, you need the context to come in and say, "We need to focus on this.

We need to prioritize it. We need to pick the big things." But in very few worlds are you then going to have that org-wide span of context as well as, and now here's, you know, people are smoking, and we need to find a way to reduce that, and here's the three ways we can do that. So this is where, right, you said, like, let's figure out big priorities.

Let's ladder it down. I'm curious though 'cause, like, so much of this seems like it's probably physical problems with physical solutions or, or maybe not. Maybe it's physical problems with digital solutions. And then I would wager there are likely some small, cheap things that will have big effect and some big, expensive things that will have, like, little effect.

How did you kind of understand the two sides of that? Like, it's gonna be a big project, little project, big effect, little effect, and what is the real impact? But also, can you walk us through an example or two of things that were surprisingly high leverage? I'd say a few ways, right? So the way I look at this is, again, I come back to my remit.

My remit is all technology investments and [00:17:00] portfolio of it, right? Most of our-- So the way the rental works, everything's accessible through your app, right? And I need you in my mobile app Right? We've got actually got rid of those cards, Jeff Uh, yeah. I think this is a long time ago when I've had a car for a little while.

I had to commute a long way every day for, for a while there, so. No, that's fine. I still have mine from the old days, and I think it still works, but everything's in the app, right? So that's actually good from a control perspective. I get to control the experience. So what ends up happening is it's never, oh, I can do this via software only, or I can do this via field person, a fleet person figuring out their own stuff.

It's likely a software and operations problem to solve, right? And I say operations broadly. It could be marketing, it could be fleet management, so on and so forth. But really when we look at it, we looked at it from a investment perspective. There's so much you can spend on software and so much is gonna be needed to support whatever solution you have.

So what I had to create [00:18:00] was a standardized, I call it a standardized template, right? Your pure investment portfolio type of template where it says, "Here's a hypothesis. We're gonna move the P&L by this much. We're gonna move the member experience by this much." Mm-hmm. And we hypothesize this is the end value.

So step one, for the larger projects, think of a couple of million-dollar investments, right? We really have to evaluate what we're really gonna, what they're gonna do. But the second is really to come up with those, the teams have the freedom to try a few pilots. Mm-hmm. So something which costs like 50K, 100K, go try it.

You don't need approval from a leadership team member to go try something quick, but the goal is have that quick pilot, give it a couple of months, then come create a business case to get the actual investment- Mm-hmm ... so that we don't jerry rig everything together because that's also part of it. Like, oh, let's just do this and I will look at three tools to do my job now instead of my one tool, [00:19:00] right?

That does not scale over time. So the key for me was how do you figure out pilots but create a platform which scales and is doable over time? So back to your question, right? What's a good example? I think a good example would be we had this damages thing, and one of our ideas was, let's change the way someone checks in and out of a car.

So for example, people could just hit unlock, great. I force you to have photos of the before and after condition of your car. I'm specifically interested in the exterior to check out the dings and damages, right? And so just the fact of making someone do it, it actually from a community and behavioral standpoint, it changes the game, believe it or not.

Whether I do anything with that data or not makes no difference. Big value is in the way people perceive the product at that point on. But okay, great, it's a technology solution. We saw some great wins. It took months to roll out and figure out and get right, [00:20:00] and we can still optimize it. But the second, so we got to, let's say from A we improved ourselves to B.

We're still not at our goal, which we need to get to. So then we started saying, "Okay, how do we pilot?" There are devices in the market, right? Off the shelf devices that you install in the car. They cost you a subscription cost per unit, right? Per unit per month. But these have a gyroscope, right? These have impact detection abilities beyond what I call my brains in my cars.

Each and every car is connected and has a brain unit, which is proprietary to Zipcar. It's doing a better job than that. Mm-hmm. And it's okay, right? We don't have to build everything in-house. So we tested it in a couple of markets. Again, zero CapEx investment, zero software investment. Just get this out, test it for three months.

Tried it in two markets and the team came back, joint team, right? This is not your product person. This is my member services person who's driving this [00:21:00] 'cause they also have to take care of the damage related P&L, right? And so we come back with this device. We tried a couple. One worked better than the other Doesn't solve 100%, but we ended up using a device which gave us both damage detection and smoke detection.

Mm. So smoking has been a big problem for us, and so we really don't want you to do it. Go do it somewhere else. Just don't do it in the- Yeah. And so we got like a two-in-one solve, which we had not expected, and that was worth a few million dollars annually from a P&L perspective. So it's funny how you let the teams try stuff.

It's not funny. You know this. This works, right? This is, this is the core of, okay, unintended, we got, like, two birds with one stone, and it actually, the smoking piece works a lot better than the damages piece. So end to end, we started with damages, but we kinda solved on both sides of the coin, right? Again, it comes back to data, and we track, okay, where does it make [00:22:00] sense and where it doesn't.

We're pretty ruthless about everything backed up with data, and then three months post-launch, again, measure. How is this doing? Is this worth now just putting it in BAU, business as usual, and let it run, or is it worth reiterating and going back to the drawing board, right? Yeah. That's a success story.

There's not all which are successes, right? But you gotta let the team try to it. I love the in-car device piece because I feel like something like smoking, people are very aware of social or something akin to social pressure. Mm-hmm. And so even a light ping of like, "Nah, you're not supposed to do that.

Remember, like, no smoking," and you know, they realize they got caught, it seems like something that a small ping is probably gonna make people stop and not do it for the most part. You know, some people are just gonna do their thing no matter what. But to me at least, I feel like it wouldn't feel oppressive or it wouldn't feel negative.

Be like, "Oh, yeah, I can't. It makes sense. I can't, I can't. Okay. That's fine." And so you're like, you're having this really positive effect on, I'm sure, like, cleaning costs are down 'cause cleaning smoke out of cars is terrible. [00:23:00] Whoever the next driver is doesn't have to smell it, so you're gonna have a better experience.

And is it even reducing the original person's NPS at all or, you know, their experience? Probably not or not measurably almost, but like you said, it was also very, very small lift to do. And I think that's such an important piece that we forget, and this happens in marketing all the time too, right? Like, my background in marketing, so many people are prone to say, "We have a problem.

Let's get a tool. Let's get another thing. Let's bring in a different application or different software solution," and suddenly you have 20 things you're checking, all measure different things. But you didn't stop to think about what are the small things we can do that are gonna have bigger effect or what are the small fulcrums we can utilize to get there, and that's what I love about working here.

You know, we talked a lot about Zipcar, but at LogRock, that's one of the biggest things, is how do you kind of understand at least the digital side of those experiences and how do you surface What are the problems affecting users? How do you kinda bring that forward real time? What's really impactful?

But also, how do you assess impact and how wide it is, and what's the potential gains of fixing it, and then how do you automatically fix it? And now, how do you bring in that user [00:24:00] feedback too, and, and take that into account so, you know, kind of automatically shifting points, you know, to something like this is exactly where we sit.

So I love hearing this story because that's kind of the problem we've been focusing on a long time is how do you pick the right things to focus on really quickly? And to hear how you guys have done it is a great way for us to look at it and continue to evolve. I will say one thing, right? This is a book I'm a fan of, Amy Edmondson's The Right Kind of Wrong.

You know, she's famous Harvard professor, right? Talks about a lot of different strategies. But something which really I took away was, like, you know, how do you... Failure has happened- Mm-hmm ... right, throughout the system. But how do you recognize and encourage the right kinds of fail fast approaches? So taking my TripAdvisor days of test everything, not possible in Zipcar to test everything, right?

But, and in some contexts it's not for the different companies as well. But, like, how do you empower, like, how do you fail fast, right? And figuring out what is okay and what's just outright stupid, which could have been avoided. I have to check this book out. You know, you bring up, like, [00:25:00] let teams fail a little bit here and there.

You know, you gotta make room for the experimentation. If you have the right focus on how do you discover the right bl- problems, likelihood is, A, no one ever bats 1000. In a lot of these testing things too is if the cost of the experiment, you know, the gyroscope thing, what did I call it? Like, you said, "No CapEx, no OpEx, really."

What happened if it didn't work? You were exactly where you were, and it took one person a little bit of time. If it did work, which it did, huge gains. So you gotta make room for helping people understand we're gonna fail sometimes. Not we might. We are going to fail sometimes. It's okay. Now let's get on with winning.

Right. No, I, I agree. I love talking about the experimentation because I think it's really, really important. People need to understand this. There is one other piece I wanna hit because you've talked about this publicly, and I mean, let's be honest, we've all heard of those two little letters, AI, right now, and it's kinda all people can talk about.

But I think you have a really good take on looking at it and, like, I've heard the term Theater of AI, especially bandied [00:26:00] about, you know, maybe some content on LinkedIn. You know, I think you've taken a really pragmatic look at it's not this magic wand. It's a tool, which I think the same, you know, we all looked at mobile back in the day and cloud, and they're not magic either, we've learned.

Seems like you had a good pragmatic view on AI strategy, AI theater, and why most AI strategy is just AI theater right now. What is your kind of view on the way forward companies are gonna take advantage of this stuff? Yeah. The hoopla is real, like every day a little bit- ... yeah, a whole new thing. Look, I think it's a really useful, you know, set of tools to have in your tool set, right?

The way I think about AI maybe is at its core there are two pieces of it. One is how do I get AI into my product, right? How do I get my-- the gen AI to help from a product experience perspective? And then the second is really how do I make myself and my teams more efficient in their day to day? And I think at the core of this is you gotta have the data first.

You push it in, push it out. And so really it's about, you know, ensuring [00:27:00] we have the right data set up. We have access in our models or third party plug-ins which we use have access to those right data sets, right? Mm-hmm. So that's key. And of course, there's governance around it and not leveraging the AI, and I think people underestimate how much effort that takes to really- Yeah

create that walled garden, but that is crucial to get right. One of the things I would say, you know, on the optimizing, I think it's as a starting point, if you've not done it, but I'm assuming everyone in this market's been doing it. But internal operations, absolutely there is efficiencies to be had. I think one of the keys is connecting it back to an operating goal.

So we talked about cleaning, we talked about fuel cards, for example. I'll give you an example where we had in a certain market, right, a new fleet manager comes in. This is the person accountable to moving the cars, ensuring that cars are shipshape They actually went ahead and removed these, what we call our fuel cards.

Think of it as a credit card tied to that vehicle in the [00:28:00] visor of every car. They're like, "Oh, why, why do we need these?" So they removed it for the market, right? And what we do is we monitor feedback from the end users, look at the App Store, pump it all into one place, and this is where AI comes in, right?

So it is almost like it found customer issues, extracted that trend- Mm-hmm ... figured out, like a lot of tools will do that extraction. But the key is the generative part is, hey, what is it week on week which really changed in all these unstructured? I don't have something which says fuel card missing, by the way.

PS or customer cap- I, I assume that my fuel cards are gonna be there in the car, right? It's not much of an issue otherwise. But it's really figuring that problem out from unstructured to structured data. Then we created Slack connections and connectors to fire off a bunch of alerts with specific people in those markets.

Hey, look at XYZ type of thing [00:29:00] which is happening That's where your gen AI really is distilling and doing the execution bit, right? We've got governance around not directly letting it go back into the app, right? Saying, "Hey, I heard you, Jeff. I know about this issue." That, that is still coming. One day. I mean, Anthropic just released, what was it?

Fable. Yes. Fable 5 yesterday. So I mean, who knows? Maybe tomorrow, but not today. You know, like, so this is, uh, really key, right? It would have taken me or us at least three to four days to figure out what this problem was. Yeah. This got solved, structured, I mean, end to end got solved in under 24 hours. I'm like, okay, that was worth the cost of the tool itself.

Well, it, it's funny because you bring that up, right? Nothing was looking for is the fuel card there, but we built a couple internal things ourselves in this whole, like, company brain at this point, which is fantastic, and it can accurately answer any question about anything from sales to what's happening in the feature side and feature [00:30:00] usage.

Basically anything that's not, like, HR data 'cause it backed up PII- Yeah ... and walled gardens. But we were able to build it out. You know, we basically had someone start out in some, uh, not really but, like, back alley prototype basically. It was one growth person who just had a hunch about salespeople wanting to ask questions about their sales opportunities, and so he put this together and launched it and gave it to one or two reps.

And all of a sudden, the next day, the API that he had put on it was just getting like ping, ping, ping, ping, ping, ping 'cause it spread across the entire sales customer success, SDR, marketing organization. Like, product people were hitting it, and we had to actually build it out into a real hardened application, you know, do a little bit more and build more robust endpoints and bring in more data layers.

It was good though because literally it took us, I think... I won't say 48 hours, but maybe a day plus that to build the entire thing, but it was because we had spent a year or two before that labeling all of our data structures and explaining how they work together and having it all documented how everything relates to each other, and that's allowed us to move forward [00:31:00] really fast.

And whenever I've thrown kinda raw data at LMS like that, the problem's never that the data's messy. The problem is you don't explain what the things mean and how they relate to each other. And as soon as you do that, magic can happen. So this sounds like there, same thing, right? You know the outcomes, you know what the data was, and now it knows how to use it to find the right things.

Oh, absolutely. And- Another example a few weeks ago, right? Like every software company, we had a minor production issue, and so this is like customers can't get into a car. That's serious, serious. I call this P zero, all hands on deck, right? Yeah. Doesn't happen all that often, but it does from time to time.

And so this is where we've got these things plugged into our code base at this point, where of course you have to prompt it correctly and got... We've literally got a standardized prompt, "Help us go figure out this issue." I kid you not, it did it in three minutes. What might have taken an hour to do, it was insane to see the fix come through, and the fix is out within half an hour, which may have [00:32:00] taken on a normal basis, senior engineers at least two and a half hours to figure out.

Yeah. And th- that half extra matters. Like that's an hour and a half more where users are able to get in and just have a great experience and not even know a thing was wrong. So awesome. Well, Nishant, I could go on with this for all day 'cause we are hitting on my favorite things, AI, user experience, driving cars.

But, uh, we'll have to have you on again 'cause this is great and, you know, I wanna keep kinda understanding how this evolves over time. Great seeing you again, man. I'm glad this worked out, and next time maybe we can-- we'll film a follow on, but we'll do it from a Zipcar while we drive around or something.

Well, thank you so much for your time. This was fun. I appreciate it.