Episode 99
· 39:13
Folks are skeptical of AI and then they hear great feedback from other end users or the internal team members. Oh, my God. Like, you converted not just people within my company, but also me a little bit.
Jack Bridger:Hi, everyone. You're listening to scaling DevTools. I'm joined today by Nick Gomez, who is the cofounder and CEO of InKeep, which is a AI support tool focused on developer tools. Is that a good
Nick Gomez:That's a perfect intro. Yeah. Okay. Great.
Jack Bridger:So I know you're work I was looking through, like, your testimonials, and I think there's 3 of our guests so far are using. So A lot of people.
Nick Gomez:Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou, Lou,
Jack Bridger:Lou, Lou, Lou, and, And that was just from the bubbles, so I'm sure you're working with even more of them.
Nick Gomez:Yeah. Yeah. Yeah. Developer tools, you know, a lot of folks in the space, and, they're fun to chat with and then good customer see.
Jack Bridger:Yeah. That's, that's awesome. I don't know if you wanna tell us a bit about, like, what you're doing and how you came to, like, work on support for DevTools.
Nick Gomez:Yeah. For sure. So we do AI support, for developer products. For developer products, support looks a little bit different than traditional b two b SaaS, for a few different reasons. 1, the developers kind of live in documentation.
Nick Gomez:So even, like, the touch point is a little bit different. Right? So in in traditional support, people the way in which people interact with support is through like a help center or, like an in app form that type of thing versus for developers it's often in the documentation, right? And there's other differences. So for example, a lot of support happens in community forums because a lot of developer products are kind of PLG bottoms up, type companies.
Nick Gomez:And so it just takes a lot of different shapes and forms. So we were thinking about like what yeah what does support look like for developer products, right? Like how do you scale a dev tool? Technical support is very time consuming. Right?
Nick Gomez:So anytime you're looking, trying to help a developer troubleshoot their issue, you kind of have to account for, like, what tech stack are they using? You might need logs from them. So it often takes, like, very technically minded, support people or product and engineering time to actually be able to, help developers. So we're thinking about, like, how do you help people actually scale that up as they grow as a company? And that's really yeah.
Nick Gomez:That's where we started and kinda what what we focus on.
Jack Bridger:Yeah. And and, actually, I feel like that's actually gonna be a good theme for the episode, which is, like, how do you actually, manage support when things come in from different places? Like, on the project I was been working on recently, all of our support requests for, like, the actual, like, SaaS would come into, like, the GitHub repo, and people would create issues and stuff like that, which is, like, completely unexpected for, like, what I was thinking it would would happen. Right. Right.
Jack Bridger:And then I guess you also get, like, people emailing you. Sometimes people would, like, track it down on, like, Twitter and stuff and, like Right. All of these things take, like, quite a long time to resolve, and they're kind of, like, often, like, interrelated with, like, features and things like that. And so well, like, have you seen in, like, if someone's building, you know, a DevTool, like, they're probably already doing support, but are there things that they should be doing, like, to kinda make that process better?
Nick Gomez:Yeah. I think kinda making it clear when folks should, reach to the different channels. Right? So, eliminating, you know, what should be a GitHub issue versus what should you go to the Discord, community to go get answered. Right?
Nick Gomez:That type of stuff is is very important, just to wrap people kinda to the right place to ask questions, and also to make it more manageable for yourself. Right? So you don't get duplicative questions across separate kind of all these different channels. Right? So that that can be one thing, but also just, that that's part of the problem.
Nick Gomez:I think it's support tooling is very broken for developer products because it's it doesn't really account for all these different channels where customer communication or customer engagements happen. Yeah. So that that's also kind of part of,
Jack Bridger:what we try to help companies with. So so trying to, like, signpost, like, okay, this help them put it in the right place, first of all.
Nick Gomez:A little bit. So right now, what we focus on in terms of helping is comp helping reach users wherever they're interacting with your company, and answer their questions, in any of those touch points. So for example, we have a search bar that you can embed into your your documentation, but we also have, like, Slack and Discord bots that you can add to your community. Mhmm. So part of it right now is just like a focus of, hey.
Nick Gomez:Wherever your developers are, wherever your users are, you can add a self serve chatbot, that actually helps those users get answers to their questions. So that's one part of it. And then the other part of it is that we ingest all of these different places where knowledge about your product exists. Right? So we take your GitHub issues, we take your Discord threads, we take your documentation, and we index all of it, and that becomes part of what the InKeep AI support bot, can answer with.
Nick Gomez:So that's part of basically being able to actually tap into all of this different content and all of this different knowledge that exists across all of these different channels. So we we help companies with that. Right? And so the traditional thing there is, like, someone's asking something on the docs or looking for something in the docs, but it was answered in Discord and Google can't get that. It's buried.
Nick Gomez:Discord search is not that great. Yeah. And yeah. So someone might not even wanna join a Discord. Right?
Nick Gomez:Yeah. So it's just that knowledge exists, but the developer can't find it. So they might end up creating a support ticket and adding noise to to your team. So what we help do companies do is kinda unify all these different places where where knowledge exists into, one experience so that developers can just get answers to what they're looking for.
Jack Bridger:Yeah. And I saw that you also do this, like, the hallucinations is, like, a big part of the thing that you talk about on the website.
Nick Gomez:Yeah. For sure. So with AI, you know, it's a little bit hype y, and there's a lot of folks who are very hesitant about how, you know, if and how to incorporate AI into their especially customer facing scenarios. Right? The the primary reason there is, like, if the bot, or the AI assistant, like, hallucinates or makes something up or says something that is misleading, that might actually create more confusion for like an end user.
Nick Gomez:Mhmm. And I think in the the space, it was it's very easy to create like a 60, 70% solution Yeah. That can like pseudo answer things about your product or your documentation. And so I think companies either kind of try to quickly spin up something themselves, or they got bombarded with a bunch of requests, from from startups who were trying to create something like a solution in this space because there is so much opportunity. But a lot of the the solutions fell short or fall short because the quality wasn't quite there.
Nick Gomez:So a lot we've met with a ton of teams who have explored many solutions and just could not come to a level of confidence with any of them because of that exact issue. And so that tends to be one place where we differentiate. We spend a lot of time making sure that the quality is as good as possible, like constantly experimenting with the latest approaches and methods and models, having really robust evaluation frameworks so that we can constantly test that we're not any changes that we make don't create any any hallucinations. And also honestly part of it is all of our customers have a really high bar for quality. Yeah.
Nick Gomez:So as soon as they complain about anything we go and fix it and that generalizes and helps everyone. Yeah. So by working with a lot of companies that really care about end user experience, we've just continued to kind of raise our own bar over time. And at this point, you know, companies with with really high bars, like Pineco and Anthropic, Postman, you know, they all really care about end user experience and quality, and they've all come to trust us, because we care so
Jack Bridger:much about quality. That's cool. And what percent of, like, questions end up, like, getting resolved, like, through without, like, the human being involved?
Nick Gomez:It depends on the, like, the actual knowledge that exists about your product. Okay. Right? Yeah. So we have like a we give people in our dashboard, a percentage of questions that were answerable from the documentation.
Nick Gomez:So teams can actually track this and use that as a, as a KPI if you will.
Jack Bridger:Help improve the docs.
Nick Gomez:Exactly. Exactly. So on the upper side, it's usually, like we've seen up to 90% coverage. That's for typically more mature companies that have a lot of GitHub issues, Discord, content. But even earlier state stage, it can be anywhere between, like, 30 40% as well.
Nick Gomez:Yeah. It's so it's it's I mean, you know, you can imagine it a good portion of that might have ended up in your support teams or in your engineering team's hands Yeah. If it hadn't been caught by by AI. So it's like a huge time saver but it also makes the iteration loop and the experience for developers so much better because they can get those answers as soon as they have them. Yeah.
Jack Bridger:They have to wait for a ticket to be
Nick Gomez:They'll have to reach out to a human. You know, they might be worried about getting being sold to. They're just like, I have a scenario. I wanna get it implemented. How can how can I do that with your product?
Nick Gomez:And if so, Yeah. That's what a developer wants to know when they're trying to evaluate a new product. Yeah. And being able to kinda close that loop and and give them direct answers is kinda like the best experience.
Jack Bridger:Yeah. And do you find, like, I would imagine one of my hesitations would be that as a developer, I am often very, like, skeptical as soon as I see, oh, no. It's like an AI chatbot. Like, it's gonna be so bad. Is do you think that developers are actually, like, trusting enough to, like, be like, let me try the chatbot?
Nick Gomez:Right. So we we also hear this a fair bit where especially, founders or or engineering leaders, they they're a little bit hesitant because they're like, okay, I come from a background where I I read all the documentation. I try to really understand the product, and a lot of folks still operate that way. But what we found is with tools like Cursor or Versus Code and GitHub Copilot, ChatJPT, developers they're actually like probably one of the primary users of all these AI tools. And I think they're becoming like it's becoming an expectation, that they can move very quickly in development.
Nick Gomez:Right? So that's been one of the huge increases in productivity for folks is for developers to be able to just use these tools to generate code, modify code, and and create new features, etcetera, very quickly. And that's all through these conversational LLM experiences. And so it it kind of makes total sense for developers if in their IDE, when they're building, they get this experience when they're trying any product or trying to learn about a product, they would want that too. Yeah.
Nick Gomez:And so while there are definitely people who are still very hesitant about, like, interacting with an AI chatbot, and that's in part because a lot of them are not that great. What we find is like when someone asks a question it generates a good answer, they trust it and they're like oh why would I ever read the docs again? So we've actually heard from founders, I've heard this story a few times actually, where like they were skeptical about AI themselves and so they hadn't really considered any AI solution. But then one of their users was like, hey, you don't have a AI chatbot. This competitor does.
Nick Gomez:Could you get one? I don't I don't want to read through all your docs, or I've had a really good experience trying other things that have this. So it's almost becoming like an expectation that you have a solution for because developers are kinda getting used to that.
Jack Bridger:Yeah. Do you think there's gonna be almost like as, like, you both like, as you get more context and the chatbot gets better, and then, like, Cursor is probably gonna get, like, more like, they might start, like, trying to get, like, the Discord chats and stuff of, like, you know, of of these dev tools. Like, you might wanna, like do you think there's ever gonna be, like, a kind of, like, almost, like, deep I'm a person who's not yeah. It's like what you're describing is, like, actually seems, like, valuable beyond support, but also just, like, how I develop, you know,
Nick Gomez:for sure.
Jack Bridger:Cursor, but with more context about the dev tools.
Nick Gomez:Right. Right. Right. So curse actually does have, like, a a way to, index arbitrary websites. It's it's it's like a good quick pass, but there's, like, a few key kind of differentiators there.
Nick Gomez:One, they how they do it is they very quickly do it. Right? So, they're not, like, artificially using, like, a headless browser type of thing, crawling the website. So they're just doing whatever they can, get as quickly as possible, and that's how they index it. So the and then the other part is so sorry for that, they don't always get all the content because of that.
Nick Gomez:And so it's meant to be more like a quick, like, first pass type of thing so that there's a little bit of a mixed experience there. And then 2, from a company standpoint, they can actually with Inkkeeper, a tool like us, they control the experience and they can add knowledge to it, that you wouldn't otherwise get through like the public internet. So an example there is Discord, right? So Discord is not in the public, wide web, so things like Cursor can't index that. But also teams oftentimes use support historical support tickets or other content Yeah.
Nick Gomez:And knowledge, that they have that are that is private, to help power the the chatbot and and the knowledge. So typically the like, the you know, on AI Copilot specifically maintained by your team and overseen by your team and that uses some of these private sources, will be better than something that can only scrape, like, the the public web.
Jack Bridger:Yeah. Yeah. Yeah. That makes sense. That makes sense.
Jack Bridger:Yeah. And and what does, like, a good like, I guess you've worked with a lot of support teams now. Right? Yeah. Have you seen any patterns for, like, what, like, good support teams look like at dev tools?
Nick Gomez:I think the best ones are really good at closing the loop. Right? And it's part so oftentimes, it's, either because the the support folks are product oriented, like people who really care Yeah. About product and they think of support and self-service experience as part of the product. Mhmm.
Nick Gomez:Or they're product people who are actually leading, like basically, the the the lines blur, very heavily when when it's, like, operated well. And so part of it is like, hey. If if we're answering stuff on Discord and support tickets, is that getting reflected back into the documentation or help center or wherever that that content lives? Are we, you know, listening to the common things that people are asking for, and are we, not just answering that stuff in the documentation, but, like, are we incorporating that into our road map? Right?
Nick Gomez:Like, are we prioritizing correctly based on, what people are asking? So that's part of what we actually help companies do is we take all of the conversations people are having with the AI, and we provide them analytics on the specific things that, where there wasn't a good answer based on your existing documentation or where the bot said that something wasn't supported. So those analytics, and it doesn't just, like, flag the questions. We actually summarize for you using AI, exactly the thing that someone was looking for that was not answerable. Yeah.
Nick Gomez:So we we we make these, like, weekly and monthly reports, that documentation or product teams can use, to actually prioritize their content road map or their product road map as well. Yeah. So I think teams that, the the best teams that I've seen are really good about closely monitoring all of that and actually taking action such that the knowledge base gets better, the product gets better, and you get support questions, right, as those two things happen.
Jack Bridger:Yeah. That's interesting. So you I think, yeah, you're saying that it's not just about answering the question. It's about, like, using the information that the question gives you to improve the documentation so that question doesn't get get asked again or gets asked less often, or people just don't have that issue anymore because of, changes to the product. Exactly.
Jack Bridger:Okay. That's very cool. Yeah. And in terms of, like, actual, like, who does the job, I you mentioned, like, product teams. So, in in that case, it is, like, actual sort of, like, core engineers who are, like, answering support, or do the good teams typically have, like, you know I've seen people with, like, titles like level 2 support or, like, level 3 support, that sort of thing.
Nick Gomez:It varies a lot by by team, and there's not really, like, necessarily prescription for how you should do it. So sometimes it's DevRel folks, who are kinda like the the first line of defense for especially, like, community questions. So So they'll be the ones, like, answering stuff on on the Discord or on a forum, that type of thing. Sometimes they support team is the one that's staffing the the community channels, that type of stuff. We do see it often actually.
Nick Gomez:It's like a support team does only your your paid customers or like the, you know, proper, like ticketed, through through, like, a support form. And then the the community will be, like DevRel or product folks, stuff like that. So some folks split it that way. Some some folks, support team owns all of it, etcetera. It's really kinda like it it depends on, what your business model is.
Nick Gomez:Yeah. And just how your your team is split up. But at the end of the day I think what matters is just whoever whatever the title is or whatever work that they fall under. The the thing that matters most is just like 1, really caring about, like, answering as quickly as possible. Yeah.
Nick Gomez:And then 2, just having knowledgeable people about the product. Yeah. At any of those layers. Yeah. And if you if you have those two things, then, like
Jack Bridger:thoughts go wrong. Yeah. Exactly. Yeah. That makes sense.
Jack Bridger:This is kind of like a slightly off question, but, like, do you see many companies that are, like, charging specifically for support?
Nick Gomez:Yeah. That's a it tends to be a a a pretty big differentiator for, open source companies, or very, PLG kind of bottoms up Yeah. Type companies, where they're they're at you know, how they're part of the upsell or to a paid tier or an enterprise tier, is dedicated support. So some of those interactions happen over Slack, where for an enterprise customer, you might create, like, a 1 on 1 Slack channel for folks. I've seen cases where you actually can't create a proper support ticket unless you're you're a paid customer, because those have, like, SLAs and other kind of contractual things attached to them.
Nick Gomez:And so the if you want support, then you only you can only go to the Discord. Right? So, yeah, like, access to to higher levels of support tends to be one of the the things that, companies try to differentiate their their paid, plants or enterprise plants with.
Jack Bridger:Yeah. Okay. Yeah. That makes sense. And then sorry.
Jack Bridger:These are, like, so many, like, noob questions on support because, like, it's not something I I know much about at all. Yeah. But, like, when you also get, like you know, if you're in one time zone, if most of your team's in, like, a certain time zone, like, have you seen, like, good solutions for people in terms of, like, making sure that actually things get answered outside of your time zones?
Nick Gomez:Yeah. I mean, that that's challenging and hard. Right? Like, if you're not in person team in SF. Yeah.
Nick Gomez:It's gonna be hard to answer, like, European customers or folks in, like, Israel and just other parts, because of the time difference is is so big. I mean, honestly, part part of the answer there is if you have an AI support bot Yeah. Yeah. Users don't need to interact with a human to to to be able to get 247 help. Like, one of one of our customers, Windmill, the the founder Yeah.
Jack Bridger:Literally Ruben's been on, for sure. Yeah.
Nick Gomez:When, he he literally kinda gave gave us a feedback of, like, it feels like we have, like, an extra team member, that's online 247 and helping our users. Yeah. So that that can be, you know, one way in which you can make sure that regardless of the times when you're in, you're able to get some level of, support, regardless of, you know, whether your team's awake or not. Yeah. That's cool.
Nick Gomez:But yeah. If you're a distributed team, it it might be a little bit easier just because you can literally, you know, have folks that align better with the different time zones and stuff. Yeah.
Jack Bridger:But I guess otherwise, like, if if someone does want to, like, have you know, obviously, they're gonna use they should use something like InKeep, but then they probably also need, like, someone that for, like, the escalations and stuff where, like, there isn't a solution available, do they is it is it, like, have, like, a, rotors, like, in like, where people work late, or is it, like, you know, hire someone in another time zone that you see is, like
Nick Gomez:technical support is somewhat difficult to, like, offshore. Yeah. Right? Because it's, it requires a very high level, of knowledge of the product and engineering and software and that type of thing. So most of the time that people wouldn't I would recommend people hire, like, if they do hire internationally, it's like a like a full time person that's part of your proper team, is very intimately familiar with your your product, is like a core part of your support.
Nick Gomez:So it Mhmm. I wouldn't look to kinda offload it, to some type of external service, or like a contractor type thing. I would just look to hire someone really good. Sometimes it's it doesn't have to be a support person. If you just have, like, an engineer working in in a different area, they can help field some of these support questions.
Nick Gomez:And we see that a lot, especially in in early stage companies. It's it's really a group effort, where where founders actually want all of their team to to be answering questions, because it it builds empathy within the team
Jack Bridger:Yeah. Yeah.
Nick Gomez:And about what what people are asking for, what they're running into, and it add just adds like a layer of responsibility as well. Yeah. So, yeah, it doesn't have to be a support person. You you can hire, an engineer, product person. Just and someone who's knowledgeable about the product and can help field some of these questions when, when the kind of the rest of the team is is self flying.
Jack Bridger:Yeah. That that makes total sense. Yeah. And have there been any kind of, like, surprising things that you've, like, learned about support and, like, good support?
Nick Gomez:Surprising things. You know one really interesting pattern that I kind of realized is the answer to so community building is like a very, important thing for a lot of companies, right, like building a Discord or a Slack, creating champions and that type of thing. What I kind of realized is that part of the reason why I think this material, that approach materialized of like having a Discord community, was to help bring on other folks who are knowledgeable at your product to answer questions. It's almost like crowdsourcing support. Yeah.
Nick Gomez:Maybe most folks wouldn't frame it that way, but it is a way to get, it doesn't require you hiring someone. Yeah. It just requires you creating a community where people feel so passionate about the product that they're willing to help others, right? And developers tend to be that persona of where if they use a tool a lot they become passionate about it, want to be part of a community, want to help others, right? Like I think everyone kind of grew up using like Stack Overflow and that type of thing, and the I think there's like a built in sense of like, I've gotten help from other people so I'm gonna help other people as well.
Nick Gomez:And so community building is like a form of support as well, where the more champions, you create for your product, the more folks who are knowledgeable and can actually help offload a little bit of support to to the general community and others. So that's been like a I think that's an interesting pattern where you you don't really see that with other types of products. That's I think that's pretty unique to developer products, and in part because of the nature of developers and kind of their willingness to help others.
Jack Bridger:Yeah. Developers are super social and, yeah, very great.
Nick Gomez:Exactly.
Jack Bridger:Yeah. One of the things I I don't know if this is, like, an actual topic that you're happy talking about, but, like Yeah. Something that you told me is that, like, some of your customers have been investing in you. Is that is that a is that a topic we can talk about?
Nick Gomez:Sure. I yeah. I mean, the the TLDR there is, like, sometimes folks, are skeptical of AI, and then their product team or support team or who whoever, we end up working with, like, ends up adding it, and then they hear great feedback from other end users or the internal team members that are like, oh my god, like, we tried so many AI solutions. None of them were great, but, like, this really made a difference. We've had a few of those cases where then the founder is like, oh, actually I was so impressed by this whole thing, that you converted not just people within my company, but also me a little bit about like the value of this.
Nick Gomez:You know, would you be open to, like, an angel check and stuff like that? So, you know, we've taken on a few of those, for, folks we we find that are, like, good to have, on our on our team.
Jack Bridger:Yeah. That's very cool. Like, is that I get are there any, like is that I guess there's a lot of benefits to that or it's just like a big
Nick Gomez:Kind of benefit. It's just it I think it's good to have folks who really care Yeah. And are invested in in you and if if that makes sense. So, like, as we develop new things in the future, it's very easy to kinda get pick their brain about, like, how would you, would you find this valuable? What messaging would work?
Nick Gomez:That's a thing. So they're just good advisors, honestly. Yeah. So oftentimes that's that's why we do this is if if we feel that, it's it's someone who, really cares about the space. We have a good repertoire with, and, we feel could have interesting product insights as we, as we build new things.
Jack Bridger:Yeah. I think it's really cool. I I feel like it was
Nick Gomez:really cool. I wouldn't rely on it. Yeah. It's a great
Jack Bridger:story, though. It's like, you know, our customers, they they try to oh, you know this person? Yeah. They try to, and then they thought it was so good they wanted to invest their own money. It's like It's not just I don't know.
Jack Bridger:It feels meaningful. And then also when I was talking to, Sean Wang recently, probably this episode would be out by then. Yeah. He was talking about, like, hiring your customers and saying they're, like, like, really good DevRel opportunity is to actually just hire, like, developers that bought your like, bought and brought your tool into their company.
Nick Gomez:Mhmm.
Jack Bridger:Because, you know, they, like, believe in it so much, they put their neck on the line there. Right. And they were able to, like, actually sell it internally. Yep. And so it's kind of interesting area, like like, your customers at the early stages could be, like, actually a lot more than your customers.
Nick Gomez:Yeah. Yeah. That's, part of when I was talking about community building
Jack Bridger:Yeah.
Nick Gomez:I've seen that pattern a ton as well where someone who's super active in the community answering questions, that type of stuff, and seems like they're extremely knowledgeable of the product, and are very active. I've seen a ton of, companies end up actually kind of reaching out to those folks and be like, hey, actually, either it can be for engineering roles or support or DevRel, depending on who the person is, but that that does happen, quite a bit. So, yeah, it's I don't know. Dev tools has that kind of quality where like people are really invested and there's a little bit of like fandom or not not fandom just like appreciation for good quality craft. Yeah.
Nick Gomez:Right? And I think it goes both ways, right? So like developers appreciate really well built products. Yeah. And then vice versa.
Nick Gomez:I think companies, teams appreciate community members or or customers who appreciate that craft and, like, build on top of it or champion it to other folks. That's what the thing. Yeah. So it's it's very interesting. It's like a virtuous cycle
Jack Bridger:Yeah. Which is cool. Yeah. It's very cool. It's, I mean, it's one of the best things.
Jack Bridger:Anurag, from Rendo was kind of talking about, Intercom on the last episode for support, and I kind of thought I'd throw you a hard question of, like, how do you guys compare to Fair to Intercom? Intercom.
Nick Gomez:Yeah. We we get that question a lot. A decent amount especially actually from, like, investors or folks not in the developer space. But the the the the the short there is, like, Intercom is quite good at what it does, but what it does is doesn't apply that well to developer products. So what I mean by that is, for example, an intercom chat bubble is never gonna generate, like, a code snippet for you.
Nick Gomez:That's just not what it's, like, designed for. That's not what who the end user is designed for. So, like, intercom, probably like half of their customers are B2C or like consumer facing scenarios, right? So it's not all B2B SaaS, probably a fairly small portion of it is actually developer products that's going to so it's not just like it's not a primary use case for that product, right? It's actually one one of the inspiration points for us when we were starting was I noticed that for b to b SaaS, like traditional b to b SaaS products, something like probably like 90% of them use some type of chat bubble so like either Intercom or Drift or any of those.
Nick Gomez:But when I was looking at developer products at some point it hit me I'm like I'm not seeing that little chat bubble, right? I would see it maybe in 1 in 5 so like 20% instead of like 90%. So there is like some core value prop that intercom or these chat bubbles provide and what I ask myself is like what value are they providing to these types of companies and why doesn't that apply to developer products and what is the equivalent in developer products, right? So that that was that was actually part of the inspiration for how we get started is like what what is that gap there? Why is there that disconnect?
Nick Gomez:So yeah, in terms of how we compare to, for example, intercom, so geared specifically for technical developer products, you can embed us in a few different, like, as React components, instead of, like, iframe thing. You can add us, like, as a search bar, as a Discord bot, Slack bot. So part of it is just, like, the end user touch points are different Yeah. And we accommodate kind of for wherever you're meeting your users, or wherever your users interact with you, that that type of thing. And, yeah, just the overall end user experience is is oriented around, developers and and what they would enjoy.
Nick Gomez:Part of it too is, if you if you took poll developers, I would say a majority of them would tell you that, they actually are, like, allergic to those little chat bubbles. I don't know if you've experienced this yourself, but they a lot of people, developers in particular, associated with, like, marketing and sales because it's often used for that Yeah. By kind of traditional businesses like lead generation that type of stuff. And so developers don't want to be sold to that they don't want that experience. And also some of them feel like it's partly inauthentic and that they associate those chat bubbles with, like, I need to click a bunch of buttons to actually talk to a human and that, because the AI or the automations there are not very good and are not actually gonna answer my questions.
Nick Gomez:Right? And so I I think, developers actually don't use it that much, right, and aren't keen to to use it that much. So for us, like, everything from, our actual AI and ML pipelines and how we ingest content to the actual end user experience like the UIs, are oriented around kind of, avoiding those, issues that developers have with, you know, traditional chat bubbles.
Jack Bridger:Yeah. Okay. That's very cool. That, yeah, that that makes sense. Like, devs are skeptical of these things, I think.
Jack Bridger:Yeah. Yeah. Okay. One last question I wanted to ask you is, like, what developer tools are you guys using that you like?
Nick Gomez:Oh, I just I mean, it's my favorite pastime. Right? Let's see. We use Drizzle for, like, a RRM. We use Vercel to deploy a lot of our stuff, GCP for, like, Docker based images.
Nick Gomez:Yeah. I mean,
Jack Bridger:all all the okay. Cool. Any new ones that you've, like, any kind of, like, startups that you're excited about?
Nick Gomez:Oh, man. We started using resend, probably not that nice, shocking. So for yeah. Shout out to Zeno. The you know, they have a really sleek UI support, everything.
Nick Gomez:I think they they're pretty representative of, like, what is good modern developer experience, and support across kinda like the entire life cycle. Yeah. So resend, let's see what else other other new things that we're trying. Recently started using Langfuse for, like, a a bit of observability, like, LM observability stuff. They're they're like a YC company.
Jack Bridger:Yeah. So that's, like, when you wanted to see, like, what what kind of was the prompt and what was the response to
Nick Gomez:the Yeah. Yeah. Exactly. Exactly. It's like monitoring all and applications, that type of stuff.
Nick Gomez:Yeah. Oh, one thing it's not really a start up, but, the Vercel AI SDK, it's it's basically it's a way to call and, AI, like, models, like OpenAI, and topic, etcetera, and easily kinda turn that into, generative UI is what they're calling it.
Jack Bridger:So like the v zero stuff on an API?
Nick Gomez:Yeah, basically like being able to easily turn a LM call into like dynamic UI components.
Jack Bridger:Oh, that's cool.
Nick Gomez:So for example, we use that in our reporting layer where in real time you can come into our reports, say, I want to see the documentation gaps for the last month, right? You click a button, we do our AI stuff and then in real time we're using the Vercel SDK, we can generate, like, those insights for you, the things that, like, summarize for you, like, here's x y z, and we render that kind of in real time. Oh, wow. So we stream it to the UI.
Jack Bridger:The UI based on the response that you based on the data you get back?
Nick Gomez:Yeah. Basically, as we process it and as we say, okay, here's, one topic area that you could be better at and here's the specific things that people couldn't find, that get that gets streamed to the, to the UI. You see that, you can start reading it Yeah. And then it's continuing to do that for the the rest of the the topic areas.
Jack Bridger:Oh, that's fine.
Nick Gomez:And it gets streamed in in real time. So there's not, like, an async process there. Yeah. But there there's a ton of things that people do with it. So very another common example there is, like, I have a form.
Nick Gomez:Yeah. Right? But I want to provide an app like I don't want to have someone actually go through the entire form I want to just have give someone a text area they can describe what they want or what they're looking for. This is so this is common for like support tickets. Yeah right You can actually take that information, call, an AI model and map what they put in the text area into the forms, into into the input fields
Jack Bridger:so that you're getting the kind of benefits of, like, filtering and stuff like that, but they're getting the benefits of, like, just
Nick Gomez:Exactly. It's basically, like, turning unstructured natural language into structured objects of some sort. And then the Vercels here makes that makes it easy to take those structured objects and then stream them as UI components
Jack Bridger:Oh, that's cool.
Nick Gomez:To to Yeah. Find an application.
Jack Bridger:That's cool. Yeah. Very cool.
Nick Gomez:Yeah. And it's very interesting, application development I think is going to change a lot and what UIs look like, what apps and dashboards look like, because there's all I think untapped, or still maturing side of using AI and and these models, is around how do you how do you make all these end user experiences richer and more intuitive and dynamic. So front end development I think is changing a little bit because front end developers now can use these like it's democratized they can use these AI models to actually create really rich experiences. So another example there is like date pickers instead of having to click through and stuff like that I can just say like 1 week from now and then LLM will be able to tell convert that into like a proper date object. So there's just like a bunch of little micro places where you can really improve the experience.
Nick Gomez:And then there's use cases like the the the, you know, real time generative, reporting layer that we have, stuff like that. Yeah. But Yeah.
Jack Bridger:Super cool. For some reason, I feel like day picking, I remember doing. It's always, like, more complex than, like
Nick Gomez:Yeah. And and people are prompt.
Jack Bridger:Things never work properly. Like, it's just there's always, like, some weird bug on, like Exactly. Yeah. That's really cool. Okay.
Jack Bridger:Amazing. Yeah. Yeah. Well, thanks so much, Nick.
Nick Gomez:Absolutely.
Jack Bridger:Yeah. And where can people learn more about about you and about Incub?
Nick Gomez:Yeah. At incubed.com, you can Nice. Yeah. You you can come in there. Actually, one thing that, we do is, if you have a product, it doesn't even have to be a developer product.
Nick Gomez:We have a fair amount of non dev tools, as customers now. You can come in, drop in a URL to let your documentation or your help center, whatever it is, and we'll automatically generate a demo for you to try. You don't have to meet with me. You don't have to start, pay pay for that.
Jack Bridger:You are the best guy if they do want to meet you.
Nick Gomez:Yeah. Yeah. Hopefully I'm fun to talk to. But yeah. So if people want to check it out for their own documentation or for their own product, they can come in and very easily do that, through our website.
Jack Bridger:Yeah. Amazing. Well thank you everyone for listening and thank you Nick for joining. Appreciate it.
Listen to Scaling DevTools using one of many popular podcasting apps or directories.
