376: What To Expect From Apollo Next with Tyler Phillips

and I was wondering, you know, are there any features that are going to start showing up inside of Apollo? And so I thought I would use my annoying curiosity and pick the brain of the guy who's in charge of building these things. So Tyler, I'm excited to have you on the show. Tyler Phillips: Excited to be here. And I'm ready, to receive your feedback as a Apollo PM. Collin Stewart: Perfect. Let's start with tell me what you're building and why. Tyler Phillips: Sure. Okay. So I'm a principal PM at Apollo and I'm building our AI platform and the Apollo AI assistant for go to market teams. Recently we launched AI power ups, which is. Essentially your research assistant that will help you automate your research for any bespoke data point you're looking for fine tune your targeting and Apollo, and then build out personalized messages built on top of that with our sequencing tool. Collin Stewart: Gotcha. So you're kind of building clay inside Apollo. Tyler Phillips: I, I don't see it that way. I think Clay's definitely done a few

things right. Especially in the kind of. Use case kind of exposing that like a lot of outbound is around creativity and finding the alpha and not just like the traditional filters. So definitely they nailed that. And so it's definitely something to be inspired by. But if we have a lot more of that kind of building, that's a little bit separate from the company like clay. Collin Stewart: I love it. I love both tools. I use actively use both tools and I, I have my loves and hates of both, both of the tools, cause I use them. And so I'd love to know specifically what's unique about the direction that that you're going. Tyler Phillips: Yeah. So I think like the first thing is for Apollo is, you know, we're still preferred as like the initial list building stage. Like even yourself, like, you know, we kind of talked about, you like to filter down your audience on Apollo, you know, from like thousands of records, you know, down to a fine tune list. So I think like as being like the initial starting point for you know, you're building out your list, having AI kind of infused into that experience where you can filter, let's say like thousands of

records in the Apollo database is like super powerful as a starting point. So you kind of an AI first prospecting experience. I'd say like the second thing too, is like, you know, naturally being kind of all in one tool, you kind of get connected to, you know, all your, your data's inside of Apollo, as well as connected to all Apollo data and like kind of emails and prospecting data. And then finally, like kind of the direction that we've taken, you know, based off our user base, which is skews to them where like SMB, you know, like smaller sales teams is to make it a lot more easier to use. Like people don't. You know, want to learn how to write really complicated prompts or add a bunch of columns. I just want to get to like the data point they're looking for to then write a personalized message and target off of. So that's kind of what we've been focusing on and we've seen pretty significant growth with that focus. Collin Stewart: I love it. Even when I am using clay, I always like to start in Apollo anyway. And I know there's lots of people out there that use clay that say, oh, you're doing it wrong. I've been using clay since the time when you used to have to go into sales and do the, like the clay. Google

Chrome extension and loop through and just click and click and click. And it was really powerful, but super annoying. I really liked being able to start with like the entire universe and then apply kind of chunk it down and apply filters and see how all the numbers shift. And I find Apollo is a really great place to do that. What I was excited about was being able to run kind of prompts at scale. On top of that Apollo dataset and like use my own custom, create my own custom filters to like filter it down. Tyler Phillips: Totally. Yeah. I think we're seeing that like play out too, with a lot of like our users that kind of using like the fine tune targeting use case. Like they come into Apollo, they try to use, you know, our traditional filters, like job filter or job titles, industry company size, and they're like, okay, no, I like, this is not enough. This will still like thousands of records. And like, okay, now I want to like really get down to like a subset of audience that like has a fit for my product from like basic filters. Then also like have shown a proven pain point. That we can solve. And I think that that's kind of where the market is going in general is like the right

person. And I think that's something that Apollo uniquely can win at. I Collin Stewart: like it. I also, I think if I'm looking at kind of the differences between the two I really like clay cause it's infinitely configurable. I can do just about anything. It's sort of like having like console access or like. You know, God mode in doom, you know, it was IDDGD or something like that was like the code you typed in, you could like walk through walls and get all that. Like clay kind of feels like that. Whereas the Apollo experience is a little bit more like on rails and a little bit simpler to use. Tyler Phillips: I mean, I also think it's like kind of like the general, like, you know, there's kind of like the, like, if you think about like the user of Apollo, like it really like is, isn't really the user of clay for the mass audience. Like either loser of clay is like, you know, maybe more of a growth, you know, I guess they have a GTM engineer, but like the growth hacker and like Apollo is like more for like your traditional guests to SDRs. AEs and like also like, you know, RedOps kind of supporting them. So I think it's just like different like personas and use

cases and also segments in the market and also like the level of comfortability of using AI is very different. So I think like with that, you kind of produce like different outcomes, just why Apollo is moving more in the direction of like. You know, usability, ease of use and kind of high, you know, high quality for maybe like a more like just mass market use cases where like maybe clay is kind of going more like the integration pathway. So it's kind of a divergence in thought. But I do think like the alpha piece of like, you know, finding the right alpha and like that still exists. That's something that we still want to support, right. Where you can. Find, you know, people have, companies have gaps in translation of their website. Like that's something we want to support for our customers. Collin Stewart: I love that. And just if we dive, if we double click on the finding alpha, cause I was talking to Brendan a couple episodes ago when we had a deep conversation into this. So I'll throw a link to that in the show notes, but the explain to the audience, if they maybe didn't catch that episode, what is, what are you referring to when you're talking about alpha? Tyler Phillips: So, I mean, I think like, you know,

as you were like thinking through like building out and like an outbound campaign , obviously traditionally you're going to think through, okay, like what are the different, like company size, , job titles, maybe even job postings or whatever that like come traditional, like targeting methodologies. But I think like today with kind of like the flooding of AI, , messages, as well as just kind of like the spray and pray model, like doesn't work anymore. So there's like this, at least, , I think, , Brennan coins that is like micro campaigns. I don't, I mean, micro campaigns is what probably won't go too deep in the definition there. But how I think about it is like. Searching for that alpha using AI, where you're looking for a proven pain point that this person has, that you can uniquely solve. And I think , that's something like you're thinking through, okay. Like when someone joins a conversation with me, , what is the problem they come to you with? And like, how can I actually ascertain using AI? Like. Breadcrumbs of that problem. And that's something I think that I'm really focused on learning to like how we can build that into our product. And I think with power ups, that's like our first foray into this kind of like alpha identification outside of like your traditional filters. And also I

guess intent. Yeah. Collin Stewart: Yeah. I think put succinctly if Apollo adds a filter for, , B2B SAS, right. Then being able to generate your own B2B SAS filter in clay. Or with one of these tools isn't going to be as powerful anymore because every once everybody has access to it, then it's going to be less valuable. Tyler Phillips: I think that's true, but also, I mean, the counterpoint to that is , if you think about like a perfect market, , , there's no, I guess perfect marks don't exist. I was an economics major, but , if you do think about it, like, if you have a product that this person actually has a problem and needs to be solved and then you're able to reach out to them, That technically, like, regardless of if anyone else has that information, that will be valuable, right. For you to commit, commit the outreach. So that's like the only counterpoint to that. I do agree with alpha sentiment, but like, just because it's available, Apollo, like doesn't mean that it becomes immediately commoditized. But that's my only kind of, Collin Stewart: I strongly agree and I don't think it'll be, I don't think things get immediately compromised. I think they kind of

follow the technology adoption life cycle. So the innovators get the greatest value or benefit out of out of using the filter or the channel or, you know, that, that feature, that functionality. And then the early. Sorry, it's late innovators, then early adopters, then early majority, though, those get like the majority of the benefit. And then once it's making it to the laggards, then it's, you know, Fewer, fewer benefits. Anyway, I wanted to. I wanted to dig in on like, what's powering the models. Cause I was playing, I spent pretty much most of Friday playing with a mix of clay and Apollo kind of together. It was a fun little dance of like, it wasn't in prep for this podcast. It just happened to line up that I spent, you know, most of the day. And so I'm curious, like, I love, one thing I love about clay is I've been able to pick the model. I think it's a little bit less. Important now, like back when it was like 3. 5 or four being able to select GPT 4 was incredibly valuable. I probably don't want to scale up an O1 prompt

and have it run. Cause one, I'm not doing stuff that that's, that needs that level of complexity and reasoning. And two, I don't want to wait the like potentially 30 seconds to a minute on every one of those thousand or 10, 000. Kind of rose in the list. And so I'm curious what's going on behind the models. Cause I don't really, I can't see that. And so I was sitting there going like, am I, am I stuck using like a one mini at best, or am I using like three, like I'd love to, I'd love to understand. Tyler Phillips: Yeah. So, I mean, I will say this, like we're constantly updating the models for the given use case. It's like, we just have to migrate it to perplexity sonar model. So we, cause we've been partnering with Perplexity OpenAI. So we have , kind of like as of right now, three models or like , you know, use case models that we have, and then I'll kind of explain like what's later on top of it. So we have the AI search engine model, which is using perplexities. So in our latest model for the research use case secondly, we have the AI text generator model, which is open AI for our mini. And that's mostly for like

categorization use case. So for example, the labeling company B2B SAS, you don't really need much more than just like an open AI for our mini model. And the final one is entropic Claude Haiku 3. 5. Which we've seen to be like fairly cost efficient and also pretty good at generating messaging. It's like well known to the market to be good at like sales copywriting. That's all to say though, like. Well, we noticed a lot of our users, like we default to the search engine model and almost like 95 percent of the time everyone uses the default model. So we've kind of taken a step in the direction of, okay, so when, when, when a user, instead of a user, like selecting a model and writing a prompt and the referencing variables. What if you could just write in your own words, like what you're trying to do. And then essentially we will pick the model for you based off your use case and then generate the prompt and like attach the variables in Apollo's database to personalize the prompt. And then we noticed like with that, like. It's much easier for people to understand and use. And then if they want to go and tweak it, they can, which is more like the advanced user, which is matches more of

clay, but like that's generally what's worked for our user base. Collin Stewart: Funny story. I consider myself fairly good at clay. And when I was working at Apollo on Friday, I found I was using the like, explain the prompt and just having Apollo come up with it for me. And I wasn't going in and doing the, the work of like assembling the prompt myself and picking the variables. Yeah. I was noticing, I think it was because one, the like prompt generator piece did a fairly decent job. The area where I did get stuck is I was assuming that I had access to the full Apollo data model, which would be like any of the filters on the left hand side, so, you know, if I could run and apply one, I would assume that I would have, I would be able to reference that information in the model somehow or in my prompt. I get that's a more nerd advanced level feature. Is that something that will be coming eventually or is there a workaround right now? Tyler Phillips: Yeah. So, I mean the workaround, so we have not every single like data point is supported right

now within the like prompt recommendations. But yeah, I, I, over time we'll continue to index more of our, like, you know, you know, database, like all the different data points we have, and even like things like conversations, notes deals, et cetera. But yeah, at this current moment, like not all of it's available within the prompt recommendation tool. The reason why that the case is because we've been training our models to be able to predict Like what variable to use? It just doesn't, it only covers like 80 percent of the cases. And so for any specific variables, for example, like buyer intent, that's something that right now we don't cover within the recommendation. Collin Stewart: Is that something that I could say, I think the one I was looking at was like, I wanted to filter by the number of people that had I wanted to filter by have like X number of salespeople. And so I wanted to apply that filter. And then I wanted to use that data in the in the prompt. Tyler Phillips: Gotcha. Yeah. So the, yeah, so the, the, that that data point is not available at this current moment, but we are working to add all the data points to be available to the AI and I think like a broader, Collin Stewart: what, come on, man, why not,

Yeah. And I think that's also like generally like the trend array of like, kind of, if you're, if you're optimizing more for usability and ease of use, covering that was like, you know, 99 percentile cases will become a challenge, but I do agree with you, like the, all the data points of how it should be available within. Collin Stewart: Cool. I'm curious from what you've seen so far, and I can talk a bit about some of my favorite things, but what have you, what have been your favorite things that you've seen used with the, the, some, like some of the workflows and the assistant there? Tyler Phillips: Yeah, sure. So like, I guess one example is we've been, you know, we partnered with this AI translation company called smartling and they were using our power ups to basically identify accounts that had a gap in translation on their website. So you'd look at, you know, what are the different websites it's translated into, and then you kind of look through, okay, as many gap in translation. And I thought that was a pretty powerful use case. Kind of goes back to like the proven pain point or like the alpha that you could really call into and like the message basically writes the message itself. That's one

example. Another example is really interesting is like kind of this, like a competitive advantage where you basically use a power up to determine. You know, what is like, you know, what is a competitor of this part of this company doing in like, you know, your service area. So in case for Apollo, it's like the generation and you basically like in the email call out that like, Hey,, you know, if you're targeting a company, like, Hey, like your competitors doing, , X, Y, Z and lead generation , working, let's say with like, , whatever, some ASDR company or whatever, and then like, Oh, like, , maybe Apollo would be a good fit for you to kind of compete with them to get ahead. And that kind of creates this like. Almost feeling of like you're behind. And when you feel behind, you want to kind of claw your way back. And I think that is a pretty interesting use case as well. Collin Stewart: Interesting. I want to go back to the Smartling example. So with, when they were, and if you can't say exactly, if you can't share exactly how that I totally understand, but help me. So they were talking about gaps in translation and there was like a before and after. So how, how were they doing that specifically? Tyler Phillips:

Before the power ups? Collin Stewart: Sorry, when they were using it. Cause the, I thought the gaps in translation was like, they had a snapshot of before and after and they were doing, they weren't comparing, so. Tyler Phillips: Yeah. So, , like basically, we pretty much just coded their like manual workflow where we basically look at on the website, they typically have like a toggle on the top, right. Of the different languages. And you basically first extract, and AI is a lot better if you do it in multiple steps, so first you extract, you can hear all the, the. The language is translated into, and then you say, okay, like basically for these languages, let's say like, you know, Spanish is any like language or is any like, content on the Spanish webpage in English? And then that's kind of how we would identify the gaps. We'd have a list of languages and we just kind of cycle through the website. Collin Stewart: And so were they able to do this in the Apollo UI, or was that something that had to be kind of. A little custom coded or custom collusion. No, Tyler Phillips: yeah. So we, yeah, we did it all in the poly UI. The used first step was to like, you know, use a search engine to identify the languages, the translate, the

websites translated into, and then the second step is basically just traverse the websites and determine if there's a gap or you notice his English on like the Spanish or Italian or whatever. But giving it kind of like the, the boundaries of like the languages to look at yeah. Giving it the languages to identify. So that was all be able to do inside of Apollo. It's just multiple columns. Collin Stewart: Gotcha. So first column would be go and find the languages supported on this site. Kind of look for that website. Second step would be like, okay. And maybe you're outputting the links. Second step would be, okay, find the Spanish page. Go and scrape the Spanish page. And then step three would be like, look at the text from the Spanish page. See if there's any, if there are any English words in there. Tyler Phillips: Exactly. Yeah. And then that's the end of it. If you see there was a gap, it's a fairly easy email to write, right? You're just like, Hey, , notice that your Spanish web page has like, , bleed through with an English, you know, our service can help you translate your website content and, you know, for like whatever, 30 percent of like the cost of paying an outsource

tool that kind of, it honestly writes itself when you kind of build out in that way. Collin Stewart: Cool. Sweet. I appreciate you sharing. I've got some, this is the selfish part of the interview where I get to ask you all my favorite questions and things that I would like to, that I think you're going in the direction of and I suspect other people are kind of struggling with. And so the first thing that I hate doing with Apollo is the, the CSV dance. I always build a list, download it, upload it into zero bounce, upload that into Google sheets or upload the output from zero bounce into Google sheets. Then I do the filter of like, I grab all the valids and I grab all the catchalls that are I Collin Stewart: think it's category. I can't remember the scoring, but they've got a scoring system out of 10. And so I grabbed two and above at a zero bounce. And then I take that and I upload it back into Apollo and I'm like, here's my list. And so I've heard that you're working with Zero Bounce. Is there something that I'll be able to

do to prevent or to save me this CSV dance? Tyler Phillips: Yes. A short answer is yes. You'll be able to kind of filter on verification status using zero bounce. So that's something that that's definitely is rolling out right now. In terms of like more broad, I feel like just to dig a little bit deeper into like the problem that you just shared though, is that like, we recognize that like, You know, importing, enriching and exporting data and Apollo is not the easiest thing to do with the planet. Arguably, probably one of the harder things, although it shouldn't be. And so I think we're really trying to invest in like dire, especially to like, make it a lot easier to kind of complete these enrichments very quickly. You know, with a combination of, , Additional database providers, not just Apollo, and then also using like the validation source. So inside of Apollo, but then if you also want to export, you can do that fairly easily and also add columns. So that will be built out. Collin Stewart: Cool. One random observation. I use Apollo. I wouldn't say daily, but multiple times every week and I Collin Stewart: have downloaded. So many

files from Apollo onto my computer. And every time I'm in settings, I can never find the exports button. I sit there and I look at it like an idiot, especially if I'm showing a client, be like, okay, now I'm going to go download the list. And then I look at it and I always zoom past it. I'm like, I know this is just user error. Like problem exists between keyboard and chair. I wish there was a simpler way of like, I've exported the list, like. Give me a better way to like track it down. Cause I'm often working with three or four lists at a time. And I did this on Friday. I was doing something for a client. I had three lists and then I downloaded them. I went and had lunch. I came back and I was like, shit, which list did I download first? Which one is list 86, which one's 87, which one's 88. And I went, fuck it. I don't want to screw this up. So I went and redownloaded them and sat and had to wait. . Tyler Phillips: Well, I, I'm kind of curious if you don't mind me asking, like, Collin Stewart: yeah. Tyler Phillips: What would your ideal experience be there? Collin Stewart: The list, Jen Ferry comes and just like puts them, does all the things for me. Okay.

pasted into a Google doc as a hyperlink where I named the search criteria. And I do that for the account and for the contact list. So I don't screw anything up. And then I go and I download them and I go, I downloaded this one first, I click download first, download second, download third, so that I don't mix anything up. Because for me, if I'm sending a client, a list to a client or loading up a campaign and doing that kind of dance, then. I need to make sure that I'm not going to send them the list with the wrong context or upload the list into back into the sequence with the wrong context. Right. Cause often I'm not just downloading. I'm not just working in Apollo and then downloading and then sending it. Right. I'm going to download it. I'm going to go do some other stuff in clay and then I'm going to re upload it, but I'm going to the client first for approval. So name it, make it easier, make it auto download with like the name of the search. Tyler Phillips: Gotcha. All right. Stay tuned. Collin Stewart: Awesome. I would love a global, I want, I want clippy, but for email deliverability. You

know, like, I want an annoying little thing at the bottom right that's like, yeah, it looks like you're trying to add some. Emails to the sequence. Can I zero bounce them for you? So like any time, any of the people on my team add any emails from Apollo into a sequence, I would love to have something that automatically zero bounces them. Tyler Phillips: Yeah. I think so that comment on that specific point, I think you'll be excited about what we're going to release in a few months related to the idea of Clippy and giving you suggestions and stuff like that. So I don't want to Reveal too much, but definitely something in the works related to kind of assisting you as use the Apollo product. Collin Stewart: I don't know what I can say or can't say about the beta, some of the beta stuff we're testing, but we've seen some things kind of in that direction and it's, it's encouraging. Tyler Phillips: Okay. Yeah, that's good. Yeah. I think deliverability overall is definitely another big focus of ours. We recognize that like, you know, we need to really improve there. And so there's a lot of investment happening in that front. Collin Stewart: Cool. I haven't used it in a while.

Remind me, are you, can I still do warming through Apollo? Tyler Phillips: I believe so. Yes. Yeah. Collin Stewart: Okay. Is that something that's a focus for you all right now? Or is that something that Tyler Phillips: I think for that one, I think we're starting to explore a little bit more, like, you know, maybe Partnering with different companies in that direction. Yeah, definitely. Deliverability is, you know, it's, it's a beast in its own right. So I think in that direction, I think we're definitely looking to, you know, kind of expand out our partnerships there. Collin Stewart: One thing as you're progressing that I would love is this is such a, my, like a, I almost feel silly asking for it, but it is like a power user thing that I like in some other platforms. Okay. If you, if you get, if you're doing zero bounce or even if you're pinging the mail servers, you understand, like in the response, you'll know if it's Gmail or office. And I'd love to do like to like, so when I upload my when I, when I upload, when I add my, or sync my email accounts, I'd love to have, you know, three Gmails and three outlooks. And then Apollo would automatically

send up, you know, keep Gmail and Gmail and outlook and outlook. Tyler Phillips: Yeah. So you basically want to like, be able to have like domain buying on Apollo for like both outlook and Gmail, then match, right? Collin Stewart: I I don't care if I need to buy it. Like if you could buy them through Apollo, sure. But I just want them to match. Like if I'm, if I'm sending to an Apollo or I'm sending through Apollo to a Gmail account, I want that to go from one of my Gmail accounts. Yeah. If I'm sending to an office. So definitely cool. Can you get, I guess I have to ask on behalf of everybody who's listening. Can you get, get rid of our cap on the 10, 000 enrichments? Yeah. You know, if you're nice to me, Tyler Phillips: I think yeah, I think for, you know, I. You know, I think when you're building a great products, it's always nice to break it. So we definitely are going to be moving beyond the cap at some point. And I, but I think for us, we just got to figure out the right, like pricing model for that. And that's something that as we're kind of moving to, you know, more unified credit

system should be good. Yeah. But yeah, I totally get it. You know, people want to, you know, break the bank and, you know, Enrich everything with B2B SaaS or not. Collin Stewart: Yeah. I mean, to me, it's, it's more about the workflow of like, I love that. The best thing about Apollo is for me to be able to start, like you said, start with a huge market and then carve it down. And the most annoying part of the new features that I was running into was I had to click on, I think I double, I, I suspect I used a bunch of credits twice because I would run a column. And then I say, I started, I think the one I did on Friday, there was like 30, 000 accounts in there and I run and then I came back five minutes later and I was like, okay I think I hit run all 10, 000 again. And it was just another 10, 000. And what I had to do was. Filter, like filter on that column. And then I think it was like, is unknown. And then it got me another list of 10, 000. And I think I did it a couple of times. Like, I don't necessarily care about processing more than 10,

000 at a time. I just don't want to have to come and do this little dance of like, click around, click around, click around. So like if we're running an enrichment and there's 30, 000, I'm fine. If you batch it and just let me know, like when it runs, I just don't want to. I want to be able to carve up the market and I don't want to have to do the, like a whole bunch of filtering and remembering and coming back and waiting for things to finish. Tyler Phillips: Yeah, I think that's a fair point, especially because like, you know, if we are saying that we want to be like the initial lead list building place which will have like, you know, large, much larger, like TAM, for example, which is where people start. I definitely think like filtering down to the SAM could be much, much more like simpler experience. That we could, we could have, I'm just curious though, what were you running on like 30, 000 accounts? Collin Stewart: I want to say it was, it was giving, oh, it was does, does this company sell to a technical buyer? Okay, Collin Stewart: gotcha. Okay. And so I, I basically said, Hey, here's the description. I think I'd already filtered down, like sat, like, do they sell

software? And then like, are they B2B? And then do they sell to a technical buyer? I think that was the three steps. I can't remember which one it was. Yeah. But I remember popping into our interview template or doc, and I was like, I'm going to, I'm going to give Tyler shit about this. Tyler Phillips: I'm sure that won't be the last time. Collin Stewart: No, definitely not. No. Matt gave me your email address, your personal home address, your parents address. Like there's going to be letters. I'm screwed. Tyler Phillips: I'm so screwed. Collin Stewart: Is there anywhere I can see where enrichments are running? Like is there any dashboard that shows like, Hey, if I've got 10 sheets that are going to where I can like check in on the statuses? Tyler Phillips: Yeah. So in the settings, it was like power ups usage and you can see basically all the lists that you've run it on and you can click into individual like fields that you Collin Stewart: created. Gotcha. So I can see how many credits I wasted on a particular exactly. Tyler Phillips: Okay. Collin Stewart: And that's in that, that's in the settings area. Tyler Phillips: Yeah. It's under power ups usage. Collin Stewart: Yeah. Yeah.

Fantastic. Any advice on running multiple enrichments at a time? Because I find I'd like to do these like stacked enrichments where late, like we talked about for in, in our previous earlier example smartling, I'd love to do that, but I find if I'm starting with like 10, 000, then. I've got to click it, even if it's just 10, 000, I got to click it and then come back and wait and then build the other one. And I can't manipulate the data model easily enough to like, I guess, finish my job. Like in clay, I can do the first 10 and I know I can do that in Apollo, but I can do, I guess the first 10 and then do the next 10 and I can pass the data into one prompt and pass the data. I guess it's, yeah. And then I want to run all. And then once I've got it, got that chain figured out, then I want to run them all. And I, that's where I got stuck with Apollo. Tyler Phillips: Yeah. So, okay. That makes sense. So like generally, like my recommendation is like, well, before we go into like bulk enrichment is like, start with like the

first, , 25 and then make sure like, , , before you like running like thousands of records, cause you know, once you do that, you're consuming a good amount of credits. Just make sure that it's like, you have a good, like what I would like to call like a test. Data set of like different companies, then even like valid each of them, like check the sources that we display. And just make sure that it looks correct and have like a different thing. And then you obviously can enter in the prompt to make sure it kind of covers those edge cases. Yeah. Tyler Phillips: From there though, let's say you have like three columns, like, you know, let's just take a smartly example of number, like, you know, list of languages, then you like extract the text. And then from there you have you know, you want to like basically have a flag for English play through between like, let's say with Spanish, English. So if you run just the third column, it will actually run all three columns because you have dynamic enrichment. So you don't actually need to run on the first two, it will run, like let's say you run 10, 000 on the third, it will actually try to run on the first two as well. Although based off your credits, if you run out, it will stop. But that's kind of how that works. That's how we solve that problem

right now. But. I definitely get the sentiment that you just kind of want like a run all, but Collin Stewart: yeah, what I really want was like, I want to run it with 25. And I think where I kept running into problems was every time I created and ran, I would click like run enrichment. It Collin Stewart: would, the search view would snap back to saved contacts. And like my first view didn't have, was all like net new. I always click on like always net new. I never work on a saved when I'm doing stuff like this. Cause I want to grab the whole market, not just a small piece. Yeah. And so I, I think what was happening is I would I would go, I'd run the enrichment, it would snap back to saved and they would only have like 50 percent or 30 percent of them that would have run. And then a bunch of them would say, okay, not running. And so I'd have to, or like not hadn't run at all. And so I'd rerun them. So I kind of got into that dance. So is there advice on like, just save a list of 25 and then work on that. And then once you've perfected those three, then push a bigger list in there. Okay. Tyler Phillips: Yeah,

I think that's probably the best. So we, every time, so every power up, you do save a contact or I'll say accounts for free, but you do have to save the contact to your own database within Apollo. So that does, that does happen. So that's probably why you saw that, like you, you went to save, you only saw like 50 percent of them cause it does save and all safe. So yeah, I would create like a list. And then I would run Tyler Phillips: it on that list and then, then I would just add more to that list and just keep running on that list. It's a little bit easier to like maintain the state in that format versus like doing it on like menu. Cause then you'll, it will snap to saved and you know, they will, the basically over time we'll add all of them to it, but it takes time to save all this context and then run the enrichment. Collin Stewart: Gotcha. Okay. I follow you there. I'll do that next time. I didn't realize it had to be saved. So that was my screw up there. Or my, I missed that. Kind of last geeky technical piece here. The I find with clay, like I've got a templates folder that my team reuses. We run, we have similar templates that I've built and that we reuse them across

clients. What, what would you say are the best practices for, you know, folks that want to redo that? Like once I put all this time into say these three and these three columns that do this enrichment, is there a way that I can pull them into like a template or have, have, make it easy for my team to recall that view or that, that kind of stack? Tyler Phillips: Yeah. So yeah, basically any anyone on your team, if you're an admin when you create a power up, it creates a custom field and you can just add that to any view. So you just click on add to view, and then you can click on the fields, search the fields that you have. So it's kind of similar to like, you know, like a notion type or if, you know, notion for like tables. Yeah. You just add the property that exists. So you can do that any view and then that that person can rerun it. We also do have the concept of like a private field. So let's say like, you know, one of your SDRs wants to just kind of create their own power up and have their own alpha, you know, those people generally are pretty competitive. They can also create their own and like, you know, it doesn't have to like share costs and like

kind of muddy the custom fields implementation of your instance. But yeah, that's how it works. Collin Stewart: Okay. And then can I set custom views with like, these three power ups for this view and those four powers for that view? Tyler Phillips: Yeah. And then basically anyone could go off of that and they can either duplicate the view and always have it. Collin Stewart: Gotcha. Okay. So I can do that. And then I can run a list through that view. If I want to replicate that and do it over again, Tyler Phillips: for sure. You just filter on the list and then you have it there. Collin Stewart: I love it. Okay. I appreciate that. If anybody's curious about any of this and they're listening, they're going to call and I wish you had some visuals, drop me an email. This is my, my, my first name at predictable revenue. com and I'll, I'll do another video follow up or maybe some companion content with some screenshots or maybe a workflow. Tyler Phillips: I'd love to see a demo from you. Collin Stewart: Yeah, absolutely. I've got something coming. I've been testing out a couple of new tools that kind of do walkthroughs and things. So I'm Collin Stewart: trying to walk the line between, I use a clay and Apollo all the time. And I want to be clear that like Apollo isn't paying us for this. There's

no like benefit. I'm not even, I'm not even putting our referral link in here because I always forget to put a referral link in places. Maybe this time I'll remember and it'll go in the show notes. So if I did remember. Click on that and maybe we get some money. That's not like a significant source of revenue for us. I just like to use the access I have through the podcast to learn things from people like Tyler, because most folks you're like, if a thousand sales leaders reached out to you, you're not gonna be able to. You know, have this conversation with everybody. And so that's what I wanted to use this geeky sales conversation to get into the details so that I could share a little bit more about some of the cool things that we enjoy using with Paul. Awesome. I really appreciate you coming on the show. If people want to get in touch, beg you for features, if they're going to be real kind and send you flowers and ask you to, you know, build something for them. What's the best way for them to get in touch? Tyler Phillips: Yeah. I mean, feel free to reach out to me and connect with me on LinkedIn. It's probably the best. Yeah, it's probably the best to reach there. Collin Stewart: Fantastic. And if anybody's looking to get in touch with