In this episode of the Pipelineology podcast, host Gary welcomes AI pioneer Jeremy Schiff, who shares his extensive background in AI and machine learning. Schiff discusses his journey from developing an AI-enhanced photo editor and a personalized restaurant recommendation system to founding RecruitBot and Salesbot. He emphasizes the importance of integrating AI in sales to increase productivity and accuracy, addressing the resistance and enthusiasm toward AI in the industry. Schiff highlights Salesbot's advanced AI capabilities for improving data quality and enhancing outbound sales strategies. He also explains how Salesbot's unique data pipeline and enrichment waterfalls ensure high-quality, comprehensive, and up-to-date contact information, saving sales teams significant time and effort. Listeners are encouraged to explore Salesbot.io and try the platform's features through a free trial.

00:00 Introduction and Guest Welcome

00:29 Jeremy Schiff's Background and Career Journey

01:52 Introduction to Salesbot

02:46 AI in Sales: Opportunities and Challenges

06:24 Salesbot's AI-Driven Search and Data Quality

12:58 Salesbot's Unique Data Pipeline and Enrichment

24:09 How to Get Started with Salesbot

25:31 Conclusion and Contact Information

[00:00:25] Jeremy Schiff: Thanks so much. It's wonderful to be here.

[00:00:39] Jeremy Schiff: Yeah, I'd love to. Background on myself.

I've been, as you've said before, I've been doing AI for, it's funny to say almost yeah, it's been, renamed three or four times since I've started this stuff. Started out doing my PhD in robotics and AI at Berkeley. My first company was like, an AI enhanced photo editor that powered companies like MySpace and Photobucket, eHarmony and was the number one photo editing app on Facebook.

It grew to about 50 million people using it a month. My second company was personalized restaurant recommendations. So think of Netflix's AI algorithm, applied to sort of Yelp data. Sold that one to OpenTable. Helped basically rebuild their search engine and solve, worked on a bunch of AI problems while I was there.

And then I was like, what happens if I can go and, recommend people instead of recommending places? So we started out actually in the recruiting space with RecruitBot. But then, we've actually pivoted into Salesbot and I've found. It's really fun, and it's really fun to sell to salespeople.

Lots of high energy, lots of people willing to try new stuff. And it's been really, fun sort of building out Salesbot. And, I guess some quick context about Salesbot. I'm sure we'll dig into it in more detail, but basically, Salesbot is what we call an AI first version of an outbound platform.

So imagine reimagining all of the AI technologies like Chat GT and that sort of stuff. But how do I reimagine workflows and data around that. So our database is three times bigger than Apollo. Double the size of ZoomInfo. It's more comprehensive, we have much higher quality contact information. And then AI will do things like automatically generating your search, or automatically generating an email campaign and personalizing it.

Based on all of that sort of, meaty data we have in our database to help salespeople spend less time chasing and more time in meetings.

I don't, I don't wanna touch it. I don't want it to do that. It's, I don't want it taking my job. And then there's the other one, who says, just give me all the AI, I wanna automate everything and I don't, I, I'm like that guy who takes out the billboards that say don't hire employees, just hire AI. I, I fired

[00:03:09] Gary Ruplinger: Or like the LinkedIn influencers who are, I fired my entire team and replaced them with AI. Where, so I know there's those two camps. I'd love to get your take on where a lot of this is going because I talked to a lot of business owners and they think, are we still, especially for those of us in the digital space, are we still gonna be here in a year or two?

But if you're putting your head in the sand and you're like, I'm not gonna spend any time learning about AI, it's saying, I'm gonna put my head in the sand and not learn about how to use a power dialer or to use LinkedIn or to use email automation, right? It's the sort of cliche at the moment. Which I, I said many years ago, but now it's become, I feel like a trope that a lot of people just say is you're not gonna lose. AI isn't going to take your job, but someone using AI is going to take your job. And I very strongly believe that is true. So if you're not experimenting and adopting some of these technologies, I think you've got a problem.

That said, I feel like the pendulum is swinging in a bunch of different directions, right? So there's the sort of companies that are sprinkling AI on top of their existing sort of workflows. I, we call these like the outbound 1.0 companies. Companies like ZoomInfo. Then there's companies that are trying to completely replace salespeople with AI.

And I think there are use cases for hyper transactional sales where that's gonna make sense. But I think for the majority of the market for like mid-market or selling to mid-market or enterprise or even any sort of slightly more complicated SMB sale, salespeople are gonna want more control over what these sort of platforms are doing, right? And so you, there's all sorts of advantages of humans and there's all sorts of advantages of AI and you want to lean into the strengths of both. AI is gonna take a very long time in terms of being able to build a sense of connection and a relationship, which we all know is essential for closing any sort of significant deal, right?

You have to build trust, you have to build like mutual respect., and you have to come from a place of authority, and it's very hard to do that if it's some AI chat bot that's just like talking to you. But why would I go and spend hours of time doing research about a prospect when I can have AI do that for me in a couple of seconds and then I can go and tweak that messaging to be more relevant to the individual?

Or why would I go and spend a ton of time iterating through search fields to go and try and find the right people in my market when AI can get me 90% of the way there in just a couple of seconds? At least Salesbot's view is we want, we're looking to make salespeople way more productive and ultimately make them a lot more money.

That's a, that's a sort of a mission that like other people can get behind and one that we're excited to bring forth into the world.

Yeah. Building lists and exporting them and clicking page after page and meticulously making my negative keyword searches and block, filtering all this stuff out where

[00:06:45] Gary Ruplinger: Sometimes it's been hours.

[00:06:47] Gary Ruplinger: I'm really curious, how your AI kind of first architecture improves upon this or how you guys have done this.

I know it's, it's one of those things where probably it's one of the bigger time sucks out of my day and it's been impossible to outsource it to other people as well. So I'm curious how you guys approach it.

So what's actually interesting is, so like lots of people have different approaches, but our approach again leans into this- how do we do the yin and the yang of the human and the AI? So you literally describe in natural language who your ICP is. So I'm looking for people in marketing leadership in Austin, Texas, in the construction industry with, that are in the enterprise in Austin, Texas, or something like that, like pretty- it can be as detailed or as vague as you want it to be, and then you give a link to your website, or you just give a blurb about the thing that you're trying to sell, and the AI can go and merge all of that information together and generate a search for you. So it's probably gonna be about 95%, right?

But there's, to your point, there's probably a couple of negative tokens or whatever we need to do, and the platform's flexible enough to go and do that, but it saves a huge amount of time. And the reason it's been so interesting is the way the search engine works is a little bit like a human actually.

So it runs a search and then it looks at the number of results and reviews, the first couple of results. And then it says, am I in the, am I in the ballpark of what I think I'm looking for? If I am, let the, let people go and see the search and go and modify it. If I'm not, look at what's wrong, tell the AI to go and correct it. Change the search tokens and do it again. So it's funny, sometimes people will run a type in a search and it'll, we'll give you an answer in like, a couple of seconds and others it'll take almost like half a minute or more. And it's because it's actually like redoing the search multiple times the way a human would.

And it ultimately gives you much, much higher quality results. Also a really interesting side effect, which has been a big bonus that we didn't anticipate, but our customers have been really enjoying. It is, to your point, there's always this problem of balance. Where I want lots of people in my TAM, but I don't wanna waste a bunch of time with people who are irrelevant, right? There's like this, I want it broad and narrow at the same time. And that's difficult to reconcile. And one thing that's been really powerful about the AI is it's been really good at generating other people that you might want to talk to that you're not by default gonna go and type in.

So like one of my favorite, like couple favorite examples. 20% of people look for co-founder when they're looking for founders to sell to. There's no difference. A co-founder is just as good to sell to as a founder. Like literally same frigging thing, just a slightly different name for the same thing. Or VP of revenue when you're trying to sell to a VP of sales. So I could give a hundred of these examples. But, so there are all these titles, for example, industries as well, where someone will find one industry and type it in. It turns out there are like six other industries that are also good enough that the sort of person doesn't know to include.

So there are all these examples where we can go and expand your TAM and what this means. And I mean dramatically, so we've been able to see 5 to 10 x expansion of the people that you might wanna talk to. So if I like, conservatively, let's say I can find five times more people in your market. That means that not only all of your salespeople, but all of your competitor's salespeople are only talking to 20% of the market. So 80% of the market is sitting there twiddling their thumbs and hasn't been oversaturated, hasn't bombarded by a thousand phone calls about your product, hasn't heard the same pitch 112 times.

Like we've all been on that call, hey, I do X, Y, or Z, and they're like, I'm sorry, I've already spoken to four, four vendors this week have told me that they do this. Thank you, but no, thank you. It's really nice when you're like, hey I, they're like, you're the, this sounds like a real problem for me and I've never heard of anyone who solves it. Really big difference , right? So, the AI is gonna save you a ton of time and it's gonna expand your TAM so it'll make the outreach that you're doing a lot more effective and efficient than doing it yourself, which has been an interesting side effect.

Whereas the ones flying under the radar have maybe named themselves something a little different. People aren't bothering them because everybody goes for the click, click easy. Okay, done. Let's

[00:11:47] Gary Ruplinger: let's do it.

But then it's, but it's always more fun to talk to someone who's just like this is obviously something that I need and I don't have something that does this today. Let's have a conversation. That's always, the fastest close is always gonna be that sort of meeting.

And I know most people who are doing any type of outreach or cold calling would really prefer not to. They'd like, they want good, they want a good list. So I'm curious how, what your approach to something like that is.

So just to like, get a little bit nerdy for a second. When you're building businesses that are based on AI, you don't wanna build a business where anyone else can just knock off your product in six months, right? We've all seen these oh, there's this cool product that does X, Y, or Z, but in six months it's gonna be twice as easy to go and build that product.

So eventually, all of that's gonna be commoditized and, there's no length long-term defensible business here. Then, for us, we've always thought of our data as this sort of unique moat that allows us to provide value to our customers in a way that's gonna take a long time. And specifically that moat is data quality, like, very exactly what you're saying is the point.

So we've spent many years working on this problem. And it's sort of a very unique one. But, and the reason it's so hard is you want to, again, it's, like a lot of AI problems have these sort of yin yang or like maximizing and minimizing problems at the same time. Like very similar to I wanna search for the right person, but I wanna also broaden my, broaden my approach. Which is I want to merge people who are the same person into the same profile, so I don't have six duplicates of Gary in the system. But at the same time, if there's two Jeremy Schiffs that are different, I don't wanna merge them together and confuse all of the data. And so there's a Jeremy Schiff in New York who's a professor in AI, and about a third of the sites I go to are convinced that he and I are actually the same person.

We've spent a lot of time and energy investing in a next generation data pipeline that takes in many, many of these sort of different sources and merges it together. And that is like one of our sort of core IP. So the reason that our database is so much larger, the reason that our data is so much more fresh, the reason that our data is more comprehensive is because this pipeline makes it very easy to integrate lots of different redundant or complimentary signals about people into a single coherent database that has minimal duplicates, but still has lots of value. So that's on the lead quality side.

Then there's also implicitly in your question, the context quality, right? Like I want, like not only do I want the right lead, I really don't, there's nothing worse than finding I, I've got the exact person that I'm so excited to call and I call them and I get their mother or their roommate or, that person actually switched jobs six months ago, right? It's infuriating. So, we do, because of our outbound 2.0 approach, there are other companies that are, will also allow you to do what's called an enrichment waterfall. So something like Clay is a pretty good example. But, with Clay, it's a pretty complicated, sophisticated product. So, like you're probably hiring a Clay agency to help you go and sort of, go and do the email waterfall or the phone waterfall to go and get the much higher quality data. And the data's higher quality because you're going to vendors that can give you this information in real time, as opposed to whatever like ZoomInfo or whatever other company, captured six months ago that's might, may or may not be relevant. So getting timely data is really important. Salesbot has built an enrichment waterfall into Salesbot with one really big caveat, which is it doesn't feel any different than just saying, reveal the email or reveal the phone number.

So you say, gimme any a phone number and we'll go to the, the best vendor for that specific market. And then we'll go to the second best vendor and the third best vendor. And we'll go through all of them and say, okay, here you go. And so our coverage rates are often up to double. We've seen on phone or connect rates are, like, we've seen, like one of our biggest customers saw, connect rate go from 8% to 12% by moving to us. Most companies accuracy rates, so that is like when I call someone, is the person that I'm hoping to speak to, the person I'm gonna speak to? Almost every other platform is at about 60%, we're over 80%. So you can see that all of this, because all of this stuff compounds. Having a more accurate database, better contact information, the AI that helps you target. All of that stuff comes together and can be really, to like save you a ton of time and spend more, you spend more time in front of people you wanna sell to.

[00:18:26] Jeremy Schiff: Yeah, so our job is basically just to build out really strong tools that allow you to target the right person and then get the contact information.

So we don't care if that's work email or phone. Those are our two most common. We have a couple customers that have now asked for personal email, so we'll do that as well. But we're ambivalent, right? And there are lots of different go to market strategies. Actually, Salesbot, we haven't talked about it yet, but Salesbot not only has like LinkedIns for almost everyone in the system, it also has other social profiles.

So you might want to go and spend 30 seconds on Twitter going and researching them, or poking them on Twitter if they didn't respond to the email that you wrote or something like that, or X whatever you wanna call it now, it's or Facebook or Instagram or what have you. And so all of those extra layers of different channels can be really helpful at getting certain people to respond.

And so, but like plenty of customers are email only. Plenty of customers are phone only. The majority of people are doing multi-channel though, and they're doing email and phone. At a bare minimum.

[00:19:45] Jeremy Schiff: I would agree. I, think our email data is better, but our phone data is much, much better. Like when I'm talking about 50% when, and again, it also all compounds when I can go and find 5 times to 10 times more people. And for each of that cohort of people I can find 80 to 90% of the phone numbers instead of 50% of the phone numbers, and then 12% connect instead of 8% connect. And then, 80% are relevant instead of 60% are relevant. All of a sudden, it's like we're, we are talking about three or four times more meetings in the day because of all of those sort of things compounding on top of each other.

[00:20:50] Jeremy Schiff: Garbage Yeah.

[00:20:56] Jeremy Schiff: I'm actually curious, we should talk afterwards. I'd love you to run that same list through Salesbot. It'll take us like 30 seconds and I can see what sort of coverage difference there is.

[00:21:08] Jeremy Schiff: Let's do it. For sure.

Because I know especially, like some searches like, what was I doing? I was doing one for a new client yesterday in the hospitality space. Which encompasses anything from hotels to restaurants to resorts to golf courses. And I just wanted hotels. And it's like, how on earth do- you know it's like trying to like figure, basically struggle my way through this search of keywords of

[00:22:09] Gary Ruplinger: Now how do you give me just the hotels, I don't need any of the golf courses. And it's does, your, can your AI handle that? Can it actually figure it out?

So take that sort of hotel example. You might go and the AI will probably do something like not only search the industry that you said, but it might type things like, I'm gonna search for hotel or variants of hotel in the company description of what it is that you're searching for. that's not a way that people are typically used to searching, but the AI will know how to do it and will plug in, hey, here's 20 different keywords about hotels into the company description, and go and find you a ton of relevant people that might wanna come up.

And so the challenge that a lot of people have is ultimately your point that the industry search is too general. And I want something that's a little more narrow. And so, there's a bunch of different ways you can do that. But yeah, effectively, you wanna get creative about it and the sort of world's best SDRs already know how to go and do this, but the AI can bottle that sort of intuition and do it for you.

And what's nice is, again, that sort of 90 10% rule where the AI will get most of the company description perfect. And then they'll add one term that you're like, that's not exactly what I want. And then you go and take it out and then you have exactly the search that you want and it took you 10 seconds.

And now you can go and go after just the people who are inside of your ideal customer profile and ignore all of those golf people who are not relevant.

[00:24:17] Jeremy Schiff: Yeah, just come to Salesbot.io. or you can send us an email at sales@Salesbot.io and happy to talk to you. Love to show you that AI search. Love to show you our AI based email generation or personalization, all of that sort of wonderful stuff. And there's a free trial, too. Just so people, people often get nervous about like, how do I try this thing? And so we have a free trial. You can just sign up and try it yourself. And then, for most larger customers, we'll do a paid pilot so you're not locked in for a year until you're convinced that the data is legitimately better. And then we go from there. But, it's been pretty fun. We just put our data where our mouth is and it makes it easy for people to get convinced that this thing is better.

[00:25:07] Jeremy Schiff: Yeah.

So that's very cool. So

[00:25:30] Gary Ruplinger: Very cool. So yes, if you're curious about getting some better data quality for your list, check out Salesbot.io, email Jeremy, or email the sales team and they'll hook you up with, with some info.

Jeremy, really appreciate you coming on today. Thanks for sharing today and have a great rest of your day.

[00:25:50] Gary Ruplinger: Alright, take care. Thank you.