GTM Infrastructure for the Agentic Era

Julia Nimchinski:

Welcome to the show, Daniel Sachs, CEO and co-founder of Lendbase. What a treat, how are you doing, and what’s new in the Lendbase world?

Daniel Saks:

Julia, thank you so much, and really great to be here. Love speaking in front of, groups of people who are excited about go-to-market automation and AI, so really excited to take you through it. I’m gonna first, let me share here… I’m gonna first share a quick presentation, and then we’ll go through some live scenarios.

I’m also joined by Max Wu, who leads, GTM Engineering and, and sales development on our side. It can be, like, the live example of what we’re doing.

But just by way of background, Landbase was founded with the vision of what GTM looks like in the agentic era, and we wanted to create the infrastructure really leading with data that can allow you to power your go-to-market engine. And we started a couple years ago by really founding the company as the first AI lab for go-to-market.

And our focus has been to look at the ways machine learning, AI, and different, agents, tools, and data can be most unlocked in this new era of Agentic AI. My background, was the founder and co-CEO of AppDirect, which is approaching a billion in revenue. We’re the largest commerce platform for selling tech, raised over half a billion in equity.

And are the largest, resellers of tools like Microsoft 365, AWS, AT&T, Telecom, and others. And my co-founder, Hua Gao, is a Stanford PhD in neural networks, who started the first machine learning, startup called EverString that was bought by ZoomInfo, and ran Gen AI at ZoomInfo most recently.

And, just a call out, we do want to offer free credits everyone here, so feel free to scan the QR code. I’ll also show this at the end, but this will give you a link to sign up and free credits. So if you want, you can follow along as we go, and we can help, really, guide you through this process.

So, what we’re starting with here is from, from the onset of Landbase, what we did is we partnered with agencies to be able to get a huge set, like tens of millions of, campaigns. And we did an analysis of, like, what drives success, for go-to-market across different types of channels.

And what we realized is the most important thing for driving success, particularly in outbound but also inbound campaigns, is the value proposition and the trust of your company brand or your individual brand.

But far and away, once you have that, or limiting that, the next most important thing is starting from a base where you have a very targeted list.

So therefore, you can have an accurate group of people that you’re identifying, and that increases the chance that you’re gonna have better deliverability rates, lower spam, higher conversion rates, higher connect rates.

And what we found in terms of the type of campaigns that need to be sent, or the lists that need to be built, they’re not defined by traditional filters, so… SDRs for the last, you know, 20 years have been in an interface, looking at, filtering by saying, you know, information technology industry across, city, with X number of employees.

But that tends to not be the way to get a really high, accurate list. you want to actually allow agents to reason over the data to create very customized lists with customized messaging. And that’s exactly what we’re doing with Landbase.

And we’ve processed the world’s information around, businesses, companies, contacts, and we can enable, agents natively in Cloud Code and Codex to reason over that to execute omnichannel campaigns that have far better connect rates and deliverability.

So, really, the key finding, that would be the reality for any of you, in the GTM, in the Agentic era, is how do you get a better list that’s more accurate, both with contact info, but also the caliber of who you’re targeting.

And how do you make sure that, when you start a campaign, whether it’s Dial2Connect’s campaign, or whether it’s an email campaign or LinkedIn campaign.

You’re starting from the basis of really valid, relevant value prop, so that way, the momentum of the campaign grows, because, you’re not getting flagged as spam, you’re having better open rates, you’re having better connect rates.

But in order to do this behind the scenes, Landbase has had to process all this information in a gentic way, so agents can read over the… reason over the data, and therefore, all you have to do is chat in natural language to execute complex workflows And the AI, does the work for you.

So what we found is, like, the… state of the world today, is that you either need to have a human, essentially filtering, in ZoomInfo or Apollo, or you need a GTM engineer in Clay, which consumes a lot of credits and is you know, there’s only so many GTM engineers in the world.

Or you need to, you know, if you’re very sophisticated, you would probably have a data scientist writing SQL queries. So what we’ve essentially been able to do is, automate all these flows into natural language, where you can judge the output of the list by, do you see better conversion rates and better connect rates instantly?

And we, were only recently launching this product, so you guys are early to the test. But over the last few weeks of testing this with B2B influencers. We’ve had people who have seen 2 to 3X the connect rates on calling campaigns, which really means that they’re 2-3x more productive.

We’ve seen much higher connect rates on email campaigns, and these are the type of the outcomes that we want to drive for you. So the key, as well, is that you’re spending less time on list building, you’re spending more time on, you know, executing, and you’re getting better results from those campaigns.

The key thing that we’ve seen, so we do have a web interface, but we’ve seen that once you put our data into Cloud Code or Codex, it can reason over, our data, but also have more context on you to be able to do much more complex workflows.

And our AI and command line interface is also designed in a way that, we can… take the judgment of knowing the trade-off between accuracy, precision, cost, credits, etc. So, by leveraging our AI, or Command Line Interface in Cloud Code. It knows a lot about you if you’ve integrated different solutions to it.

And conversely, it also can make decisions based on the network that we have to optimize for things like cost and accuracy, which are ultimately going to lead to far better outcomes for you across your campaigns. So with that, I’m gonna just quickly, share, the interface.

So… As I mentioned, everything can be accessed and observed through a web-based interface, so if you want, you can define something. So you could say, I’m a salesperson for Brita water filter. And I’m… Visiting… Colorado. who should… I… I… Should I visit, and why?

So these are the types of questions that you can ask landbase, and because we can reason, we can, you know, manage this. You could do this in our web interface, or you could do this in… Claude Code. Obviously Claude Code will have more, current awareness of you, but, that’s, you know, the way it works.

So this is interpreted already, like, okay, here are the different types of people, here are the cities you can go to, and it will help you narrow down some of the prioritization. So, let’s go 1, 2… As this is running, what I can show is, like, you can also… And see all. you also can use our web interface as a, visibility, layer.

So one of the pitfalls of working only in Cloud Code natively is you can’t, you know, easily view and observe things in a spreadsheet format. So you can leverage our web interface to do that alongside Cloud Code. And then, what you can also do is start creating, really advanced workflows, native in Cloud Code.

So I can do things like, find… lookalikes of, of a different, company, like, you could put in, so what I did here is actually I asked it. who, based on what you know about me from, like, HubSpot data and other data, should I be targeting? You can then narrow down these people that you want.

You can then have, like, Landbase automatically expand this to find lookalike companies. It then can create, like, a top list of who you should go focus on based on, who you’ve won at lately, and it helps you narrow down, your ICP. So there’s a lot of complex workflows.

I’m going to turn it over to Max for a second to share one more, and then we can open it up to questions.

Max Woo:

Yeah, so back to what Dan was saying, our founding premise is to really enable non-technical go-to-market operators, so regular sales guys like me, to do GTM engineering directly in your AI. So behind the hood, we have over 230 million contacts and 1,500 unique data attributes to build Tier 1 lists. So, for example.

Let’s say I’m building a list for a lead generation agency. You can basically ask your AI to find me a list of 3PLs, which are third-party logistics companies that have 3 AEs, and no SDRs, and are hiring for 2 SDRs at the same time. Then, image using our waterfall system of 20 vetted providers to get 15% connect rating on average.

So, with that said, I’m going to take a quick 2-3 minutes to show you, how I went from having zero clue who a company sells to, what they do. to producing a lead list for their fundraising efforts with a full TAM app and personalized message.

So yeah, end-to-end from research, list building, personalized messages, to TAM mapping, all within Cloud Code. So let me share my screen here. Perfect. So, hopefully you can see my screen. Essentially, I just prompted Cloud Code to say I work for Syncuria.ai, and they are a document processing platform.

Yeah, to tell me their ideal customer profile and signals I can use to build Tier 1 list using a land-based CLI. So Claude just did some research as to what Syncura sells, CDP, CPA, ICP, we’re able to pick up the fact that they sell to COOs, chief automation officer, the champion is typically the head of intelligent automation.

And then from there, our special software. So these are signals that, only landbase would have. So for instance, aggregate actual software spend, specifically in the AP automation category, we’re able to pick up the fact that Bill.com actually makes up 77% of their mid-market, market share.

AI adoption index by sector, so obviously technology companies would have the highest, tech adoption. Job posting signals, we can uncover conference attendees months in advance. We can get tough-to-find data attributes, like a sales team size, SDR count, as I’ve mentioned. How distributed their team is, recent funding, M&A activity, and so on.

Now, we somehow changed the conversation into finding a list of VCs and Angel Seneca networks that specialize in investing in companies like Syncura, because they’re trying to raise money. So then, we’re able to pick up, the fact that they were founded in 2024, based in Toronto, and we have tiered up different VC firms.

terms, depending on how relevant they are. So, for instance, these are vertical AI seed specialists, meaning they only, invest in vertical AI companies in the C stage, prioritizing Toronto-based VCs. Tier B are early backers of direct IDP automation comparables.

And I asked it to actually tier this list up with scoring, and it came up with a scoring rubric based on thesis, stage, vertical, and geofit, and track record, so each lead gets a score of 100.

And then I ask for personalized emails and personalized LinkedIn messages, so if I show you the final output in a CSV, obviously you can also view that within our web interface.

But you can see the investor here, it’s been tiered up, the type of, you know, investor they are, stage focus, and they’ve got all of the, scoring, Results right here, so each lead gets a score of 100. And if we take a look at the email, you can see that it’s been personalized.

Hi, first name, Radical has been the clearest conviction bet on applied AI to come out of Canada, so as a Toronto team building exactly that, you are our first call. Second one’s a bit different, so does AI first from day one thesis is essentially our founding premise.

Skincare isn’t a SaaS tool with AI bolted on, the reasoning layer is the product. And you can see the LinkedIn message here as well, and then we have also enriched them with the appropriate contacts, which are the founders and managing directors, mobile phone number, email, and so on.

And I also did a full TAM app, so, if I scroll down a little bit here… Yeah, going back to the list of VCs, build me a TAM estimate and segmentation for VC and Angel Seneca networks that would be interested in investing in companies like Skira. And you can see that there’s a nice graphic to go from TAM, SAM, to SOM.

And then from there, we’re also able to see within the SAM, the different segments. So there’s institutional VCs, there’s 300, A16C is one of them, micro VCs, followed by angel syndicates, and so on. So, yeah. That’s, what I wanted to go through for today.

Daniel Saks:

Thanks so much, Max. Julie, any questions?

Julia Nimchinski:

Yeah, phenomenal demo. Thank you so much, Max and Daniel. One question is, you showcased a little bit how its land base is different compared to Apollo, ZoomInfo, and the like, but folks are asking, like, are there any other criterias? or use cases that are super unique compared to Place, XN’s Demandbase.

Daniel Saks:

Definitely, I mean, I think the biggest difference is that if you’re using a database like ZoomInfo or Apollo, you need to have someone filter over it, or if you bring it in via MCP, the agents can’t reason over it, so you can ask certain things, like, who’s the CEO of a company, and normally they’ll be fake CEOs, or you could ask things like, yeah, who should be my ICP, and maybe Claude Code would answer, but it wouldn’t be from the intelligence of the data provider.

Whereas we’ve really focused on creating data that can be reasoned upon based on knowing what campaigns execute. And really, the secret sauce is our model, GTM Omni, which has been trained on millions of campaigns and helped identify patterns to be able to have better conversions.

Julia Nimchinski:

Thank you, Danielle. And one more question, tricky one, and we’re transitioning to the next demo. So, what does the AI do when the data could be incomplete, conflicting, wrong?

Daniel Saks:

Yeah, so data is naturally imperfect, so the key is that when we reason over data, we try to understand, you know, where are there coverage gaps, versus, where are there, like, creative ways to find similar signal.

So some people may ask, oftentimes people will ask us things that are not possible, or the data’s not available anywhere in private databases or online, but we can try to use proxy signals.

So again, because our models have been trained on so much data, they can identify a type of signal that could act as an alternative to a world where we don’t have real data, and then be able to suggest a solution For a signal-based campaign.

Julia Nimchinski:

Love it. We’ll share the slide you shared in the beginning in our Slack, and yeah, follow-up sequence. Thank you so much.