How I Generate B2B Leads With AI Agents

0:00I build AI agents that monitor what

0:01people are saying and doing on social

0:03media. They figure out exactly who's

0:05reaching out to who and contact them

0:07with messaging that's insanely relevant

0:09to what they care about, completely

0:11automated. Now, one of these agents

0:13booked three meetings for me in one

0:15single day with zero manual work my end.

0:18And in this video, I'm going to walk

0:20through exactly how I set that up, the

0:22exact tools and workflows I'm using so

0:24that you can start generating this like

0:26clockwork without building a single

0:28static list or doing any manual work

0:31your end. So, let's get to it. So, the

0:33biggest mistake that I see with people

0:35making without bound is they start with

0:37a static list. An example of a static

0:39list is a list that never updates. Once

0:42you create the list, it is set in stone.

0:46So, typically you go to use tools like

0:48Apollo or ZoomInfo, you filter via job

0:51title, experience, uh location, things

0:54like that. But, the big problem with

0:56this is it tells you nothing about

0:57whether someone actually cares about

0:59what they're doing right now. The only

1:00thing that you know is that they hold a

1:02certain title, they work for a company

1:04with a certain size, and so on and so

1:06on. You're essentially guessing that

1:08because someone is the VP of marketing

1:10at a 200-person SaaS company, they might

1:13want to hear from you, which 9 times out

1:15of 10 just isn't the case. And so does

1:17every other company selling to that

1:18exact same person, by the way. And I

1:20think that's why cold outbound is

1:22getting harder and harder because

1:23everyone has the same tools. It is

1:26easier than ever to go get, you know,

1:28the ZoomInfo or Apollo's of the world,

1:30set up email sequences, and every

1:32company is doing the exact same thing.

1:34So, in a world where everything is

1:36highly commoditized like that, it means

1:38that no one really stands out as a

1:41result. And I'm seeing inboxes fail,

1:43they burn quickly because they get

1:45reported as spam. I'm seeing reply rates

1:47plummet. I'm seeing people get rejected

1:49on cold calling because there is just no

1:51relevancy there. And so, this is why I

1:52think social signals has a huge

1:54opportunity to play. So, here's what I

1:57do instead. Now, is I let people tell me

2:00that they're interested through their

2:02own behavior. I call this digital

2:05breadcrumbs. And with digital

2:07breadcrumbs, I'm looking at the social

2:10signals. What I mean by this is, let's

2:12say Joe Blog liked XYZ post three times.

2:15They commented on XYZ post three times.

2:17And they also posted about, you know,

2:19XYZ. I can put all of these things

2:21together, these digital breadcrumbs, to

2:23create a social signal. And there are

2:26two types that matter the most. The

2:27first is what someone is saying, meaning

2:29they're actually posting content about a

2:32topic related to what we do. They might

2:34also be writing about, for us, for

2:37example, social listening, intent data,

2:38go-to-market workflows, or whatever it

2:40is. The point is they went out of their

2:41way to create something about that

2:44topic, which tells you it's top of mind

2:46for them. Second is someone engaging,

2:48right? Meaning they're liking or

2:50commenting on posts from perhaps a

2:51competitor, perhaps a thought leader, or

2:53even your own content. They didn't just

2:56scroll past it, they stopped, interacted

2:58with it, and that is a massive signal in

3:00of itself. No one forces someone to like

3:04or do something on social media. So,

3:05someone's true inhibitions, whether you

3:07love social media or hate it, is their

3:10true inhibitions. And what I mean by

3:11that is you've got to sell on what is

3:14that person interested in as opposed to

3:16what is the company trying to push

3:18because at the end of the day, even if

3:19the company is going towards a certain

3:20direction, if the ICP that you're

3:22selling into wants to go slightly

3:23different, they're going to side with

3:24those providers that are kind of angling

3:26themselves to make that person look

3:28good. As a result, we're stopping

3:29guessing anymore. They've already shown

3:32through their behavior that they're

3:33interested, and that changes everything

3:35about how the outreach actually performs

3:38and the responses that we're going to

3:39get, which I'm going to get to you later

3:40in this video. So, the tool that I use

3:42to capture all of this engagement and

3:44social activity is Triggerfy, which is

3:46my company's product. I built it

3:48specifically to solve this problem.

3:49Short little bit of history, but

3:51Triggerfy used to be an all signal

3:53platform. And then I realized that

3:55social signals were absolute key for us

3:58to focus on because the results were

4:00night and day versus any other intent or

4:03any other signal. So, here we are. We're

4:05inside of the Triggerfy platform. Now,

4:07if we think about the use cases with

4:09Triggerfy, you can monitor anyone who is

4:12posting about XYZ, and that is across

4:15multiple different social media sites,

4:18right? It doesn't just start with

4:19LinkedIn, albeit that is a very, very

4:22popular and a good one. We're also

4:23looking at things across X and Reddit

4:25and YouTube and podcasts and Instagram,

4:27and I think we're just about to add

4:29Substack and TikTok as well as GitHub

4:31and Stack Overflow. So, it's like a

4:32really rich, diverse area of data. And

4:35I'm really actually seeing Twitter and X

4:36gaming steam at the moment. But, let's

4:38stick to LinkedIn. So, how am I using

4:41it? Well, I'm basically searching for

4:44anytime someone is talking about a level

4:48of intent, right? So, an example would

4:50be anyone talking about social signals

4:54or social signal linked to maybe like

4:57GTM. GTM, let's do that. So, here we go.

5:00So, we have now these people who are

5:02posting about social signals. Very good.

5:06They're actually peddling a competitor.

5:08So, what I might want to do is this

5:10person it there's two signals in this,

5:11right? One, this person could be

5:13interesting to reach out to if they're

5:14ICP. And two, the individuals who are

5:17engaging with this competitor might also

5:19be interesting. Here's another one. So,

5:21they're peddling social signals again as

5:23like a third-party signal to to use. And

5:27again, there's six people that are

5:28engaging with this, 18 people engaging

5:30with this, plus all of these people,

5:32right? We can see all of this data here

5:33who are engaging with this content. And

5:37there we go again, a massive one here,

5:39157 likes, 65 comments. That's going to

5:42get us a lot of data inside of itself.

5:44So, all I have to do then is set up and

5:46save this search, and this is going to

5:48run then every single day finding anyone

5:50who is talking about high-intent-based

5:52topics. From there, what we can do is we

5:54can go and actually monitor and build a

5:57workflow to pull this engagement out.

5:59Okay. So, now we're having a look at the

6:01actual workflows that I'm building out

6:02of the back of this. So, let me break

6:04this down step by step. Earlier, we just

6:06created a search finding anyone talking

6:08about high-level-intent-based

6:11searches. Now, one thing actually that I

6:12didn't show, which do you want me to go

6:14straight back and show you quickly, is

6:15we can also set up searches to monitor

6:19our competitors and our thought leaders,

6:21right? So, if I search this search, I

6:22can input any company or profile URL

6:25into here and pull any of their posts

6:27that they're doing. Now, why would we do

6:29that? Well, let's go back to the

6:30workflows as well. So, the use case thus

6:32far, to reiterate, is we're tracking

6:34intent-based searches, we're monitoring

6:35our own brand, we're tracking thought

6:37leaders, and we're tracking competitors.

6:39Now, this is the workflow that I have

6:41set up off the back of that. First of

6:43all, I'm using an agent if I'm running

6:45an intent-based search. So, for example,

6:47the social signal one that we did

6:48earlier. This agent is going to review

6:51that post and ensure that it's actually

6:52talking about social signals specific to

6:55GTM. Boolean gets you so far, we need to

6:57use AI for the last little bit leg just

6:59to qualify if it is. If it is, what I

7:01then want to do is I want to go ahead

7:04and I want to then actually start to

7:05collect the engagement. So, if the post

7:08has more than 10 likes, I want to go

7:11ahead and start to pull all of those

7:13likes on that social media post before I

7:16enrich them, and then I can send them

7:18through to HeyReach or Instantly or

7:20whatever tool that you're using, right?

7:22What I'm also doing over here is I'm

7:23wanting to let the engagement cook. So,

7:26if the post has got less than 10

7:28engagements, maybe I picked it up you

7:30know, straight away, right? And I want

7:31to wait and let that post get more

7:33steam, then I'm going to wait 24 hours

7:36before I go ahead and get the engagement

7:37from that post, and I'll do the same

7:39thing. I'll then enrich those

7:41individuals and send them through to a

7:43campaign. Now, that doesn't just run on

7:46intent-based searches, that's running

7:48anytime someone's engaging with my

7:49competitor, that's running anytime

7:51someone's engaging with thought leaders,

7:54that's running anytime someone engages

7:56with my brand. And this is what I mean

7:58when I say unlimited leads. The signals

8:00never stop flowing because people are

8:03always posting and always engaging. In

8:06B2B SaaS, people are posting and

8:08engaging on social media every single

8:10day. So, the volume of the signal coming

8:12in is consistent. Sure, some days it

8:14might be massively high for a big post,

8:16and some days it might be a little bit

8:18less, but it's an evergreen campaign

8:20that we're running. So, I don't need to

8:21go and build a new list every single

8:23month or every single day or buy another

8:26database. The system essentially is

8:28feeding itself, and the leads keep

8:30coming in as long as the conversions are

8:33happening. By the way, if you want to

8:34check out Triggerfy, it's totally free.

8:36It's a pay-as-you-go plan. I'll leave

8:38the link in the description below, but

8:40let's keep going. So, how do these

8:42agents qualify and enrich the leads

8:44automatically? So, now I've got the

8:46signals, but the raw signals aren't just

8:48leads yet, right? Just because someone

8:50is posting about intent data, that

8:52doesn't mean that they're my ideal

8:53customer. They might just be

8:55name-dropping a topic without actually

8:57caring about it. I need to make sure

8:58that they're ICP, and I need to make

8:59sure that the people engaging with that

9:01content are ICP. And this is essentially

9:05where I built this entire flow, which is

9:07running this particular process. Anytime

9:09I'm enriching an individual, I can

9:11actually start to filter out these

9:13individuals based on job title, based on

9:15company size. That's when I'm using what

9:17I call static parameters to actually

9:19work out if these individuals are more

9:22criteria- or qualification-based fits,

9:25which is typically how people are

9:26building their list generally, right?

9:28And this entire flow that we have,

9:29right? It's like a 15-16 step node, is

9:32everything that I need. I don't need an

9:34A10, I don't need an additional workflow

9:37because this qualifies everything inside

9:40of itself. I could also set up a webhook

9:42to see if they ever reply back inside of

9:43HeyReach, and if they don't, I then go

9:46and build or send them through to

9:47Instantly. So, I can go over email and

9:49then into Instantly as well. So, let's

9:51break this flow down into a little bit

9:54more detail. The first step that we have

9:56is an agent step. So in this one, I'm

9:58monitoring are people talking about

9:59Trigger Fire or triggerfire.io, right?

10:01So this is where I'm monitoring my own

10:02brand and I want to pull the engagement.

10:04Now, I'm using just 5 mini. You can use

10:07any model that you would want to inside

10:09of here, right? So you can go use any

10:10model from Anthropic, any model from an

10:11OpenAI or Gemini. And you can also add

10:14your own API key in here to run these

10:16agent steps. These agents have the

10:18ability to search the web as well. So

10:19you can get them to do research should

10:21you wish uh that's happen. Now, the

10:23output formula for this agent is a

10:25boolean, which basically then is hooked

10:27up to this if node that I have here.

10:29From there, I'm then going and pulling

10:31all of the uh engagement of that post is

10:33getting if the likes are greater than

10:3610. If they're not, I'm waiting 24

10:39hours, I'm letting it cook, and then I'm

10:40going ahead and getting all of that

10:42engagement from there as well. Now, the

10:44next step in this process is if you

10:47wanted to add more qualification because

10:48they're engaging with my brand on this

10:49one, I don't I like to reach out to

10:51anyone who is engaging with me. But

10:53let's say you were tracking a thought

10:54leader, what a lot of people are doing

10:56is they're using this enrichment node.

10:58This gets you their comprehensive

10:59LinkedIn profile, right? Company size,

11:01domain, website, their experience,

11:03everything that you would want to then

11:04run again another if node inside of here

11:07to check to see if they're ICP. Or we

11:09could use an agent node on low model to

11:11check to see if they're senior within a

11:13certain kind of department. So, senior

11:15marketing leaders for example. If they

11:17are, then we're going to send them

11:19through to a HeyReach campaign, then

11:20we're going to send them through to a

11:22SmartLead campaign or Instantly or

11:24whatever email system that you're using.

11:26So, what happens from here? Well, when

11:27they go into a sequence of some sort,

11:29whether it's email or LinkedIn

11:31automation, we're not using generic

11:34crappy content. We're not saying, "Hey,

11:36I noticed that you engaged with X." That

11:38comes across as a little bit stalkery,

11:41right? We want to focus on the actual

11:44topic that they're engaging with. So, a

11:46very common thing that I'm doing within

11:47these workflows is I'm using an agent

11:49node to also classify what the post is

11:51about. Give me the actual topic. So,

11:55yes, it might be a competitor posting,

11:57but tell me what specifically they're

11:58posting about. For example, in our use

12:00case, it could be something around

12:02social signals. That's the topic of the

12:04post. And so, in this case, we want to

12:06say, "Hey, I can see that you're really

12:08interested in social signals." And then

12:10you build value off the back of that.

12:12How I typically do that is I have a lead

12:14magnet associated to the topic that

12:16they're engaging with, and I share that

12:18lead magnet. Or I ask them, "Do we want

12:20to see this kind of three-pager that I

12:22built on how you can leverage social

12:24signals?" Typically, everyone comes back

12:26saying yes because you know that they're

12:28interested, and therefore why wouldn't

12:29they want to see something which goes

12:31even further than the post that they

12:33engaged with. And then, obviously,

12:35you're subtly selling throughout that as

12:37well. What this means is the campaign

12:39messaging is tailored to intent data cuz

12:42that's this entire flow is built around.

12:44So, these people we're already thinking

12:46about the topic, and now they're hearing

12:48from someone who can actually help them

12:50with it, right? Like that's the whole

12:51idea. I told you guys at the start of

12:52this video, but this agent was live for

12:551 day and booked me three meetings

12:57completely on autopilot without any

12:59manual prospecting, list building, or

13:02copying pasting messages, which is

13:03incredibly high leverage. And this is

13:05just one agent I built around one

13:07particular use case, tracking

13:09high-intent social listening searches. I

13:12have 10 different social listening

13:13searches, 10 different competitors, and

13:16I'm also monitoring my own brand. So,

13:17you do the math on the amount of

13:18meetings that we're booking. Now, if you

13:20want to get started building something

13:22like this, the first step is getting the

13:24signal creator that makes the whole

13:26thing work. So, I'll leave a link to

13:27Trigger Fire at the top of the

13:28description. You can start with 250 free

13:30credits and set up your first social

13:33listening search today. Again, it'll be

13:34at the top link in the description. So,

13:37go click it. And if you want to see how

13:38I book over 100 sales calls per month

13:40using AI, then watch this video next.