AI Agents + Clay: How I Automate My 5-Person SaaS Company

0:00What is going on guys? I wanted to go

0:02over another deep dive into how I am

0:05scaling my company, Triggery. Now,

0:07Triggery is only five people big. It's

0:10really, really small, but everything

0:12that we're doing is 90% automated, and

0:15the majority of that automations are

0:17being powered by AI agents. Now, today

0:19it's going to be a little bit different.

0:21We're actually going to deep dive into

0:23how I'm using a tool called Clay to

0:26effectively run a load of enrichment uh

0:29which is updating my CRM ATIO and it's

0:33also updating like essentially

0:35everything right and controlling the uh

0:37reach out that we do to individuals and

0:39everything like that. So without further

0:41ado let's dive in right so we are using

0:46uh Clay. play is like a spreadsheet on

0:48steroids. It allows you to integrate a

0:50load of different other tools into it.

0:52Um, complex, not needed for everyone,

0:55but good for what we're trying to do

0:57here. So, this is some of our users over

1:00the past month that have signed up. And

1:04we're receiving this data coming in from

1:05a web hook via a tool called Clerk,

1:08which manages our user authentication.

1:11So the first thing that we're doing is I

1:13actually have an agent that is

1:15essentially going and finding

1:18um the LinkedIn URLs of these

1:22individuals. So it finds the company's

1:24uh uh find the company LinkedIn URL and

1:27it gets the email domain. From there we

1:30then get the company information. We

1:32then go and get the website URL h and

1:34then we get all the information on that

1:37company and then we have another agent

1:39that gives us a summarization on what

1:41that company actually does. So we can

1:44see here who they are cyber security

1:46DDOS migration cloud services things

1:48like that. This is someone who just

1:50signed up quite a large customer. um

1:53we're then getting employee count and

1:54the locations of these uh individuals as

1:58well where we're then going and actually

2:00finding the LinkedIn profile of the user

2:04themselves as well. So this is then

2:06going finding the LinkedIn profile of

2:08the person that signed up so that we can

2:11go and get all of the details on who

2:13that person is i.e. their job title, the

2:16location this particular person is based

2:18in, their LinkedIn URL and everything

2:21that we want from there. What happens

2:23next is we then go and basically get the

2:27uh ID of this particular individual from

2:30Atio because this user will already be

2:32updated in Atio through other

2:34automations that I've built. Now what

2:36I'm then doing is I'm then essentially

2:40running and updating my ATIO to include

2:44all the bits of information around the

2:46company, around the individual because

2:49it's no good having the individual in

2:51Atio because they've signed up for my

2:53platform. I want to know everything

2:54about them, their job title, the

2:55company, the company size, everything

2:57like that because Atio is my uh central

3:00knowledge base and like almost like my

3:02data center, right? So then we're

3:05updating both the individual and the the

3:09the company

3:11itself. Then what we're doing is we're

3:13doing a um an ICP uh check. So here

3:18we're basically saying okay are they a

3:20manager and a a manager level and above

3:24specifically within the growth demand

3:26gen or marketing space or seale

3:29executive where they're based in Europe

3:31or the US. If they are, it triggers this

3:35uh NAT workflow and this NA10 workflow

3:40basically is uh controlling a email

3:43reach out which is going from one of my

3:46co-founders. So you can see here if

3:48they're an ICP fit it then basically

3:50reaches out via one of my co-founders

3:52saying hey like do you want to do a

3:53custom onboarding uh link? Then what

3:58we're doing is we're querying what we

3:59call our um our well we're quering

4:02Stripe essentially to find uh the

4:06customer information. So are they still

4:07on trial? Have they purchased? If they

4:10have purchased, what plans have they

4:12gotten? And then we're adding this

4:14information through to um uh essentially

4:19a uh Google Sheets where we store all

4:22the data. Now, the reason that we're

4:23doing that is we then have another agent

4:26running across the Google Sheets because

4:28you can't run agents across Clay um to

4:31query all of this information to

4:33understand who they are, how they've

4:35been mentioned and so on and so on. And

4:38that is essentially how we are using

4:40Clay to enrich all of our new users

4:42coming on which by proxy is then

4:44enriching our whole CRM data at the same

4:47Bye.