How to Write Custom Messages Based on Intent Signals

0:00

All right, so in this video, I'm going to show you how to write custom email/

0:03

linkedin messaging

0:05

lines to the people that visited your website based on certain intent signals.

0:10

You can do this by using a tool called clay, which directly integrates with RB2

0:13

B, allowing

0:13

you to automatically write these first lines, automatically qualify them for

0:16

these intent

0:17

signals, and send those first lines to an email tool.

0:20

So in this video, I'm just going to show you how to qualify based on those

0:22

intent signals

0:23

and how to add them to a clay table automatically.

0:26

So how do you connect clay in RB2B?

0:28

So if you go to RB2B itself, you can actually go down this integrations column,

0:31

and you'll

0:32

notice that there are a bunch of them, but of course we are looking for clay.

0:35

If you click on it, it's asking for a webhook URL.

0:38

So in order to get that webhook URL, we actually go over to clay and click

0:42

Create New, and

0:43

we can go to Table and then scroll down to import data from webhook.

0:47

And once we click New Blank Table, it will actually give us a webhook URL that

0:51

we can

0:51

then copy, go back to RB2B and paste right here.

0:56

And that's all you have to do.

0:57

Now all of your website visitors will automatically stream into your clay table

1:00

very, very easily.

1:02

From there, we can actually go into the table that I've already built out

1:05

because of course

1:06

I've had my RB2B running for a little while, and it's been able to populate

1:09

some people

1:10

into my clay table.

1:12

So the way in which this flow works is that it's looking to see if the people

1:17

that reached

1:19

or visited my website qualify for any of the intent signals that I want to

1:23

qualify by.

1:24

So here are the lists of the intent signals I am looking for from these

1:28

prospects.

1:28

The first one is that whether they are hiring for a sales position at their

1:32

company, the

1:33

second is whether the person that visited my website was recently hired for

1:37

their role.

1:38

The third is whether the company that the person works for recently posted

1:42

about AI,

1:42

and of course we have a backup just in case nobody matches one of those intent

1:47

signals.

1:48

So we have a first line right now for each one of these.

1:50

This one is saw company names open sales roles and figured you were looking for

1:53

ways

1:53

to scale your outbound.

1:54

If they were just hired for the role, congrats on the new role, figure that as

1:57

company names

1:57

new title, you're looking for a way to increase top of funnel without spending

1:59

too many resources.

2:01

The third one is saw your post about post details and figured you might be

2:04

interested

2:05

in another cool AI tool.

2:06

And of course the backup is just saying, hey, I came across your LinkedIn, I

2:08

thought you

2:08

might be interested in clay.

2:10

In this case, of course, I decided to do a campaign selling people on clay

2:14

itself.

2:14

So how do we actually find the data to back these intent signals to determine

2:18

whether a

2:19

company actually qualifies?

2:21

Well, you'd have to do a ton of manual research to determine this, but luckily,

2:24

clay is able

2:25

to do this all for you automatically, saving you a ton of time, and of course,

2:28

creating

2:29

an automatic flow that has a website visitor, auto important to clay, and

2:34

completely enrich

2:35

the person to determine whether they fit the intent signal.

2:38

And as you'll see in a moment, write that first line for us without us having

2:41

to do any

2:42

work.

2:43

All right, so let's get into the actual clay table now and I'll walk you

2:45

through the entire

2:46

flow end to end.

2:47

So a lot of this data was already populated by RB to be the first name, the

2:51

last name,

2:51

their email, their title, their company name, and their LinkedIn URL.

2:55

All of this was given to me by RB to be automatically.

2:57

So I didn't even need to do anything to get to this data.

2:59

It just flew into my table right when somebody visited my website.

3:02

However, there are other pieces of data that I might need in order to make

3:05

these qualifications

3:07

to match these intent signals.

3:09

So in order to do that, I always like to do a base level enrichment of the

3:12

person and

3:13

their LinkedIn profile.

3:14

This is pretty much a tool that allows you to find every data point of a person

3:17

's LinkedIn

3:18

profile very quickly.

3:19

All you have to do is click add enrichment and enrich person.

3:24

And you'll see here that if you click on this and throw in their LinkedIn URL,

3:26

you can then

3:27

get data on the person by just clicking on this cell.

3:31

So these enrichments in clay work all pretty similarly, which is they ask you

3:35

for a piece

3:35

of data about a certain person.

3:37

You give them that data and they pop out something new for you, whether it's

3:40

their LinkedIn profile

3:41

data, their email, their phone number, their company's revenue.

3:44

It's really, I mean, the sky's a limit.

3:47

So ultimately, all I had to do was give them my person's LinkedIn profile and

3:50

it found

3:51

me all of this information, their summary, where they used to work, their

3:54

latest experience,

3:56

their headline, their title, all of this.

3:58

Now I have the reason I enriched the person in this case is because I needed to

4:02

find some

4:03

company data to make these qualifications.

4:06

So actually, when I went and found their company's LinkedIn by just going to

4:09

latest experience

4:10

and popping out this URL, I then decided to enrich the company as well.

4:14

So same process, I just typed in enrich company and clicked on this and threw

4:18

in their company's

4:19

LinkedIn URL.

4:20

And you can see here, I have a bunch of company data as well, their founding

4:24

date, the country

4:25

they're based in, their general employee size, their industry.

4:28

Again, it goes on and on.

4:30

The reason I needed to do this was so I could run this next integration.

4:34

So as I mentioned before, I'm looking to see if the person's company posted

4:38

about AI recently

4:39

on LinkedIn.

4:40

That is one of my intent signals and I'm planning on writing this custom first

4:43

line.

4:44

Which that is the case.

4:46

So I went to add enrichment and I just looked up recent posts and you can see

4:49

here, we can

4:50

find the recent posts by a person or a company.

4:53

I threw in the company's LinkedIn profile and you can see here that some posts

4:57

popped up

4:58

for some of the companies, not all of them, but some of them.

5:01

And that's all we needed because again, we're just trying to see the ones that

5:04

have it.

5:05

However, I don't know whether these posts, which can be accessed right here,

5:09

actually

5:10

tell me anything about AI.

5:12

What if these aren't AI based posts, maybe it's posted about something else in

5:16

the org.

5:16

And again, the only thing I'm looking for is whether the company's posts

5:19

mention AI

5:20

and some capacity.

5:22

So the way I can figure this out is actually by using the open AI integrations

5:25

within clay,

5:26

which you can kind of think of as chat GPT being applied to 50, 100, 1000 rows

5:31

at any given

5:32

point and allowing you to prompt at scale.

5:36

So all I have to do there is click Adam Richmond, type in open AI and you will

5:38

see this generate

5:39

text section right here.

5:41

Very easy.

5:42

And when you click on it, you have a prompt just like you would in chat GPT.

5:44

However, the difference is you can say something like, tell me whether and

5:48

instead of putting

5:49

the person's name in, you're actually going to put the entire column in, thus

5:53

allowing

5:54

you to customize your prompt for each one of the rows.

5:58

So if I exit this prompt, you can see that I actually already did this for our

6:01

posts.

6:02

I said pick the LinkedIn post that most relates to AI from the LinkedIn posts

6:06

and print out

6:06

the text.

6:07

I find an AI related post and print no, and I went and ran the integration.

6:11

And you can see here that a lot of the time it couldn't find any AI related

6:14

posts, which

6:15

again is fine, but sometimes it could.

6:17

So it would print out the texts for all of the AI related posts that I found,

6:20

which is

6:21

great.

6:22

So we now know that there are quite a few companies here that actually match

6:25

the criteria

6:26

we're looking for.

6:27

However, we haven't written a first line.

6:29

We need to be able to reference this post in some capacity to make it seem like

6:32

we're actually

6:34

doing our research when reaching out to prospects.

6:37

Luckily, open AI is quite versatile.

6:39

So we are able to use yet another open AI integration to write our congrat

6:43

ulatory sentence

6:44

based on that post.

6:46

I said use the following news article info, which was pretty much the post to

6:49

create a

6:50

congratulatory into intro sentence.

6:53

It should be formatted as congrats on text, keep the output under 10 words.

6:55

And I fed a little example in here.

6:58

Then I just gave the post text from right here into the open AI integration.

7:02

And voila, we now have ourselves a first line based on the post.

7:07

So congrats on closing your first investment round.

7:09

Congrats on the smart cat templates launch.

7:11

Congrats on being a CNBC disruptor 50 company.

7:14

We have an awesome first line to feed into our eventual email or LinkedIn

7:20

message.

7:21

Now I wanted to make sure that I knew which companies qualified for the intent

7:26

signal that

7:27

I was looking for.

7:28

Luckily, clay has this really cool thing called formulas and you can actually

7:33

check boxes

7:33

based specifically on certain attributes.

7:36

Let me show you how to do that.

7:38

So if you click add column and you go to formula, you can type in something

7:41

long lines of like,

7:42

only check the box if.

7:46

And then we can go and find a column.

7:48

So for example, we can say AI post line, this congratulatory line we just wrote

7:53

is not empty.

7:55

And generate formula.

7:58

You'll see here that the ones that have it will say true.

8:01

I can save it.

8:03

And suddenly we have all checked boxes based on whether this first line was

8:09

written.

8:10

As you can probably imagine, we only wanted the first lines written that didn't

8:14

contain

8:15

no, which we were able to do very easily.

8:19

And you can see here that this post check is checked every time we have

8:22

ourselves a first

8:23

line.

8:24

So that is our first intent signal completely done.

8:28

That is also the hardest one of this flow.

8:30

So the next one is whether they started their job recently.

8:34

So I think for this one, I did the past six months.

8:37

Again, I use another formula and just told the formula using plain English.

8:40

If the person start, they was in the last six months and the AI post check is

8:44

not true.

8:45

So they didn't qualify for the previous intent signal and check this box.

8:49

And voila, we have this box checked for everybody whose role started in the

8:54

past six months,

8:55

which will again prompt this line right here, congrats on the new role, et c

8:59

etera, et cetera.

9:01

The next one I'm looking for and the final one I'm looking for is whether the

9:03

company

9:03

is hiring for a sales position.

9:06

As you can imagine, Clay has an integration for this.

9:08

You'd have to click on add enrichment and then type in open jobs.

9:14

And we have job openings right here.

9:16

I ran that integration, found that there were quite a few sales jobs in some of

9:19

these companies.

9:20

And every time there was a sales job found, one, more than one doesn't matter.

9:25

I had the formula check this box.

9:26

So I went to edit column.

9:28

If the job count is over zero, check the box voila.

9:31

So you might be wondering yourself, why do I have all these check boxes,

9:35

especially even

9:36

the backup one, right?

9:37

So if they don't fit these three, we have this fourth one being checked.

9:40

Why do I have these check boxes?

9:41

It seems like kind of overkill.

9:43

Like why am I doing it?

9:44

The reason for that is because I'm having this first line print out based on

9:48

which box is

9:49

checked.

9:50

So again, we have these first lines for each one of our intent signals.

9:53

What I'm doing here is I'm writing a first line based on this intent signal.

9:57

So for first line final, I'm using yet another formula.

10:00

As I mentioned before, these formulas are awesome and you don't even have to

10:03

use code

10:04

or anything.

10:05

You can just use plain English as you'll see in a second.

10:07

So I clicked add formula, which is right down here.

10:11

And we'll ask you to describe the formula you want and in that area, I put if

10:16

AI post

10:17

check is checked, then print this line.

10:19

If the new role check is checked and print this line, if open roles check is

10:22

checked

10:22

and print this line and the backup is checked and print this line.

10:26

So it's pretty much saying depending on the intent signal, write the first line

10:29

that matches

10:30

the intent signal that we found for the person or company.

10:34

From there, you can look at the results here and see some pretty cool stuff.

10:37

So saw extension open sales roles came across your LinkedIn.

10:41

So that's our backup.

10:43

We can go down to saw sure codes open sales roles again.

10:47

Congrats on the smart cat templates launch referring to the AI post again.

10:51

Pretty sweet.

10:52

It has a custom first line based on the intent signals.

10:55

And of course, you can do this with any lines, not just opening lines.

10:58

You can do it with entire emails if you'd like.

10:59

And it's a really cool way of customizing your emails and segmenting people and

11:03

companies

11:03

in a way that really wasn't able to be done before.

11:05

And again, this can be done automatically with all of your website visitors and

11:08

sent into

11:09

a campaign.

11:10

So I hope this was helpful.

11:11

Feel free to reach out to any questions and yeah, appreciate it.

11:13

[BLANK_AUDIO]