HomeVideos

How I'd Automate An Agency With AI (2026)

Now Playing

How I'd Automate An Agency With AI (2026)

Transcript

752 segments

0:00

Throughout this year, I've worked with

0:01

dozens of agencies to automate their

0:03

business. And in this video, I will map

0:05

all of the 47 workflows and systems that

0:07

I would automate in an agency, whether

0:09

you're a marketing agency, a content

0:11

agency, or a lead genen agency. For each

0:13

system, I'll go into what exactly it

0:15

does and also how to build it, like

0:17

which AI tools to use, that you get an

0:19

idea of all of the workflows that you

0:20

can automate in a service business, and

0:22

then hopefully inspire you to want to

0:24

build some of these automations for your

0:26

own business. And if you build all of

0:27

these, you will automate your business

0:29

almost 100%. If you don't know who I am,

0:31

my name is Moritz. I'm a German engineer

0:32

with over 10 years experience in tech.

0:34

I've built my own software business to

0:35

over 200k in revenue. And now my mission

0:37

is to help transform service businesses

0:39

to run more like software businesses. So

0:41

let's dive in. Okay, now let's go into

0:43

the automation systems that I would

0:46

build for an agency. And I put them into

0:49

several buckets here. First one is

0:52

acquisition, then sales automations,

0:54

fulfillment automations, client

0:56

management and retention, operations and

1:00

team and internal. These are pretty much

1:02

all of the automations I've seen working

1:05

in agencies and that bring a positive

1:07

ROI for your business. Hey, real quick.

1:09

If you want access to this entire list

1:11

of the automation systems, then I've

1:14

also just put that into a notion

1:15

database, you can download it directly

1:17

from the link that I'll put below this

1:19

video and just get it there for free.

1:20

All right, back to the video. Okay, so

1:22

let's start with the acquisition

1:23

systems. And I've again split this into

1:26

content here, paid ads systems, and then

1:29

outreach systems. And so let's start

1:32

with the content systems. Now, the first

1:34

thing that you can do is to build some

1:36

kind of idea scraper. Obviously, when

1:38

you make content, you need the first

1:40

step is always to get ideas for the

1:42

content. And what you can build here is

1:44

some way to monitor YouTube or Instagram

1:47

or other social media platforms for

1:49

topics and scrape those topics and give

1:52

them to you as ideas. The way you would

1:54

do that is if you want to do it through

1:56

YouTube, which is a good way to do it,

1:57

you can do it directly through the

1:58

YouTube API. And if you want to access

2:02

other social media platforms, for

2:03

example, Instagram, I would recommend to

2:05

do it rather through Appify because

2:07

Appify lets you access these platforms a

2:09

little bit easier and you can get

2:11

information out a little bit easier. And

2:13

so you would basically access it this

2:14

way. Pull out the information that you

2:16

need and have some prompt to design so

2:19

that you can generate ideas based on

2:21

what you've pulled out. Okay. The next

2:23

one is once you have the ideas you would

2:25

then want to write maybe you know help

2:28

in the script and postwriting process.

2:31

Now obviously you would just use an AI

2:33

tool to do this. You can define some

2:35

custom skills to to do that. And the

2:38

best way to do that is to typically have

2:40

a lot of like a library of examples of

2:43

how exactly you want your video scripts

2:45

or your posts to be written because this

2:47

way you can reference those examples,

2:50

give it some inputs and then based on

2:52

that combination, it will give you a

2:54

much higher quality output. And what

2:56

I'll say about this is that this

2:58

automation is an example of the human

3:01

approval automation. So what I mentioned

3:03

before which means that you can set up

3:04

this automation and it can maybe get

3:06

your content to like 80 or 90% there

3:09

depending on what type of content but

3:12

usually if you want to make sure that

3:13

the content is very good and not too

3:16

AIish then you will want a human in the

3:19

end to add the remaining 10%. Now, for

3:21

the exact same reason, an automation

3:24

like a content repurposer is actually

3:26

also very useful because AI is very bad

3:29

at coming up with stuff by itself. If

3:32

you just tell it to write a post for a

3:34

social media platform, it will do a

3:35

horrible job. But what it's good at is

3:38

to take something and then just like

3:40

transform or change the format of it,

3:42

which basically content repurposing

3:44

does, right? And so there are several

3:46

different ways that you can do that. For

3:48

example, you can take a long form

3:49

YouTube video and just cut it into

3:52

clips. If you have a certain format of

3:54

your long form video, so for example, if

3:56

you have a podcast format, then what

3:58

works really well is to use something

4:00

like Opus clip or you have that also

4:02

built this functionality basically to

4:04

clip on the long form video. You have

4:06

that functionality also in in Riverside,

4:08

but I think Opus clip is is the leader

4:10

there. So, this is one way to do it. You

4:12

also have a different way which is you

4:14

can take a long form video and kind of

4:16

just look at the contents of that which

4:18

means you would take the transcription

4:20

auto transcribe it and you can use a

4:22

tool called YT the LP for that which is

4:25

hooked into your cloud and then uh when

4:27

it reads the transcription it will then

4:29

understand what this piece of content is

4:31

about and based on that it can then

4:33

suggest topics to extract from that and

4:36

help you write these new pieces of

4:38

content. But the key thing here is that

4:41

the original piece of content is the

4:44

content that gives the context for

4:46

everything. So the AI doesn't come up

4:47

with some random stuff. It just uses

4:50

what you've already produced. So

4:52

therefore, it's a content repurposer

4:53

automation. And then for the actual

4:56

scheduling of the content, you can use

4:58

something like buffer. Connect that to

4:59

your cloud through MCP and this way you

5:02

can easily schedule out your content.

5:04

Now another automation that you can

5:05

build is to create lead magnets in an

5:07

automatic way. You just like have also

5:10

feed it the topic of your content or you

5:12

know whatever context you need to create

5:14

the lead magnet. Have the lead magnet

5:16

created. Let's say you're creating a

5:18

document and just put it into notion. So

5:19

that's the giveaway. The lead magnet

5:21

that you're giving away can of course be

5:23

in any other format like a presentation

5:24

format or something like that. And then

5:26

you will usually want to create some

5:28

kind of opt-in page where you can CTA

5:31

after your video onto this opt-in page.

5:33

And when people enter their information

5:34

on this opt-in page after that they

5:36

would get access to the lead magnet. And

5:38

the way you can build this opt-in page

5:40

in a very easy way is with lovable. So

5:43

you can set up one as a template and

5:45

then if you create more you would

5:46

basically automate the process how it

5:48

just creates use that template to create

5:50

a new opt-in page and you would hook

5:52

that up to your email sending tool or

5:56

your tool basically where you collect

5:58

the email addresses of the people that

6:00

opt in and their typical tools to use

6:02

are kit or beehive. Okay, next up in

6:05

this list here are the paid ads

6:07

automations. We're still inside of the

6:08

acquisition category here. Now, one of

6:10

the things that you can do here is to

6:12

basically have an ad creative generator.

6:14

And here it depends what kind of ads

6:16

exactly you want to create. If you want

6:18

to create static image ads, those are

6:20

relatively easy to do with cloud code.

6:22

You can have a template of a static

6:24

image basically being created or you can

6:27

have Canva templates and have Canva

6:29

hooked into your cloud. And this way

6:32

these templates are basically used. And

6:33

then the benefit here is that you can

6:35

make minor edits then also in Canva

6:37

before you publish them. Now, that's for

6:39

image ads, which is easier to create.

6:40

And then if you want to create video

6:42

ads, you could hook it up directly to a

6:44

tool like Hicksfield, which is, you

6:47

know, basically just a tool that, you

6:49

know, aggregates a bunch of AI video

6:51

generators and, you know, you can just

6:53

type a prompt and it will generate the

6:55

video for you. That's one way to do it.

6:56

Or you can also, if you do more like

6:58

talking head videos and you want them to

7:00

be authentic and you actually want to

7:01

record it yourself, then you can also

7:03

just have the AI create the scripts for

7:05

you and you can then speak the scripts

7:07

into the camera and have some automated

7:10

workflows that come after this step

7:11

which do like the post-processing and

7:13

help you do the upload and so on.

7:15

Another automation system that you can

7:16

then build is essentially have your meta

7:20

ads manager, for example, it works for

7:22

Google Ads as well. you can have that

7:23

hooked up into your your cloud to the

7:26

MCP and this way it can on the one hand

7:29

monitor and give you reporting on what

7:32

is working right now when you're running

7:33

these campaigns but also it allows you

7:36

to save a lot of time that you spend on

7:39

first of all checking the account then

7:42

also uploading creating new campaigns

7:44

creating new adsets uh things like that

7:46

this can also be automated okay and the

7:48

third bucket here in this acquisition

7:49

category is basically outreach and I'll

7:52

focus on cold email here. Now, the first

7:54

step that you can do here is essentially

7:56

have it automated that you're scraping

7:58

the leads. Again, here I would use

8:01

something like Amplify is really good if

8:03

you want to, for example, tap into

8:04

Google Maps leads. If you're targeting

8:06

local service businesses or just local

8:09

businesses in general, then you can go

8:10

through Ampify and scrape these leads

8:12

from their Google Maps actor. LinkedIn

8:15

is a good source as well. And then there

8:17

are these basically databases like

8:19

Apollo where you can essentially get

8:21

leads but then also enrich leads. So you

8:24

once you have a list for example from

8:25

appy you could push it back into Apollo

8:28

to enrich them to get more details about

8:30

these leads. What you can do then do

8:31

after that after you have this list of

8:33

leads you can then basically score them.

8:35

So you just put them into a sheet and

8:37

have AI score them based on uh one to

8:40

100 fit based on your criteria and what

8:42

quality of leads they are. After that

8:44

you can have an automation that writes

8:47

the uh for you and this is similar to

8:49

the way that you would write content. So

8:51

you would also give it templates. You

8:53

would give it like a document which has

8:55

some AI writing rules and what sentences

8:58

not to use just to make it sound

8:59

supernatural. And this way you can, you

9:02

know, make sure that there's no like AI

9:04

slop sounding writing in there and you

9:06

can like remove like 95% of that. And

9:08

then finally what you can do is you can

9:10

hook it up to a tool like instantly and

9:13

have that copyrightiting essentially

9:16

just be uploaded for you. And on top of

9:18

that you can also build it in a way

9:19

because you know instantly is pretty

9:22

much like they have an API and you can

9:23

just access it or your claude or your

9:26

charge can basically access it directly.

9:28

You can also have it manage your inbox

9:29

for you. So if something comes in, you

9:31

can have it respond to a potential lead

9:34

for you through AI and you would just

9:36

essentially do the same thing. Hook it

9:38

up to to instantly and hook it up to an

9:40

intelligence and and have like Claude or

9:43

or ChadBt read the replies and then

9:46

draft a reply. Okay, let's go into this

9:48

next category here which is sales

9:50

automations. And the first one I would

9:52

like to start with is essentially DM

9:55

automation. So, you know, it handles DMs

9:58

like inbound comments. This is like one

10:00

of the examples where it's pretty hard

10:02

to build a custom tool. And so, I would

10:04

always try to use an existing software

10:06

for that. And the two best ones that I

10:09

know if you're doing it through

10:10

Instagram is just Manyhat. You can build

10:12

out these AI workflows inside of ManyHat

10:15

so that it answers for you, can qualify

10:18

the leads for you. And you can do

10:20

something similar inside of GoHigh Level

10:22

if you want to tie it to your WhatsApp.

10:23

Another automation that you can build

10:25

here on the sales side is a sales call

10:27

prep brief. So you can create something

10:29

that your prospects will be researched

10:32

before the sales call actually happens.

10:34

And the way this would work is it would

10:36

tie into your CRM. It would tie into

10:38

your calendar and then the AI has the

10:41

ability to pull that information and do

10:43

a web search about the prospect and give

10:45

you a document so that you are briefed

10:48

or you or your salesperson is briefed on

10:51

the details of this prospect and can

10:53

personalize your answers a bit better

10:55

and you would get that document before

10:56

the sales call. The next thing here is a

10:59

sales script generator. So this

11:00

technically is kind of a custom tool.

11:03

I've seen some companies do this. You

11:04

can imagine it like a variable sales

11:07

script like a dynamic sales script

11:09

basically and this sales script is a

11:12

website and you can go through it and as

11:14

you're having the sales call the the

11:16

person doing the sales call can have the

11:18

script in front of them but then can

11:19

also click on different things fill out

11:21

things and so the sales script like

11:22

adapts for you on the fly and so you can

11:25

give better answers during the sales

11:27

call. Then there is also the sales call

11:29

reviewer. So that's to review your sales

11:32

call after they have happened. If you've

11:34

done a sales call, right after it

11:36

happens, it would tie into the call

11:38

recording software, which in my case

11:40

would be Fathom, but you can use any

11:42

other like Fireflies also. It would tie

11:45

into that, pull the transcript from

11:47

there, analyze it based on any criteria

11:50

that you've given. So given it, for

11:52

example, some sales frameworks or, you

11:54

know, you can pull these also from let's

11:56

say some Alex Hermosi videos. pull these

11:59

frameworks and then use that as a

12:01

criteria to evaluate your sales calls

12:04

and then give you a report and just

12:06

output that into your Google Drive or

12:07

notion for your sales people to see. So,

12:10

you know, every day they kind of have

12:12

like live feedback on what they did well

12:14

in their sales calls and also of course

12:16

what they need to improve. You can do a

12:17

similar thing here which I call the

12:19

objection handling report which is more

12:21

on a longer time frame. It's also kind

12:23

of a review, but it would work the same

12:25

way that let's say at the end of the

12:27

month, you want to go in and have all of

12:29

your sales calls being pulled directly

12:31

from your sales call recording tool like

12:33

Fathom and just have like a batch

12:35

analysis of all of the sales call to

12:38

pull out the top objections. Using those

12:40

top objections, you can obviously like

12:42

then refine your funnel and make sure

12:44

that you're speaking the actual language

12:46

of your prospects in your funnel and

12:47

this way improve your marketing. Next

12:49

up, we have a proposal generator. So

12:51

this works in a way that again it can

12:53

pull the context from the calls that

12:55

you're having with your clients or

12:56

prospects in this case and then based on

12:59

the call transcript it can generate a

13:01

proposal. Of course this will be

13:03

customized to how exactly you want the

13:05

proposal but through the combination of

13:07

your let's say template for your

13:09

proposal and the call transcript it can

13:12

be customized to your prospect and this

13:14

of course helps increase the closing

13:16

rate. And then these two automations

13:18

both kind of have to do with your CRM.

13:21

And so the follow-up sequencer is

13:22

essentially just like sequences to

13:25

follow up with leads or people that

13:26

maybe haven't booked a call. And these

13:28

typically like you don't need to build

13:30

them yourself. You have features within

13:32

tools like if you're using kit or

13:34

beehive or if you're using for CRM for

13:37

example I'm using atio you have some

13:39

workflows that you can define in there.

13:40

A very good email sending tool is also

13:42

resend which you can connect to your

13:44

cloud code or you know uh other tool and

13:47

you can use these to build these

13:49

follow-up sequences and then you'll want

13:51

your CRM to also be able to auto update

13:54

and some of the modern CRM like atio

13:57

another example is folkier are very good

14:00

in a sense that they've opened up to

14:02

your AI tool like claude and so you can

14:05

connect atio via mcp to claude and this

14:07

way you can essentially build

14:09

automations around how your CRM gets

14:11

updated. So you can chat with your CRM,

14:13

you can just ask like which leads are

14:16

scheduled for today or something like

14:17

that. Uh and you can say like update the

14:19

status and then you can of course

14:21

automate these steps depending on what

14:23

exactly you need. Okay, next we're in

14:24

the fulfillment category. And what I

14:27

would start off here with is the

14:29

automations around onboarding. So a lot

14:32

of agencies have a very lengthy

14:33

onboarding process which costs them a

14:36

lot of time but also gives the clients a

14:38

very bad experience from the moment that

14:41

they end our program or or your services

14:43

which is bad because the first 30 days

14:45

actually determine how long your

14:48

customer is going to stay and so the

14:50

onboarding is a huge part of that. The

14:51

onboarding automations that you can

14:53

build around that is essentially you can

14:55

create custom onboarding forms. This

14:58

then allows you to just skip using tools

15:00

like tally or type form and you know

15:03

these forms can just easily be like

15:05

created with one prompt inside of cloud

15:07

code put it on a custom website and then

15:10

after that you can essentially have it

15:13

do all kinds of automations for you

15:15

after the onboarding stuff has been

15:17

collected. So the client enters their

15:19

onboarding details then you may want to

15:22

have like a workspace being provisioned.

15:24

So this is different for every company,

15:26

but in a lot of agencies, what they do

15:29

is they would create like one Google

15:32

Drive folder and have a certain folder

15:34

structure in there for all of the assets

15:35

that the client should have and then

15:37

also create the assets that the client

15:39

would need based on the information that

15:41

they've entered based on other

15:42

information they've given. Create the

15:44

assets that they would need. So these

15:45

can be things like offer documents or

15:48

VSSL scripts or even ads or landing page

15:51

copy because a lot of these assets you

15:53

can basically create them once the c the

15:56

customer has given you some information

15:58

and if your service is a done for you

16:00

service you can pretty much create these

16:02

assets for your client in a very quick

16:04

way right after onboarding and this of

16:06

course gives them a very good experience

16:08

because they pay for something and then

16:09

they immediately get value. Now, one of

16:11

the problems here is that you often also

16:14

need to chase for some information from

16:16

the client. So, for example, chase some

16:18

login or some other brand assets that

16:21

they need to hand to you. This you can

16:23

automate by having essentially a

16:25

connection into the communication

16:27

platform that you have with the client

16:29

and then you can automate the reminders

16:32

so that your client knows okay these are

16:34

the things that are missing like they're

16:35

being reminded that they need to hand

16:37

this into you. So the other thing that

16:38

can be done here is to and this is

16:41

similar uh to what I talked about before

16:43

of the assets being created but what you

16:45

can create also is some kind of kickoff

16:47

deck a presentation of let's say the

16:49

road map and also a like to-do list all

16:53

based on what the customer has entered

16:55

into the onboarding form but also pulled

16:57

from the information that they've

16:59

already given you in the calls that

17:01

you've had with them. So it could be a

17:02

sales call or maybe the onboarding call

17:04

just so you know the customer doesn't

17:06

need to repeat themselves all the time

17:08

and you can just pull that information

17:10

and fill it out for the customer. And as

17:12

your relationship is ongoing with the

17:14

customers, of course, what you can do on

17:15

an ongoing basis is pull the calls that

17:18

you're having with your customer.

17:20

Summarize them, put them into meeting

17:21

notes, extract to-dos from those calls,

17:25

and then also document them somewhere.

17:27

And this just like keeps everything

17:29

super clean. your customer has like one

17:31

page or or one folder where everything

17:34

they have is being stored in one central

17:36

location and it's kind of like a living

17:38

document or like a living directory and

17:40

just you know delivers a lot of value

17:42

for the customer. And then there's tons

17:43

of things that you can do on the actual

17:45

delivery side. And this of course

17:47

heavily depends on what kind of agency

17:49

you are. But just to give you one

17:51

example here of let's say you're a

17:52

performance marketing agency and your

17:54

job is once your client has been

17:56

onboarded, you will build them the

17:58

funnel that they will need to to run

18:00

like let's say meta ads. Then they will

18:02

need ad creatives and then they will

18:04

also need some kind of management of the

18:06

ads that you run for them and obviously

18:09

some kind of reporting. So all of these

18:11

fulfillment stuff can be automated and

18:14

the automations that you would build

18:15

here are very similar to the the things

18:18

that I've already mentioned in the

18:20

acquisition category because these are

18:22

basically the same automations, right?

18:24

You build funnels, you would build like

18:25

ad creatives and like some kind of

18:27

reporting stuff. Now you can build these

18:29

either just directly with by having some

18:31

cloud skill do that and then you would

18:34

depending on the the creative here you

18:37

would connect them to tools like Canva

18:39

or Hicksfield the same ones that I've

18:41

mentioned before. Now these automations

18:43

would of course be different if you're

18:45

let's say a ghostwriting agency right

18:47

you would have very different

18:48

automations here but uh these

18:50

automations that you built here then if

18:52

you're ghostriting agency would be more

18:54

similar to the ones I mentioned in the

18:56

earlier content section of the

18:58

automations okay let's go into the next

19:00

bucket here which is client management

19:02

and retention so how do you make sure

19:04

that you can you know increase the

19:06

chances that your clients will stay with

19:07

you now one of the things that I

19:09

mentioned in the delivery bucket as well

19:11

also is like some kind of reporting

19:12

generator. So it just should like make

19:15

sure that the client has the visibility

19:17

and the transparency into what you're

19:19

doing and how well it's performing. So

19:22

you can automate the reports that it

19:24

generates for the clients maybe on a

19:25

weekly basis or on a monthly basis. Now

19:28

you can hook that up either into just

19:31

Google Sheets and have it be created in

19:32

in Google Sheets. And the information

19:35

can be pulled depending on what agency

19:37

you are, but it could be pulled again

19:38

from the ads MCP or some other place.

19:42

You can automate workflows around your

19:44

communications too. So if you have a lot

19:46

of clients, let's say in your Slack

19:48

channels and then you can have those

19:51

channels being monitored by Slack bots.

19:53

And the way I would do that is actually

19:56

I would use a tool called Victor.

19:58

Actually, I'll I'll add that here. And

20:00

you can essentially put Victor into your

20:03

Slack channel with your client

20:04

essentially and then set up a skill and

20:07

a automation that just monitors things

20:10

that you want to keep an eye on. So this

20:12

can be simple things like the client has

20:14

written you and you haven't responded in

20:17

you know a day and then you get a

20:18

reminder that you need to respond there

20:20

or it can obviously be more complex

20:22

workflows. It could also be done in a

20:24

way that your client is actually able to

20:26

get the simple stuff done with your bot.

20:29

So your client can just tag the bot and

20:31

get some of the simple things done

20:33

directly by the bot. So you don't even

20:34

need to do the communication yourself.

20:36

So there are several possibilities here.

20:38

You can set up a weekly updator which I

20:40

guess is similar to the reporting. You

20:42

can also set up a common question spot.

20:45

The way I would do that is either use

20:48

the custom GPT function by chat GBT or

20:51

just go into cloud and create an

20:52

artifact and then just give it all of

20:55

the information that you know you have

20:57

documented about how your processes are

20:59

or like basically like if you've

21:01

documented Q&As's in the past just put

21:02

it in there and this way your customers

21:06

if they have some question to you you

21:08

can tell them like hey you can also just

21:10

ask this bot first and the bot will most

21:13

likely have the answer and if the bot

21:15

doesn't have the answer then you can

21:16

come to me and and ask me. There's a

21:18

whole another category here which I

21:20

would label as detectors and warnings.

21:23

So you can set up turn risk detectors

21:25

upsell spotters testimonial collectors.

21:28

Now the thing about these automations is

21:30

that you need to create some place where

21:32

you collect a lot of the customers data

21:35

so to say and I don't mean data in in a

21:37

like private data or anything. I mean

21:39

like all of the interactions you've had

21:41

with the customer. This can be through

21:42

the communication channels. These can be

21:44

the sales calls. So you want to track as

21:46

much as possible and put that in one

21:48

place of everything you have about the

21:49

customer. And this way you can set up

21:51

automations that then check for these

21:54

things on a regular basis. So check

21:55

whether they are at risk of turnurning.

21:58

Maybe you have certain criterias that

22:00

would you know tell you hey this

22:01

customer is at risk of turnurning

22:03

because you know they've complained two

22:05

times about something being too slow or

22:06

something like that. The opposite is

22:08

also true. You can also track whether

22:09

this C customer is super happy and maybe

22:11

it's the right time to do an upsell.

22:13

Same thing here. You can check whether

22:15

it's the right time to ask for a

22:16

testimonial. Okay, let's go into this

22:18

bucket here of operations. And these I

22:20

would say are really pretty simple. So

22:22

obviously you want to have the boring

22:24

tasks here just automated in a way that

22:26

you don't really need to think about it.

22:27

And what's part of this is like

22:29

invoicing and chasing for payments. So

22:32

typically the accounting tools and

22:33

payment tools that you have nowadays,

22:35

you can do that within that. So you can

22:37

use something like Stripe or Wise to for

22:39

example create your invoices. And

22:41

because the invoice generators of these

22:43

tools just didn't work for me and I

22:45

wanted like a really quick way to

22:47

generate invoices. So I just built a

22:49

quick custom tool and it has like a

22:51

template of everything and I can just

22:52

like oneclick create invoices. Then you

22:55

can use contract generators and there

22:58

again you can do the same thing. You can

23:00

like create a custom tool. If you need

23:02

like some kind of e signature, a good

23:04

tool to use is e-signature.com

23:07

and they have an API where you can

23:08

basically you can also tie it into your

23:10

onboarding where every time after the

23:12

onboarding the e signature basically

23:14

triggers and the client gets an email

23:15

and then they need to sign some kind of

23:17

contract. And then another thing to do

23:19

is a financial dashboard that you can

23:22

create whether it's an UI or you can

23:25

also just create it as like a simple

23:26

Google sheet which is what I've done and

23:29

essentially once a month it triggers it

23:30

just pulls information from your Stripe

23:33

from your CRM in my case at gathers all

23:35

of that information and then fills out

23:38

all of my KPIs all of the metrics that

23:40

I'm tracking and just puts it into a

23:41

nice sheet so that I don't have to do it

23:43

and all of that is done for me. Okay.

23:45

Okay. And the final category here is

23:46

team and internal automations. So one

23:49

thing which is kind of the foundation

23:50

and it's like actually one of the most

23:52

important things to have if you have a

23:54

team or you can also actually use that

23:56

for managing the clients but like some

23:58

kind of project management tool. So you

24:00

can use notion for that you can use

24:02

clickup for that which is really good.

24:03

But yeah, whether you're using the

24:05

project management tool to help you in

24:08

the management of your internal team or

24:10

whether it is to manage the projects for

24:12

your clients, it definitely makes sense

24:14

to have these tools like I would

24:16

describe it as the back end where the

24:18

data is stored and then you have your AI

24:20

which can hook into those. So through

24:22

the notion MCP for example or through

24:24

the ClickUp MCP it can hook into Claude

24:27

and your Claude is the brain on top of

24:28

that and control these project

24:30

management tools and that just speeds up

24:32

the management of these tools. Daily

24:34

standup compiler is a good thing to to

24:36

have to keep your teammates accountable.

24:37

It can be as simple as something that

24:40

again records the standup call and then

24:42

just puts the summary of the call into a

24:45

document and everybody has visibility

24:47

into that. It can also be something like

24:50

using Victor, the tool I mentioned

24:51

before, which you can just essentially

24:54

give it access to your Slack. And this

24:56

way, Victor can help you manage your

24:58

teamwide to-dos. Then you have

25:00

automations around contractor

25:01

onboarding. You might have SOPs. The way

25:04

I would do that is usually just write

25:06

like some kind of cloud skill. I don't

25:08

know how how often you would onboard

25:09

contractors, but usually it would not be

25:11

on like a daily basis. And then you

25:13

would have one skill that you can

25:15

trigger that goes through certain steps

25:17

to just get your contractor onboarded.

25:20

You can even build out something like a

25:21

custom GPT or an artifact again for your

25:24

contractor if they have any questions

25:26

which kind of ties into this last one

25:28

here as well. Internal knowledgebased

25:30

bot. Again, you can use something like

25:33

chatbas. The easiest way is to to build

25:35

a custom GPT or an artifact with just a

25:38

bunch of your knowledge and like common

25:41

questions. And then whether it's the

25:44

customer that is able to chat with it or

25:46

whether it's someone internal that's

25:47

able to chat with it, it's the same

25:49

thing, but it helps in answering common

25:52

questions and it helps remove you as the

25:54

bottleneck of being the person that can

25:56

answer these questions. So that's it for

25:57

this video. I appreciate you watching to

25:59

the end. I hope it helped. If you want

26:01

to see more of these kinds of video,

26:02

feel free to subscribe to the channel.

26:04

And if you want help implementing these

26:07

automations into your business, then

26:09

reach out to me on my website. I'll put

26:11

the link below. It's join agentos.com.

26:14

All right, thanks for watching. See you

26:15

in the next video. Bye-bye.

Interactive Summary

This video outlines 47 specific automation systems across six key areas for service agencies: acquisition, sales, fulfillment, client management, operations, and team management. Moritz, a software-focused engineer, explains how to leverage AI tools and integrations like Claude, MCP, Notion, and various CRM platforms to streamline business processes, reduce manual work, and improve overall efficiency.

Suggested questions

4 ready-made prompts