5 AI Agents You Can Build Today (100% no-code)
Summary
TLDRIn this informative script, the speaker discusses the creation of five distinct AI agents using no-code platforms, enabling non-programmers to build AI solutions for business efficiency. The agents include Personal, Lead Nurturing, Customer Success, Lead Enrichment, and Inbox Management agents, all designed to streamline tasks, improve margins, and reduce scaling costs. The talk highlights the use of tools like Open AI API, Pine Cone, and various databases to give these agents contextual understanding, showcasing practical applications and the potential to automate and optimize business operations.
Takeaways
- π§βπ» Building AI agents is accessible without deep programming or AI expertise; platforms like nadn are user-friendly for non-engineers.
- π The use of Open AI's GPT-40 model and Pine Cone for vector databases helps AI agents understand context and complete tasks more accurately.
- π° The primary goal of using AI agents is to increase business efficiency, reduce scaling costs, and potentially eliminate the need for additional hires.
- π οΈ AI agents can be equipped with various tools such as email, calendar, Google Drive, and CRM to perform tasks like sending emails and scheduling meetings.
- π The 'Personal Agent' can execute actions based on user requests and can be accessed via platforms like Telegram for multi-step tasks.
- π± The 'Lead Nurturing Agent' is designed to engage with leads through personalized emails, potentially doubling revenue from existing leads.
- π€ The 'Customer Success Agent' assists clients by answering questions, retrieving documents, and scheduling meetings, freeing up human customer success personnel for more complex tasks.
- π The 'Lead Enrichment Agent' collects and summarizes information about leads to strengthen the effectiveness of the lead nurturing process.
- π§ The 'Inbox Management Agent' categorizes and responds to incoming emails, reducing the time spent on email management for the user.
- π§ Custom tools like 'Find LinkedIn Profiles' can be created to suit specific business needs, such as prospecting for new leads.
- π The importance of creating agents with clear, singular responsibilities to avoid overwhelming them with too many tasks and to streamline workflows.
Q & A
What are the five different AI agents mentioned in the transcript?
-The five AI agents mentioned are the Personal Agent, Lead Nurturing Agent, Customer Success Agent, Lead Enrichment Agent, and Inbox Management Agent.
What is the primary function of the Personal Agent?
-The Personal Agent is designed to execute actions upon the user's request, utilizing tools like email, calendar, Google Drive, and a database to complete tasks efficiently.
How does the Lead Nurturing Agent help in increasing revenue?
-The Lead Nurturing Agent helps in increasing revenue by sending compelling and relevant sales engagement emails to warm leads, potentially capturing revenue from leads that would otherwise be overlooked.
What is the main objective of the Customer Success Agent?
-The main objective of the Customer Success Agent is to answer questions, retrieve documents, and schedule meetings for clients, thereby allowing human customer success personnel to focus on higher-level tasks and relationship building.
What role does the Lead Enrichment Agent play in the sales process?
-The Lead Enrichment Agent is responsible for gathering information about leads, such as from LinkedIn and Google searches, and saving it into the CRM, which helps in crafting more personalized and effective sales communications.
How does the Inbox Management Agent reduce the time spent on email?
-The Inbox Management Agent categorizes incoming emails and responds to messages accordingly, reducing the time spent on email by automating responses and ensuring important emails are prioritized and attended to promptly.
What tools does the Personal Agent have access to, according to the transcript?
-The Personal Agent has access to tools such as email, calendar, Google Drive, a calculator, Wikipedia, and a window buffer memory for contextual understanding of past conversations.
What is the significance of using a vector database like Pine Cone in building AI agents?
-Pine Cone is used for the vector database to create a knowledge base for the agent, providing it with contextual understanding and awareness to accurately and effectively complete tasks.
How can AI agents help in reducing the need for hiring additional staff?
-AI agents can automate repetitive tasks, handle customer inquiries, and manage workflows, which can lead to increased efficiency and reduced reliance on hiring more staff for certain roles.
What is the potential impact of AI agents on the future of hiring practices?
-The future of hiring practices may shift towards hiring individuals with more experience and skills, as AI agents can take over roles that require less expertise, allowing human employees to focus on more strategic tasks.
Can you provide an example of a multi-step task performed by the Personal Agent?
-An example of a multi-step task performed by the Personal Agent is sending an email to a contact and scheduling a meeting. The agent uses email actions to send the email and calendar actions to schedule the meeting.
What is the role of the 'Autof fixing output parser' tool mentioned in the transcript?
-The 'Autof fixing output parser' tool is used to ensure that the agent outputs a very specific JSON package, which includes components like the recipient's email, message, subject line, etc., for the purpose of sending emails through a Gmail node.
How does the Lead Nurturing Agent determine which leads to contact?
-The Lead Nurturing Agent is triggered when the CRM management agent identifies a lead that hasn't been contacted in the last seven days, ensuring regular touchpoints with leads.
What is the purpose of the 'Google search' tool in the Lead Enrichment Agent?
-The 'Google search' tool in the Lead Enrichment Agent is used to find additional information about a lead, such as articles about the company or profiles of the founder or CEO, to provide extra context and enhance the personalization of sales communications.
How does the Customer Success Agent handle client inquiries?
-The Customer Success Agent uses tools like Slack or email, calendar, Google Drive, and project management tools to quickly answer questions, retrieve documents, or schedule meetings for clients, streamlining the process and allowing human customer success personnel to focus on more complex issues.
What is the purpose of the 'LinkedIn scraper' tool in the Lead Enrichment Agent?
-The 'LinkedIn scraper' tool is used to gather publicly available information from LinkedIn profiles of leads and summarize it, adding it to the CRM for better understanding and personalization of sales approaches.
What is the role of the 'Categorize' tool in the Inbox Management Agent?
-The 'Categorize' tool in the Inbox Management Agent is responsible for sorting incoming emails into appropriate categories, such as promotions, leads, or sales emails, to streamline the email management process.
How does the 'Notify' tool work in the Inbox Management Agent?
-The 'Notify' tool in the Inbox Management Agent is used to alert the user or the relevant person when an email requires a response, especially for high-priority or urgent messages, ensuring timely communication.
What is the purpose of the 'Find LinkedIn profiles' tool mentioned in the transcript?
-The 'Find LinkedIn profiles' tool is used to search for and compile a list of LinkedIn profiles based on specific parameters, such as job title, location, and industry, to aid in prospecting and lead generation.
Outlines
π€ Introduction to AI Agents for Business Efficiency
The video script introduces the concept of building AI agents using no-code platforms, which simplifies the process for non-programmers and non-AI engineers. It discusses the use of no-code tools like N8n for building agents and the OpenAI API for leveraging the GPT-4 model. The purpose of these AI agents is to assist businesses in increasing revenue, reducing costs, and potentially eliminating the need for hiring additional staff for certain roles. The script outlines five different types of AI agents that can be built: personal agent, lead nurturing agent, customer success agent, lead enrichment agent, and inbox management agent, each tailored to specific business tasks and processes.
π§ Personal Agent for Task Automation and Efficiency
This paragraph delves into the functionality of the personal AI agent, which is designed to execute actions based on user requests. The personal agent has access to tools like email, calendar, Google Drive, and a database to perform tasks such as sending emails, managing calendar events, and retrieving documents. The script demonstrates how the personal agent can be accessed via Telegram and how it uses the OpenAI chat model to understand and carry out multi-step tasks, such as sending an email and scheduling a meeting. The personal agent is highlighted as a valuable tool for business owners and managers to automate routine tasks and free up time for more strategic activities.
π± Lead Nurturing Agent for Engaging Warm Leads
The script explains the role of the lead nurturing AI agent, which is tasked with sending engaging and relevant sales emails to warm leads. The agent uses tools like email, Google Drive, and CRM to personalize communication and potentially increase revenue from leads who have previously engaged with the business. It discusses the common challenge faced by agency owners and solo founders in maintaining consistent touchpoints with leads and how the lead nurturing agent can automate this process, ensuring that leads receive regular, personalized communication that could lead to conversions.
πΌ Customer Success Agent for Client Support and Relationship Building
The customer success AI agent is introduced as a tool for answering client queries, retrieving documents, and scheduling meetings. This agent is designed to be customized for each client, giving them access only to relevant documents and project management tools specific to their account. The script highlights how this agent can help customer success teams focus on building relationships and providing high-level solutions by handling routine queries and document retrieval, thereby increasing client satisfaction and potentially allowing teams to manage more clients with fewer resources.
π Lead Enrichment Agent for Enhancing Lead Data
The lead enrichment agent is described as a crucial component in the lead nurturing process. This agent is responsible for gathering information about leads using tools like Google search, website scrapers, and LinkedIn scrapers, and then saving this data into the CRM. The script emphasizes the importance of this agent in providing the lead nurturing agent with the necessary context and information to send out personalized and effective emails, as well as equipping sales teams with the knowledge needed for productive sales calls.
π¬ Inbox Management Agent for Streamlining Email Workflows
The script introduces the inbox management AI agent, designed to manage incoming emails by categorizing and responding to them accordingly. The agent uses email actions and interacts with other agents like the lead qualification agent for lead-related emails. It is highlighted as a tool that can significantly reduce the time spent on email management, allowing business owners to focus on more critical tasks. The agent's role is to ensure that emails are not just responded to but also appropriately categorized and directed to the relevant agents or tools for further action.
π οΈ Custom Tools for Prospecting and Business Growth
The final paragraph showcases the flexibility and potential of creating custom tools for specific business needs. An example is given of a tool that can search for and compile LinkedIn profiles based on specific parameters, such as job title and location. The script demonstrates how this tool can quickly generate a list of prospects, which can then be used for outreach or further sales processes. The paragraph emphasizes the importance of tailoring AI agents and tools to the unique requirements of one's business to maximize efficiency and growth.
Mindmap
Keywords
π‘AI agents
π‘No-code platforms
π‘Open AI API
π‘Pine cone
π‘Lead nurturing
π‘Customer success
π‘Lead enrichment
π‘Inbox management
π‘CRM
π‘Google Drive
π‘Telegram
Highlights
No-code platforms enable anyone to build AI agents without programming skills.
NAND and Open AI API are used to build AI agents with contextual understanding.
Pine cone vector database is utilized for knowledge base creation.
AI agents can increase business margins, reduce scaling costs, and eliminate the need for hiring more staff.
Five different AI agents discussed: Personal, Lead Nurturing, Customer Success, Lead Enrichment, and Inbox Management.
Personal AI agents can perform actions like email sending, calendar management, and document retrieval.
Lead Nurturing agents send personalized emails to engage warm leads and increase revenue.
Customer Success agents handle client queries to free up human customer success personnel for higher-level tasks.
Lead Enrichment agents gather information about leads to enhance the effectiveness of Lead Nurturing agents.
Inbox Management agents categorize and respond to emails, reducing the need for manual inbox management.
Agents can be triggered by other agents or direct user commands to perform specific tasks.
Common tools across agents include email, calendar, Google Drive, and CRM for streamlined workflows.
Agents can be customized with additional tools for specific business needs.
AI agents can improve efficiency, allowing for better task management and time savings.
Agents can be built to work together, creating a network of AI that enhances business operations.
Examples of unique tools like LinkedIn profile scrapers for lead generation are highlighted.
The flexibility of AI agents allows for the creation of custom workflows tailored to specific business processes.
AI agents represent a new space with evolving frameworks and processes, offering immense potential for business optimization.
Transcripts
talk about five different AI agents that
you can build today and theoretically
anybody can build these agents we use no
code platforms so you don't need to be a
programmer you don't need to know how to
code you don't need to be an AI engineer
um you can learn these platforms pretty
quickly we use nadn to primarily build
our agents and the tools that go along
with them we use open AI API to get
their gp40 model and we use pine cone
for the vector database where we create
the knowledge base for the agent so it
has the contextual understanding and
awareness to accurately and effectively
and reliably complete tasks right
obviously if you did know how to code
obviously if you were an engineer you
can probably get more out of these
platforms but we're just looking to use
these agents to help our businesses make
more money increase our margins decrease
the cost to scale overall and hopefully
eliminate the need to increase headcount
in order to scale right we want to get
to a point where we're no longer hiring
people
for roles where it's almost like you're
hiring them to maintain where you're
currently at right like you've reached a
certain point where the bandwidth is
getting insane maybe you're taking on
more clients or maybe you uh there's
like some initiatives that you want to
do like combing through a lead list
following up with leads you know some
things that you just that are on your
to-do list but you haven't got to yet
and you're thinking oh man maybe I need
to hire a virtual assistant or maybe I
need to hire an SDR or maybe I need to
hire a personal assistant right instead
of hiring people like that
I think the future is going to be AI
agents instead and then the people that
you do hire are people that probably
will cost more but you'll have the
margins to pay for to pay for them right
to afford them um but they come with
more experience and more skills and
themselves can be a lot more leverage on
your business than hiring maybe like a
virtual assistant or someone where we
can kind of replace their role um by
using Ai and AI agents right so the five
different agents that we're going to get
into today are the personal agent lead
nurturing agent customer success agent
lead enrichment agent and inbox
management agent okay we'll get into the
details on all of these so the personal
agent is U one that I use um almost
every day it's supposed to execute
actions upon the users's request and the
tools that it has are email calendar
Google Drive and update database I'm
going to add some other tools to it like
CRM and project management as well um
but for now these are the tools that it
has and it looks like this
and we can actually just go ahead and
take a look at it in nadn so we can see
under the hood what it looks
like um and before I do that I created
this project here just to put kind of
demo agents and tools um so we can just
like see them right but you can see the
ones with Bots are agents with the bot
emoji and then the ones with the hammer
and wrench are the actual tools right so
we build our agents and our tools all in
one place makes it super super easy to
just get in there um add add a new tool
to an agent build an agent really
quickly if you need like a workflow to
be fulfilled you can build an agent in
an afternoon right um so it's just like
a One-Stop shop this is inen just a
One-Stop shop for building agents so
let's get into it personal
agent so let's get into each of the
components
here I primarily access this agent via
telegram so I just have a telegram
Channel every time I send a message it
triggers this agent to go and do
something um so I'll send a message it
triggers the agent the agent decides
like what it needs to do right and we
give it access to obviously the open AI
chat model here so it can you know be an
ai ai agent um and then I give it like
calculator and Wikipedia just you know
just in case like I want to calculate
something or pull something from
Wikipedia why not um and the window
buffer memory this actually gives it
context of the past conversations that
I've had with the agent um so like up to
like 20 messages at least that's where I
said it it's like 20 messages in the
telegram Channel where I talk with the
agent it still has access to those
conversations so it it like if there's a
normal flow like if I said like um hey
what's Andrew's email okay it pulled it
it'd be like okay um can you actually
like send him uh can you actually send
his email to um you know John you know
let y'all know that this is an email or
something right so like something where
it's like um you know you asked to to do
something and then it's like using that
context to go and do the next thing
right uh and then we give it its tools
so here where it says context data pine
cone Vector store this is the knowledge
base uh for the agent and primarily the
agent is accessing this this knowledge
base to retrieve like contact
information uh that what we found is is
contact information is like by far the
most uh requested like piece of
information in order to complete tasks
um so in this case I get I give it
specifically like contact information
right and then we have obviously email
actions so send an email uh get an email
reply to an email Etc calendar actions
updated datab or update a calendar event
create a calender event delete an event
Etc updated database so this allows me
to like manually add things to the uh
Vector store to the knowledge base so
comes in handy when I'm at like a
networking event or like a conference or
something and I'm talking with some
and you know they've indicated some
interest in like getting involved with
agents or whatever and working with our
team and I want to like save their
information in our CRM and also in the
knowledge base so I'll just say like in
the telegram like hey can you add uh you
know Jane Smith here's her email to the
uh to the database and it'll go and like
update the database with their
information right um and then Google
Drive so obviously it'll pull documents
from our Google Drive if I need the
documents right so like hey can you get
me the YouTube description template boom
gives me the template like right
away um and like so let's see this thing
in action all right so I have telegram
open here on my
desktop um let's let's do a uh a
multi-step task so we can kind of see
the scope of what it's capable of doing
so let me say
um send an email to
Andrew and ask him if he finished
the elite digital
proposal Osco schedule
time to meet with
him uh for tomorrow at let's call it 6
PM okay so multi-step task right we want
to send an email to him and we also want
to schedule a time to meet so the agent
should be working with uh email actions
and calendar actions okay so let me test
workflow
so the agent is running it's using the
open AI chat model to obviously
understand what I just said right now
it's using email action so it's going to
send him an email and ask him if he
finished the proposal for elite digital
looks like that worked great now it's
going to schedule the calendar event for
tomorrow at 6 I think is what I said
yeah boom that worked awesome and let's
see what it said it said I've sent an
email to Andrew Lewis asking if you
finish the elite digital proposal a
meeting has been scheduled for him
tomorrow at 6m you can view that event
here so let me click
it just to check its work why
not okay it looks like it did I've done
this a couple of times sering a couple
um but obviously this was the 6m one
discussed the completion status of elite
digital awesome cool so it did it um and
then obviously like I could show we
could show more things let's see if the
was sent I think it was because it has
it
there and it was awesome I hope this
message finds you well I finished the
have you finished the elite digital
proposal also can we schedule a time to
meet tomorrow at 6m best regards awesome
Okay cool so everything was sent
properly um and that should give you an
idea of how the personal agent Works um
I like this one a lot because it it only
acts when you tell it to act and um you
can give it as many tools as you want
right there is a moment where it becomes
like kind of tool overload for the agent
um but for this scenario like the fact
that you're kind of telling it
explicitly what to do uh it makes it
easier for the agent to kind of
understand which tools to use and when
right uh so that's the personal agent I
think any any founder any uh manager
even um and pretty much anybody like any
team member period probably should have
one of these just as a personal agent to
retrieve documents retrieve information
um and then like quickly do easy tasks
um that you know they would have to
like and they would have to like log
into their inbox or log into their
calendar but what what I was going to
say is like it's not like these things
are taking up too much of your time but
they do add up right a majority of our
time is spent in our inbox and in our
calendar so if we can kind of eliminate
a couple of different things here and
there it does start to add up as a as a
positive so let's move on that was the
personal agents again something I use uh
almost every day especally when I'm on
the road or or moving around uh lead
nurturing agent so this one is designed
to send compelling and relevant sales
engagement emails to warm leads tools
are email Google Drive and CRM so this
one we created because a a very common
um paino from agency owners or solo
Founders or people with kind of like
small teams is they have this lead list
they have this this uh somewhat massive
list of leads of people who have like
somewhat engaged with them or their
content or lead magnets or their ads at
some point right and they know that if
they just like went through that list
consistently like every week and emailed
everybody and sent a very personalized
touch point and maybe sent um pieces of
content or case studies that they think
would be
relevant to the lead that they would be
able to capture more revenue from their
leads right too many agency owners or
Sol Founders end up in this cycle and it
makes sense because you just don't have
the timer bandwidth but they end up in
the cycle where you run an ad or you run
a piece of content or a lead Magnet or
something and you kind of expect that uh
people see that and then they
immediately book a call and then you hop
on the call and then you close them um
but the reality is that a lot of people
aren't ready to buy right then there's a
nurturing window that needs to happen
and if you just don't have the bandwidth
or or the time you're not getting those
touch points in you're not nurturing
properly right um also what can happen
is maybe you have a lot of leads in
there that aren't qualified for your
recurring service offer but they might
be buy they might be able to buy like
the ebook that you offer the $20 ebook
right so like your lower ticket offer or
maybe that they'd be willing to buy your
$2,500 course on what you do right
medium ticket offer um there's a lot of
cases where people are fans of yours or
like what you're doing but they just
can't afford your high ticket stuff
right or they just don't need it but
they want to learn more and they're
willing to to pay money right so the
lead nurturing agent is really good at
capturing that Revenue that's in your CR
right it can almost double your Revenue
at times because you don't know how many
people are in there as silent fans who
are watching uh you know every piece of
content that you post and they're just
not going to buy they're just never
going to buy the high ticket thing from
you but they're learning a lot from you
and they're willing to buy something um
that might be a lot cheaper but still
provides value and so we want to make
sure that we get get that in front of
them right um and that one looks like
this uh and we'll hop into nadn and just
take a look at it
as well why
not lead nurturing agent
cool um I don't know if I said this but
I basically just like put a bunch of the
agents that we have into like this uh
one project here as demo versions so um
that's why you see some of these as like
deactivated and stuff we just don't want
them running or anything to like break
just because I'm I'm showing demos but
anyway the lead nurturing agent gets
triggered from other agents so we have
an inbox management agent that we'll get
into later and whenever an email comes
in that's from a lead it kicks it to the
lead nurturing agent um actually that's
not true it it kicks it to the lead
qualification agent the lead nurturing
agent gets triggered whenever the CRM
management agent sees a lead that hasn't
been contacted in the last seven days so
that one is designed to uh review the
CRM like every hour if there's any lead
in there who hasn't been who hasn't
received a touch point in 7 days it
sends that lead and all of their
information from the CRM it sends that
to the lead nurturing agent and then the
lead nurturing agent obviously uses the
database uh to retrieve any data that it
needs right like it it probably like
usually it doesn't actually need to
access the database in this case but um
sometimes it does because you know you
have like specific it depends how many
Like rules you have or um what your
content preferences are or how large
your brand document is um so like you
know in some cases you can kind of like
tell the agent how to craft emails and
and messages just like in the prompt but
a lot of times like it's it's a lot of
data and you just want to put that in a
database so it's more efficient so
that's why we have the database here um
this right here the autof fixing output
parser so we need the agent to Output a
very specific uh Json package so that
way we can actually like put each of
those components like who's the email
going to like what's their address
what's the message what's the subject
line Etc so we can put that stuff into
like a Gmail node that'll actually like
send the email right so we just got to
make sure that the output comes out
perfectly so we use this and then here
are the tools right CRM Google Drive
email actions and you're probably going
to see a pattern here we reuse the same
tools for a lot of these agents um you
the reality is like every agency has
almost the same text stack which is like
email calendar project management stuff
like software um and then like a
database of sorts for their documents
like Google drive or notion or one drive
or Dropbox or whatever um and then you
have like platform specific stuff just
depending on what kind of agency you are
like what exactly you offer but we all
have the same stuff right email and
calendar like we all have that so you're
going to see a lot of the same tools
being used with different agents um so
yeah that's the lead nurturing agent
super crucial to capturing more Revenue
that's in your lead list that's just
hidden there right people who are ready
to buy but they just haven't got enough
touch points from you they haven't
gotten enough touch points from you uh
nor have you gotten the right offer in
front of them at the right time customer
success agent so this one's objective is
to answer questions retrieve documents
and schedule meetings for a client uh
the tools that it has is slack or email
usually both uh maybe not both calendar
Google Drive and then trer or like
project management stuff so this one's
designed to be uh a custom agent for
each client that you have so if you have
30 clients all 30 clients will have
their own customer success agent that
only has access to the documents or like
the folder in Google Drive that has all
of their stuff right and then um the
actual like project management stuff
that has to do with that specific client
and then any kind of like other
additional information uh just
specifically about that client so if
ever the client is asking like for just
an easy easily answerable qu question or
to retrieve some kind of document or
anything that can be like really quick
and automated the agent steps in and
handles that and it allows your actual
customer success person to focus more on
like building the relationship or
providing uh higher level solutions to
whatever problems that are happening um
it also allows them to have the
bandwidth to handle more clients so you
don't need to hire as many customer
success people because a lot probably
about like 40 to 50% of the queries that
are coming in from your clients are
being handled by their specific agent
right it also reduces turn they're
getting their answers they're getting
their documents that they need much much
quicker from you because they're
speaking with an agent right um and
there might be a little bit of a a habit
change or learning curve or process
change to some clients like people
aren't necessarily used to working with
AI agents you know directly um but like
what I've seen is like after a couple
days or like a week of like using it and
the agent just provid in the answer or
providing the document right away people
really start to um like enjoy it and and
and almost be upset if we were to like
take it away right it's such a huge
asset um and this one looks like this
let's just go ahead and hop in I don't
even know why I put the pictures on
here um because I'd rather just show it
in the
actual platform like where it is um
let's see where's this one CX agent yeah
okay so um we primarily use slack
deactivated it because I don't want
anything whatever I just don't want this
agent to get triggered um so if anything
is like sent into the slack this gets
triggered and the agent kind of decides
like first of all do I need to answer
this um if I do need to answer it can I
actually answer it do I have the tools
and the ability and the knowledge base
to answer it um and if I can let me go
ahead and answer it right uh and then we
have execute workflow trigger so
sometimes your customer success person
which a lot of times is you but your
customer success person will have a
personal agent of their own and so
they'll tell their personal agent like
uh oh man I forgot to send the uh
proposal or the uh the new scope to the
new s so to uh you know this this client
um can you make sure that that gets sent
and so the personal agent will trigger
the uh specific customer success agent
to send that document
right um so that's how we like there's
there's ways to connect these agents
together to have them work together but
we try to keep that as simplified as
possible um and but we'll get into some
of that later when I do videos on like
architecture and and designing how they
should work together and whatnot best
practices um but the components of this
are the same like we give all of our
agents a database and then calendar
actions again Google Drive actions again
a SAA for like project management and
then email actions again right like I
said these tools they start to repeat
themselves because there's just like
common workflows common tools that
everybody
uses lead enrichment agent so this one's
key because it helps your Le uh lead
nurturing agent actually do a good job
right and not just send like a generic
uh like marketing email that just gets
sent to spam and nobody really reads the
lead enrichment agent is basically
responsible for finding information
about that lead uh about all of our
leads and saving that into the CRM and
so the tools we give it are Google
search website scraper LinkedIn scraper
and
CRM um and this one looks like this so
let's let's go back into inate
in and let's take a
look uh which one which one are we doing
enrichment lead
enrichment okay so this one
we're using Google Sheets um for our CRM
for now just no need to there just
hasn't been a need to like get a real
CRM yet um but anyway same thing here we
give it the vector stored the database
then we give it a LinkedIn scraper so
this tool will obviously go and scrape
uh any information you can find on
LinkedIn LinkedIn has some um pretty
harsh restrictions on scraping their
website understandably so um but they do
allow like public like even if you're
not signed in to to LinkedIn there is
publicly available information that they
they put um so we can get we can at
least get that information then there's
some third party tools where you can
gain access like extensions and whatnot
but I swear like every extension that
I've used um eventually just gets my
account like somewhat flagged or or
temporarily banned or something um and
so it's it's kind of it's it's um it
scared me away from using like third
party tools uh for LinkedIn specifically
but you know I've been I've been trying
to do that LinkedIn stuff for a while so
maybe things have gotten better but
LinkedIn scraper obviously pulls
information about LinkedIn summarizes it
and puts it into the CRM CRM this is
just to put stuff into the CRM right um
website scraper this will go to their
website scrape it summarize put it into
CRM and then the Ser API this is just to
be able to do like Google search so
sometimes we just have it Go and like do
a Google search of the person and like
the company see if there's any like
articles that were written about the
company or profiles about the founder or
CEO um that we might want to reference
or pull in for just additional contexts
about the person um and then obviously
add that information to the CRM so this
one's job is like very key in the whole
process and that it actually gives us
that that extra Edge that extra those
extra pieces of information that make
the emails that are being sent out by
the lead nurturing agent that much
stronger right and then also like for
anybody who's on our sales
team all this information is being saved
in the CRM so like you know you're
hopping on a sales call they're the
salesperson's like personal agent is
drafting like a brief about the person
and about the the account and sending it
to the to the salesperson 30 minutes
before the meeting and all the
information from the CRM is very
important right like the salesperson
needs to know this stuff in order to
have a really good uh and productive
call so this one's really really key
just in terms of information gathering
and making each of us smarter when we
hop on sales calls or hop on calls with
our clients
okay last one here inbox management
agent I think everybody should have this
one um I I'm like hardly in my inbox now
because of the agents that I've set up
and uh it's been nice um it used to be
like every day I would just like the
first thing I do is go into my email and
answer a bunch of emails and um eyes
constantly on my email account and like
uh or my inbox and I'm like you know
trying to get in there and answer answer
emails really fast and uh I still like
to answer emails fast but
I'm hardly actually in my inbox anymore
if I get an email goes to my telegram my
agent lets me know um if it's an email
that I don't need to respond to and an
agent can respond to the agent will
respond right and it's also
categorizing like okay this is promo
let's get that out of here this is a
lead let's make sure someone responds to
this uh this is a sales email like
someone's sending us let's put this in a
label in a folder uh let's label it tag
it appropriately in case we want to like
reach out to that vendor at some point
right um and so obviously this agent is
managing the founders inbox not just the
founder but anybody's inbox and
categorize and responds to messages
accordingly we give it email actions we
also give it other agents so this one is
really responsible for uh dishing out
what needs to happen so like a leag
comes in it sends it to a lead
qualification agent so it can like book
The call right if the leag comes in
interested leag qualification agent is
the one who will actually have the
conversation and then schedule the thing
at the end the inbox management agent
will not do that it's not actually
responsible for handling anything to do
with leads it just handles like very
basic responses that need to happen and
frankly it hardly responses like let's
be real it's hardly respond responding
to emails um it's mainly just kind of
like triggering agents other agents that
need to be triggered to do something or
just categorizing the uh emails and then
we obviously give it CRM and
categorization tools so we I say
categorization tools because um there's
like specific ways we want to categorize
everything and in instead of like giving
the agent a ton of like rule sets for
how to categorize stuff we're just going
to create a tool that can that will
handle those rules for it right like we
want to like the agent's only job
realistically is to um
identify identify the email and then
determine what exactly it's supposed to
do like reason what exactly is supposed
to happen once it's read that email or
what it needs to do with that email
right um it's not supposed to like if
it's like okay we need to categorize
this tool we need to categorize this
email okay go send it categorize and
then we also need to um trigger the lead
uh nurturing or the lead enrichment
agent or no the lead qualification agent
um maybe we need better names for these
but the lead qualification agent to like
respond to this lead like okay go do
that instead of like oh we need to
respond to this lead with a really um
wellth thought out compelling message
with content pieces um as well as C
categorizing the email in this perfectly
in this exact way like we basically what
I'm trying to say is we just don't want
to give it too much to handle we want to
give it a very clear like this this is
your this is exactly what you're
supposed to do and any of the additional
stuff the details that are part of that
are going to be handled by the tools
themselves right not by the agent um
that seems to be work the best for us if
any of you guys are building agents and
you have better processes around this
stuff please let me know this is
obviously a new space people are
learning building their own frame works
and their own processes and and Sops
around it so um you know things are
constantly evolving
here but this one looks like this um
and I'm not going to go into it I can
just do it from here um actually I will
it'll just be bigger for you guys to
see uh you can see lead qualification
this was the one I was talking about uh
kept getting mixed up so inbox
management
okay so this one gets triggered off
Gmail right so any message that comes in
boom this one's the first thing that
like this this is the agent that sees it
first right sees it before me sees it
before anybody else usually right and
then it kind of it has the vector store
it understands like contact data so if
it's it if it uh is a lead or something
that information will be in the database
and so it'll be able to say like oh okay
this is a lead right um
let me kick that to the lead
qualification
agent uh then we also have reply to an
email this is basically email actions
but I've kind of simplified this tool in
particular for this agent because all
it's doing is replying to an email it's
not going to be like it's just not going
to be doing like that much else um but I
actually might change this to all email
actions just in case it like we want to
add some new rules to it um I might
change it to email actions but like I
said it it's actually hardly replying to
an email anyway so maybe it won't matter
uh categorize so this handles all the
categorization so when you're when it's
like okay the the uh inbox management
agent its job is to categorize all the
emails that come in it's like okay the
agent itself isn't actually doing the
categorizing it's not it's not saying
like oh this was a promo email like
Market as promo like it's just saying
like oh I need to categorize send it to
the categorize tool and the categorize
tool will handle all that right uh CRM
obviously lead qualification agent that
was mentioning and then notify so this
is just to notify me or whoever's inbox
it is that they need to respond right so
after it's categorize it's like oh man
this is high priority super urgent
notify me right so it has those rules in
place to do
that um and I kind of just quickly that
was the last one but just like before we
end the video I just want to dive into
the tools just a little bit um like some
of the tools I didn't even talk about
like fin LinkedIn profiles this is a
good one like this this tool right here
will basically be like if I say
um if I
say let me just show you let me just
show you uh let me pull up the sheet so
basically what okay I got the sheet
pulled up this is like so basically what
this uh tool does is it I I basically
say like um hey can you give me a list
of um LinkedIn profiles
for uh Founders in AI who are located in
Los
Angeles um and it'll go and it'll it'll
basically just like do a Google search
for LinkedIn profiles based on those
parameters grab as many as I said to get
like let's say I said 100 it would grab
100 um and it would put them in the
sheet and it it would also like Define
like what I did so like I did a couple
of test runs earlier and it's like okay
LinkedIn search results for founder
United States search results for founder
new y city um for project manager for
agency so like let's say let's see it in
action let's say I said
um I'm to test it out and say
like give me a list of let's say
300 uh
Founders located in let's say
Dubai um who work in AI
um yeah that works so we do
chat and then we can see the tool
working
here and then it's obviously going to
get to the uh the sheet eventually um
and then we'll be able to see it it will
have
created an additional sheet here so okay
look LinkedIn search results founder
Dubai cool let's open it up and soon
we're going to see like it's probably
going to be like 10 at a time that just
go boop boop boop boop boop and get us a
nice list of high quality prospects I
mean um I wonder you know these are
probably good leads maybe for us um but
we're in AI as well so like maybe these
people already know like
what you know maybe they're not actually
good fits for us um but you can see like
how easy it is to just like pull a list
of LinkedIn Prof files and then like
take that list and put it into some kind
of like automated like Outreach thing or
automated like connection request thing
or something like that um but like super
easy super quick like prospecting tool
that you know I'm probably going to be
using quite a bit uh in the future I
kind of just built this one a few days
ago but I can I can see it being like
super super helpful down the road um so
this is this is an example of like a a
unique
tool that you're not necessarily going
to be attaching to every single agent
that you have it's not like email or
calendar or Google drive but it's a
really helpful tool for maybe like a an
all-encompassing like prospecting agent
like someone who does Google searches
and not only finds like uh LinkedIn but
like also gets email accounts for the
lead gets their name like full name and
it it really fills out um an entire like
leadless like column set you know that
that you would have like in CRM like
first name last name email address
location uh job description you know job
title like all of that information like
we could build this out even further or
or we could just like build an agent and
attach the proper tools and the agent
would just like constantly be
prospecting leads for us right um and so
that's a pretty cool tool and I just
show you that to just say
like the flexibility here is immense and
outside of kind of your common
tools there's a lot of opportunity to
create your own workflows based on your
actual business right your specific
business your specific offer your
specific ICP um and so that's why we
love agents and if you want to learn
more about it I got a video that dies
deep into kind of our process how we've
even gotten to this point um and some of
the common agents that we're seeing how
we build our agents what components we
use kind of like all of the details
around it should be the first link in in
the uh description but other than that
these are five agents that you could get
get started building today if you have
any questions let me know in the
comments um or feel free to email me but
otherwise that's it
Browse More Related Video
How I Automated Phone Calls With AI | STEP BY STEP
How you should think about AI Agents this 2024. (Early Mover Advantage)
Chatbot or AI Agent Setting up crewai framework for scaling tasks
I Taught AI Agents How to Build Ads for Me [5 FREE ChatGPT Scripts]
Jira Service Management & AI
"I tested 400 AI Agents, these are the best" - Adam Silverman
5.0 / 5 (0 votes)