0:00
hey everyone my name is Janice and in
0:02
today's video I'm going to show you how
0:04
you can build an AI agent This AI agent
0:08
will help your customers assist them
0:10
answer to their questions on your
0:12
website or your e-commerce store You
0:15
don't need to know coding and this is
0:17
also a free option so everyone can use
0:20
it and implement it on their website Of
0:22
course if you have a lot of traffic on
0:24
your website you would need to have a
0:26
paid version of it Other tutorials about
0:29
AI agent show very complex steps that
0:33
you need to take like build a workflow
0:34
in NA10 or other no code platforms but I
0:38
will show you a way how to do it on
0:42
vector ship.ai I will leave the link
0:44
into the description so you will be able
0:47
to do it in less than 10 minutes and I
0:49
will explain every step that you need to
0:51
take to create the chatbot So as I
0:56
mentioned this is vector ship website Um
1:00
you will click get started to create an
1:03
account There are three options
1:06
available for the pricing You can use
1:08
the free version but if you have more
1:11
customers more visitors into your
1:13
website you can take a look at this
1:15
under $20 option as well Here is the
1:18
back end of web ship In this video we
1:21
will go through three different um
1:24
options that we're going to use to
1:26
create our AI agent So first thing is
1:30
pipelines We're going to create a new
1:33
Click on create pipeline
1:36
And here is uh the area where we going
1:40
to add our nodes So the first node that
1:44
we're going to add is an input node This
1:46
is the node that's going to receive the
1:49
text entered by your website visitor
1:53
I'm going to close this and we're going
1:55
to add also an output node that's going
1:58
to output a response to what customer
2:02
enters So if I connect these now and
2:08
output is output text from this one uh
2:12
sorry input text from this node and if I
2:16
click on run then I click on chatbot
2:20
and I can say hello for example
2:24
and the assistant will respond with
2:27
hello because there is no uh nodes in
2:31
between them So basically what you input
2:34
is the same in output Now we're going to
2:37
make it a little bit more interesting by
2:39
adding AI into this pipeline So I'm
2:46
and there are as you can see many
2:48
options available You can add open AI
2:51
entropic perplexity etc I'm going to use
2:55
open AI for this I'm just going to drag
2:59
and going to close this for now Um and
3:04
I'm going to disconnect these two nodes
3:06
that I connected previously And I'm
3:10
going to put the input here into the
3:13
open AI node And I'm going to put the
3:17
of the open AI node to the output So now
3:22
we need some basic system instructions
3:25
So to save time I'm just going to paste
3:27
a prompt here This is a simple prompt So
3:31
you are an intelligent and friendly
3:32
virtual assistant for workspace looks a
3:35
premium home office decor store Help
3:39
customers find products answer questions
3:41
about orders payments returns and
3:43
shipping Keep responses consense helpful
3:47
and onbrand calm Use natural
3:49
conversional language Don't speak about
3:52
anything else only about the business
3:55
By the way workspace looks is like a
3:58
simple website that I created with clog
4:02
Uh it has some products about us section
4:05
frequently asked question section and
4:08
yeah we're going to use this for our
4:11
chatbot So these are the system
4:13
instructions Now we're going to add a
4:15
prompt We're going to add the variables
4:18
So for this case we're going to add the
4:20
input what our visitor will enter into
4:27
We're going to just drag this text here
4:30
And on the output we're now going to
4:33
delete this input text from input And
4:38
we're going to use the output that open
4:40
AI provides So basically input is what
4:44
what our customer enters then open AAI
4:48
prepares an answer and then this answer
4:50
is displayed into output Yeah I forgot
4:54
to tell you You can choose models Um if
4:58
you are on a free plan you have some
5:00
credits you can use it as well for free
5:02
You can add your own API key if you
5:06
going to use it more broadly and if you
5:09
have more visitors to your website
5:13
I'm going to clear this chat and let's
5:25
so as you can see now um the assistant
5:29
the AI chatbot now responds Hello at
5:33
workspace looks we offer premium home
5:35
office decor and accessories Our
5:39
collection includes stylish desks
5:41
ergonomic chairs lightning storage
5:43
solutions decorative accents and
5:45
organization tools all designed to
5:48
elevate your workspace If you're looking
5:50
for something specific let me know and
5:52
I'll be happy to help But we now have
5:55
one problem The problem is
6:06
Okay the assistant will now respond
6:10
And now if I ask him what
6:19
he he says I don't have access to your
6:21
name because he doesn't have access to
6:23
the history of the chat that we already
6:26
had conversation about So what we're
6:29
going to do is we're going to add a
6:35
the history So if we go to data sorry to
6:41
chat and we can add chat memory node
6:45
We don't have to um do anything here but
6:50
we need to connect it to our open AI mod
6:53
model and then we need to adjust our
7:01
I'm going to um do it like this So this
7:14
need to let's delete this customer
7:22
and the customer message will be
7:40
and the chat history is the memory of
7:49
And now let's test this out
7:53
Hello my name is Jiannis
8:05
and he says "Hi Jiannis welcome blah
8:07
blah blah." And now if I ask him "What
8:14
He knows my name Your name is Jiannis
8:16
because I already told him that my name
8:22
and now there's another problem because
8:28
he can now think of answers based on
8:32
himself He doesn't know anything about
8:34
the company He doesn't know anything
8:36
about what products we sell what are the
8:41
um etc So we need to add knowledge to
8:49
we're going to add knowledge node here
9:05
menu Click on knowledge I'm going to
9:11
And here we can create a knowledge base
9:15
Basically this will be our uh looks
9:23
And we're just going to click create
9:27
and we're going to add documents here As
9:30
you can see I have information about
9:33
shipping about payments
9:36
Here I have about the company This is a
9:40
text file basically about the company
9:43
These are uh frequently asked question
9:48
knowledge base and these is about the
9:52
products we sell So I'm going to give
9:57
So we're going to add document Choose
10:00
upload files and click here browse files
10:04
So I'm going to select
10:06
four text files that I have here Open
10:10
And now these files are uploaded I'm
10:13
going to click finish
10:15
Now it's going to process these files So
10:19
the knowledge base has been created
10:22
Now we go to our pipeline
10:26
and we're going to add input to our
10:29
knowledge base reader The input will be
10:32
the text what our visitor enters on our
10:36
website into to our chatbot And we're
10:40
going to have this search query that
10:43
input knowledge base We're going to
10:45
choose the looks premium that we just
10:47
created And then we're going to pass
10:51
knowledge base to our open AI node
10:58
And we just need to insert the
11:01
knowledge base variable here So
11:06
three stars and chunks of the knowledge
11:13
So these are three variables customer
11:15
message chat history and knowledge base
11:19
just need to type in
11:24
So open II does know
11:28
what we are providing So let's run this
11:32
Click on chatbot again Let's clear this
11:36
and let's test this So
11:40
where is the product so I have executive
11:44
desk organizer So it costs $89.99
11:50
pipeline to the chatbot I write a
11:54
message How much does
12:03
so as you can see the executive
12:09
as in our doc um knowledge uh document
12:16
it also gives the description of it Uh
12:19
let's ask another question What are the
12:29
we accept all major credit cards Visa
12:32
and etc because here
12:36
we have all the payment options here and
12:39
all transactions secured with SSL
12:41
encryption And that's what our assistant
12:44
responded So I'm going to ask him one
12:47
more question How long
13:01
and it says orders are processed within
13:04
one day Uh standard is 3 to 5 business
13:08
days and express is 1 to2 Let's check So
13:12
processing is 1 day standard 3 to 5
13:16
express is 1 to2 So everything is
13:19
correct using our documents
13:22
So how do we get this chatbot on our
13:29
we're going to just deploy
13:33
We're going to go to our menu
13:37
And here we have a chatbot
13:43
here is the chatbot I've already like
13:49
Um it's using the pipeline
13:54
Uh and we can click on export here but
13:57
I'm quickly going to just show you You
13:59
can completely customize how this looks
14:02
You can add colors logo
14:08
uh welcome display etc And then you just
14:11
click export and you have this script
14:14
code that you can add to your website or
14:16
you can give it to your web master and
14:18
he will edit it for you I've implemented
14:21
the chatbot here So I'm going to refresh
14:29
as you can see workspace looks how can I
14:32
assist you so yeah this is how you
14:34
create a chatbot very easily You can run
14:38
it for free This will help to engage
14:41
your customers and I hope you find this
14:43
video helpful If you do please like it
14:46
leave a comment and make sure you
14:49
subscribe to the channel If you need
14:51
assistance implementing chatbots into
14:53
your websites please book a meeting with
14:57
me on my website jxautomates.com
15:00
I will see you in the next video Thank
15:02
you for watching Bye-bye