today we're going to learn how to craft instructions for your AI agent instructions are the difference between a good result and a bad result you know a successful operation and a failed operation if you kind of try with your instructions then you're going to kind of get the results you want but if you follow what I'm going to tell you in this video by the end of it you are going to be able to craft instructions for any use case any agent and you're never going to have a problem with your agent outputting information wrong
ever again I'm telling you once I learn these two techniques that I'm going to share with you in today's video everything changed I don't have to worry about my instructions now not only do I not have to worry where they're organized at but I just know that they're going to work because of the way that I operate and the way that I craft them so that's what we're going to be learning in today's video So today we're going to be learning about crafting instructions for your AI agent it's going to be an in-depth guide on
creating instructions now for those of you who actually watch this video take notes and apply what you're about to learn you will see success with your AI agents in the future you're not going to have to worry about them going off on their own little tangents maybe hallucinating not doing the exact output that you want you're going to learn literally everything about how to get structured outputs and consistent outputs as well so let's dive right into it first of all what are the benefits of good instructions number one you're going to get consistency when you
learn how to craft instructions your outputs become much more predictable now I don't know about you but if I'm creating an automation or a workflow an agentic workflow I want my results to be predictable I want to know that it's going to work especially if I'm building it for someone else or if it's like a post product purchase email I want that output to be predictable so how do we do that well that's what we're going to learn number two you get structured outputs now good instructions can turn one AI agent output into a structured
output with multiple Fields number three organization you can keep all of your instructions easily downloaded on your computer allowing for a single source of Truth and easily editable documents don't we all love that don't we love when we know where something is we can just go and simply edit it we're not scrambling around in the back end of one of our old prompts in chat gbt from who knows how long ago in today's video I'm going to show you a setup that I use in order to craft instructions keep them all organized that way if
I'm ever want to change something with my agent I can edit that file first and then upload it now some people say it's Overkill some people don't think it's necessary but personally if I'm building things for people and people are paying me money to build them automations or even if I'm building things for myself I want to know where all my data is right I want to be able to edit that I want to be able to easily store that so that's what we're going to learn today and we're also going to learn how to
craft instructions I'm going to break everything down step by step so what are the instruction types well we're going to be learning two things XML and Json we're going to be learning XML because llms can read this better and humans can learn to write in this style quickly now this is a more robust and structured way of creating any type of instruction and then Json this is going to allow you to create structured outputs with your instructions so that you can take one output and break it down into multiple Fields this is useful again when
you want multiple outputs from a single execution of your AI agent and you want to be able to take bits and pieces of that entire output this unlocks more opportunities with any agent building software that you choose whether you're using nadn make it really doesn't matter being able to craft instructions and add structure Json outputs to your agent is a huge unlock so let's dive into creating instructions for an AI agent project now if you get some value out of this video and you want to keep expanding your knowledge in the world of AI agents
and just AI in general then consider joining our AI foundations Community now in this community we like focusing on the three C's courses that get you up to speed and also way ahead of the curve with AI we have a full agent building course in here the next C is calls we love doing live calls as a community live q&as live automation support and much more and then finally the lovely Community the people that make this community up we just released this new nadn agent building course and Jason here said the course is great and
it's increased my knowledge tfold we even have community members doing networking and masterminding calls with the rest of the community so if you want to be surrounded by a group of people who think about Ai and implementation all day long you want to expand your knowledge on building agents and just using AI in general then this is the spot to be I'll leave a link to join in the description in the topend comment if you want to start crafting instructions and to do it the right way I highly recommend a tool a comfortable workspace uh
to start crafting instructions in so I recommend download loading Visual Studio code to your desktop it's free and it's going to give you a nice environment to work in in order to craft your instructions so this is what I'm going to be using today when I'm showing you how to craft your instructions and I highly recommend you do the same because it allows us to easily save the coding files to our desktop the XML files to our desktop whatever we need within our instructions and it's just the best spot to in my opinion write these
things now when you open open vs code this is the dashboard that you're presented with what I'm going to do anytime I have a new agent this is what I like doing an agent that requires deep detailed instructions I like hitting new file and then typing in whatever I want my file name to be so for this I'll do example- instructions and then I don't just want to hit create new file but what I want to do is type in dot XML and then I'm going to hit enter and then it's going to ask to
save it somewhere on your desktop so this is where you could have a folder hierarchy and really get organized what I have is a folder in this example tutorial called instructions for agents and then I have projects within my agents maybe you have multiple agents within a project that might be a cause for multiple folders within this hierarchy here so I'm just going to save my instructions to this example folder that way I can always come back and edit them in the future I'm going to hit create file and now have a nice dashboard to
work with here now I'm going to import an example of the outline in the structure of our instructions so this is an outline of what XML styled instructions look like now you may be a little nervous here you may be a little confused if you've never seen this before but I'm going to break it down for you with an analogy so think about if you are moving houses and you're packing everything away in boxes usually if you're packing certain items within a specific box you label the outside of that box that way you can pick
up the box and move it into the room that you want the materials in or the contents of the Box in so if you're moving out of a house maybe you put all of your kitchen utensils in a box and you label it utensils that is a similar structure to what we're doing here every instruction that you give your agent every goal every idea is labeled just as you would label items within a box let's say if you were moving out so as as you can see let's just break this down from the highest level
whenever you want to open a box right whenever you want to start a new section what you do is you have an open angle bracket and then a closing angle bracket so whenever I want to start writing agent instructions what I want to do is I want to have this open angle bracket and then label it agent instructions and then have that closing angle bracket this opens up the tag so this is like me opening the box and saying okay I'm am inviting new contents within agent instructions whenever you're done with something whenever you're done
with items in a box what you can do is you can close it off so we have all of this content within agent instructions this is all the items within the box and as you can see we're going to have boxes within other boxes as well nested boxes as we can call them if we are going to keep on going with our uh moving out analogy but when we're done with everything in agent instructions we then close it off so this is like opening the box putting things in and then we're closing the box down
here at line 45 so whenever you open a tag this is what it's going to look like Open Bracket close bracket and these are just like angled Chevrons or greater than or less than signs so this would be me opening the box closing the Box you want to do the same thing except before your box label what you want to do is you want to put a forward slash after the first open angle bracket so whenever you're closing a section or whenever you're closing a box or whenever you're closing a label or a tag you
know all these words can be used interchangeably this is the structure for closing it so now the agent's going to read this is the start of agent instructions this is the end of agent instructions so everything within is the agent instructions that's the idea of the setup and we're just putting boxes within boxes within boxes and that's the aide of XML computers love reading this and humans can easily write it so any time that I create an agent and I have instructions the main things that I'm putting Within These instructions no matter what are number
one a roll you have to clearly Define the purpose of your agent and why it's here with a name and a description and then you're giving it a goal you can have primary goals you can have secondary goals you can have tertiary goals it really doesn't matter but I recommend keeping it to you know maybe just one to three goals because you don't want to flood your agent with multiple ideas for what it needs to complete usually One agent needs to be good at doing one thing very well so we have role that's an absolute
necessity goal absolute necessity and the last thing that's absolutely crucial for your AI agent to perform successfully is instructions and what you want to do is you want to label your instruction one at a time so as you can see I have opening the instructions box and closing the instructions box what I want you to notice is instructions is plural here so we can have within here multiple different instruction types right so maybe I have three four five different instructions that relate to achieving my goals but as you can see we're just using the opening
and closing tags for every piece of information that we have in here so if I'm creating instruction I'm going to make a box for instructions and then I'm going to label the items within instructions with an open tag and a closing tag so these three things are the absolute necessities of creating instructions with your AI agent especially within this XML format XML is the format we need in order to get the AI reading our instructions perfectly every single time but these three items within your box are absolutely necessary now let's get into the next thing
because the next thing is very very important as well right here is where you're giving examples and these examples I like adding three examples of how I want my agent to Output information to me that way it can see multiple different user queries and how it should respond to those queries so you can add in more if you'd like but I just recommend starting with one to three examples now this setup what it does is it allows for user input examples and agent output examples and this is where you're going to need to come in
basically give the agent how it should respond to certain queries and what I mean by query is a question or statement made by you to your AI agent or the information that it's going to receive and then the output you want based on that information it's receiving and I like doing that three times with kind of different examples so it can see how to handle pretty much every situation maybe you want to add two examples for each scenario you know and have six agent output examples so it knows how to handle certain situations what to
say in certain times and guys you can leave a lot of this up to AI that's the beauty of AI is it can pick up on patterns quite quickly a lot faster than a human can so just providing a couple of examples to start is good and that's what I recommend with everyone if you're creating these instructions like don't go over the top and create a thousand lines worth of instructions start small make it precise make it absolutely necessary every piece of information in there try not to add too many filler words want to be
concise when giving instructions as well so let's start filling this out maybe what I want to do is I want to have a form input and I want to be able to type in a certain topic that I want researched with my AI agent now we're not going to be developing any crazy agentic research workflow multi-chain summarization type of agent but just for the sake of example let's say I'm creating a research agent okay let's start from the top now I recommend writing these out for the first time yourself because it's really really good to
know how to do it and once you understand what you need and the terminology that goes into it and also how to write it it becomes much easier to use AI to craft instructions as well so I'll leave this template for you down in the description but I highly recommend just following along and creating instructions yourself that way you understand how to type it out don't get lazy this is where you know you separate from being a okay agent Builder to being an awesome agent Builder because it looks professional it is professional and it's just
the right way to go about it so let's just take it from the top I'm going to label this agent instructions and this is my open tag so at the very end of our instructions We'll add our closing tag we're going to close the box to all of our stuff so whenever I'm creating nested elements within my XML I always hit enter and then tab over so that we can keep that organized structure so first we need to create a roll so I'm going to make a roll and what's going to be within that Ro
well there's multiple elements within a ro like name and description so I want to hit enter tab over and then I want to start a name tag I can enter down and then I can also start a description tag and now I can enter down just once and then hit backspace and I can close off the Box Roll because everything within roll is name and description now I don't want to get two ahead of myself but you can kind of predefine your structure so I'm going to do this and type out roll and now I
can give a name and description so what's the name going to be well I could just name this something like research agent you might want to get a little bit more creative but then I'm going to close it off with this structure so I am remember opening name giving name closing name just as I'm opening roll giving elements of roll and then closing roll and now we're at the description this is where you want to be hyper specific and very concise about what your agent does what are the purpose of these instructions and you want
to keep it like one sentence of you know give me your elevator pitch of what does this agent do except instead of 30 seconds we're talking about one sentence so for this agent I'm saying create structured research papers from simple user queries I can then close off that description beautiful so now we have our role complete and remember what the three biggest things to put in your instructions are role goal and instructions so that's what I'm going to do now and when I have that complete I'll report back to you I'm going to use this
exact same format so I've just got done creating some instructions and a goal my goal is very similar to the description it's clear it's concise that's all we need I say create structured in-depth research papers from simple user queries posted by the user and then I get into my instructions and I have four different steps within my instructions it's a way I like going about things like step one or instruction one do this so I said step one review the user topic selection which we're going to get into an example flow in nadn and actually
use these instructions as well step two create five subcategories of topics that relate to the main topic so I'm going to be sending in a main research topic and I want it to create five subcategories in order to do even deeper research step three do in-depth research on the main topic and each category so I could say each subcategory to be more specific and then step four follow the and then I call something called blog template format and create a blog post because you don't have to stay within the bounds of role goal and instructions
you can add even more tags up here in agent instructions before you get into the input output examples so maybe I want to create a blog template for it to follow a structure that I want it to follow for each blog post and I can call that tag or that label within my instructions in order for this agent to go okay now it's saying follow the blog template and when you have it in this format it's really good at calling different parts of your instructions so now we can create a blog template format in a
similar style to this so I created a Blog template you know I just opened my box and then I closed my box and the blog template is with inside that box now I just realized I did not finish this blog template so I'm going to create a final section I can just say uh final thoughts and then I can say say conclusion of research and so basically I just give a template of how I want it to write this research template I changed five subcategories to three subcategories and now we can get into our user
input agent output examples now that we have the main structure of what we want this agent to do the output format that we want it to follow so what I've done is I've just created the structure we're going examples example and then we get into each individual example so I'm creating a box for examples and then I'm having three different example tags within each uh examples box so we're just going into example then user input and then agent output and then we close that example so now I'm going to fill all of these out I'm
not going to spend much time on this because I do want to get to that Json output structure of this video and also testing this with our AI agent but I am going to fill these out to show you examples of how you should create your examples so just to show you an example of what this looks like I'm only going to use one example right now so we can get to those Json outputs but you just want to keep this process repeating the more examples you add usually the better it responds but again you
don't want to overload it I recommend you know two to five examples as you can see I have an example here of a user input I have maybe the user wants to do more research on Photo biom modulation which photo biom modulation is just a fancy way to say red light therapy but maybe the user wants to do more research on red light therapy and all they input is photob bio modulation that's what the agent receives and you want to give an accurate example of exactly how it's going to receive that data as well so
maybe I'm just doing raw text but then I give an agent output right using my blog template and I'm not like labeling things here I'm just using what the blog template looks like so that it understands based on a user query how a good response actually looks and I used chat GPT to generate this and then I modified it a little bit myself but what I want to do right now is I want to end these instructions okay so I'm going to keep that agent output right there but then I'm going to close off with
the entire agent instructions so we have an input output example and now these XML instructions are at a level where we can actually use them within our AI agent and we're going to be able to expect very consistent results because we've structured it that way so let's dive into n8n and I've already got an agent set up to where I can just input these instructions and then let's test some basic queries to see if it's following our blog template and all of our instructions here so let's do that now by the way I just want
to mention something make sure that you close out examples and the example itself so I can come down here and actually close out example and then I can go one step even further back and close out all of the examples so if we ever want to add in more examples we can just come in here and then add an example on before the actual examples close tag but now we are ready to go so when your instructions are ready you can within uh vs code just click in anywhere and then hit control a you can
do this on any app actually and that'll copy all of your data then you can hit contrl C and that copies it to your clipboard and I also recommend always saving these files so you can hit controls and save and it will save to that spot where you uh put this file on your desktop so I have a very BAS basic example of an AI agent here and let's test out these XML instructions now you might be wondering where are we putting these now most people won't understand and they'll just put them in the prompt
user message but this is where your user queries are going to come in so whenever I'm talking about like a user query this is basically like uh where your message is coming in from the user so right now I just have a chat node connected to this so I can chat with NN and I can really just say anything I want and when I send that off that's my query so I'll be typing in my research topics down here but I could have a form hooked up to this really whatever I want so let me
dive into the AI agent and where we want to add our instructions are within the system messages now some models will not allow for this like the 01 model I don't think allows for it at the time of this video but if I select system message and I'm just using something like GPT 40 you know I could even bump that up to 40 if I wanted to it's going to be able to accept this so within our AI agent in the system message you can just paste in that prompt and I recommend switching it to
an expression that way you can pull up this bigger window so you have more room to work with but I'm going to paste in all of our instructions and now our AI agent is going to receive these instructions before it ever receives our user query as you can see if I hover over this Mark right here on system message it says the message that will be sent to the agent before the conversation starts so think about how powerful that is the agent has all of this context before it ever receives any query this is just
a basic example of crafting instructions with XML in a structured way we've given our agent a name a description instructions templates to use and an input output example that it can also work with so this is what our input output example looks like we have our title our abstract and then we have our main topic and then all of our subtopics beneath that main topic and then a final thoughts conclusion as well so that's how every result should come out so I'm going to close out of this now and then we can actually test this
out so I'm going to chat with this and then I can just type in any topic maybe I just want to type in water only fasting for 6 days maybe I want to get a little specific I can send that off and what you're going to see is I don't have my credentials set up so I'm going to go do that now and then we'll test it out okay I have my new open aai account connected now let's dive in and let's try out this prompt one more time water only fasting for 6 days and
send it off and it's now going to run through that with our system message and it should give us a response similar to our agent output example and use our research blog template I'm going to wait for that response to come in and there we go that was fairly quick it gives us our title our abstract our main topic with two paragraphs like we asked our subtopic another subtopic and one more subtopic and then also we have well another subtopic actually as well and then we have our final thoughts so this follows the exact template
that we gave it which is very very cool now we could read this we could perfect it but this is the idea of XML instructions is we literally give it such a systemized way of completing and handling user queries or user data that enters the AI agent that it knows what to do every single time so now all I do if I ever wanted these types of post generated is I could just come in here and type out the topic that I want researched that's the idea of crafting instruction with XML is you can get
very consistent predictable outputs but the next thing we need to learn is Json because right now within our AI agent all we get in the output section is one agent output so we can extract the title we can't extract the sections we can only extract the output as a whole so if we were entering this into a database and we wanted to extract titles for our post and then the body content we couldn't do that right now right because all we have is the output and that would pull in the entire blog template so what
Json is going to allow us to do is actually structure these outputs in a way that allows us to pull multiple parts of this output uh and structure them in whatever way we want with multiple Fields so we're going to learn how to create that Json now within our AI agent so basically we can add Json within our XML instructions XML is just the way that you format your instructions now Json is something you can add in order to create objects from your instruction so we can leave pretty much all of this the same we
just want to structure our data different and also add some terminology within the system message which is our instructions that makes it understand that we want a Json object to be output so Json is just a way of structuring your data that allows you to pull multiple Fields let me show you a few changes I would make in order to create these structured outputs and if you want to go deeper on Json I've got a module within our community that goes about an hour long talking just about Json outputs and just about XML instructions so
I recommend going and checking that out if you want to go deeper but for now I'm just going to show you the idea of it and how we can start creating multiple outputs so right up here in our primary goal I would probably want to mention that this output needs to be in Json object format that's actually you know a really good practice is labeling how you want your data outputed so at the end here I'll just say in Json object format beautiful so that's the main goal now it knows that its goal is to
create these in-depth research papers in Json object format I can keep my blog template the same but the main spot you want to change your formatting in is within your agent output and I love using AI for this so what I can do is I can give my blog template I can give my agent output and say create this and make it a Json object so let me show you what I mean by that so what I want to do is I just want to go to something like chat gbt I'm using the new 03
model you could really use whatever you want for this I'm sure GPT 40 would work fine even 40 mini but what I say is break up this paragraph into Json object format and then I say creating Json fields for each section of the research paper based on my blog template now in this part of the instructions crafting you want to really think about what variables do you want to be able to pull directly maybe you just want to be able to pull title and then the body of that research paper maybe you want to be
able to pull each section individually and map it to a new field in a database or have a switch module that puts data going down different routes whatever your use case is for creating multiple outputs from a single input that's what you really want to think about right here so now I can just Define which variables I want I can say I want to pull the title and then I can put title in apostrophes so it knows that that's important and body content from this paper every single time here's my blog template and I Define
that right here so now I can paste in my blog template just like this maybe I grab it even with the tags and again like I said you could use whatever you want for this and then I could say uh here is my output example make this output example in the new Json object format with the two variables of title and the rest of the agent output in a body content variable beautiful so now I can paste in my agent output example and this is just an example of how I create Json now I go
more in depth on it in that module if you're really interested but now what I can do is just kind of copy this entire agent output example actually and hit contrl C now I can go back and then I can paste in that so now it's going instead of putting this in just regular text it's going to put it in a Json object and use the variables I want to extract and Define those so I can send this off and now it's going to get to work at this point in my you know Json object
creating career I guess I like using AI in order to help me because it just helps with speed and I don't really have to think about it I know the questions to ask I know the terminology to use in order to help give AI context into what I want it to do beautiful so now each one of these are actually individual Fields so now whenever I output something within my AI agent instead of it just being one output that I can pull and grab and use elsewhere I'll be able to pull title I'll be able
to pull body content abstract main title topic main topic content and I could even put this in just two Fields I could say you know ditch abstract ditch this entire array I just want title and body content I can just say I just want title and body content is key value pairs get rid of the array then I can send that off because right now I just want title and body content when you start pulling arrays into things like an a table database it starts to not be able to accept that data it expects you
know just strings of data so it shouldn't be that difficult maybe I used honestly too complex reasoning of a model but now I got what I want these are the variables that I'm going to be able to pull with each output title and body content so now I can copy this and now I can go put this within my agent output example instead of just the raw text because now we want Json object output so I'm going to delete all of this and then I'm going to paste in the new Json object format because this
is the way that we want it uh you know outputting all of our information so now that we gave this example each time it outputs it's going to Output an a Json object type title body content as my keys with these updated instructions what it's going to allow us to do is extract multiple variables from One agent output so let's copy our new instructions with this Json object format uh word here you know even just saying this is going to change your instructions these three words saying that I want it in Json object format completely
changes my instructions so now I'm going to hitr a contrl c so before we actually get into pting in these instructions that we have I wanted to show you an example of the problem that I'm dealing with so I'm going to connect this AI agent to an air table database and let's say we want to connect to this air table database where I want to put the research topic as a title and just extract the title over here and then extract the entire research body over here right now I wouldn't be able to do that
with my current instructions because I can only map one output I can't extract the title unless I do this Json structured output parsing so for example I have values that I want to put in here to send to my database like research topic and research body is shown in this template right here but I can't because I can only put in one entire output right and it brings in the title and then the entire research body so I'm going to delete this and we can fix that problem by creating a structured output parser and by
adding in our new instructions so first I'll just add in the new instructions so I'm going to hitr a delete all of those and then paste in our new instructions and remember our new instructions in the agent output example are showing how this Json object needs to be formatted with title and body content so what we need to do in nadn in order to get structured outputs is we need to Define that yes we want structured outputs so in order to do that we have to hit require specific output format so when I select that
it says connect an output parser on the canvas to specify the output format you require so this is where you're going to specify which variables you want to be able to pull you could pull in as many variables in whatever format you'd like but the idea is to give it an example of the Json objects you want to pull in that way it knows how to Output its answer so it doesn't just do one output but it can handle multiple so I'll select this output parser in Orange here and then I'm going to select structured
output parser right here here is where you want to give a on example of what your data is going to look like coming in so we can just paste in our agent output example that we have within our code so if we go back to our agent output example here you're just giving an example right of what is your Json going to look like what are the variables going to be called and this is where you want to label your variables so title and body content it doesn't really matter what's in the text you just
want to give it an example of the structure you know I don't have to have anything in body content except one word if I don't want I'm just saying I want title pulled in I want body content pulled in I want them in two separate fields and that's what I'm doing here is I'm just defining the structure of this output parser but now that we have everything connected right we have our new instructions in saying that hey we want the main goal of this to uh parse out our in-depth research papers in the Json object
formats and also we're giving our Json object format in an agent example then what we can do is actually test this step because as you can see right now we only have one output okay but watch what happens when we test this step when I hit test step what you're going to notice is it pulls in now two Fields it's not just pulling in output but it's going to pull in title and body content that way we can use both of them and map them within our air table database so I'm going to wait for
this node to execute and now as you can see we have two fields that we can map from one AI agent output so if I go to air table now what I can do is I can map these fields to their respective column we don't just have one output but we can map the title and AI will extract that title for us and then we can map the body content now when I hit test step what's going to happen is it's going to populate my a table in a nice organized fashion so now we have
our blog template in here and all we needed to do was create a Json object format within our instructions and you really don't even need instructions uh in order to accomplish this we're just trying to maintain instruction principles but all you really need to do do is require your specific output format and then give an example of what you want your Json to look like and then that will give your AI agent multiple outputs to to fulfill that way it's not just outputting one item each time so now I could go in here and I
could chat with this I could delete this empty record for my air table and then I could just you know list out whatever I want so we could do the photo biom modulation example again just like that I could send that off I could even get another topic ready in here maybe I want to do something like gut mic biome and I can research that topic after photoi modulation is done but as you can see it's going to populate into our air table database in the structure because we've defined we want research topic and research
body these are two separate Json objects that we wanted again I could have 20 columns in here and have 20 different variables pulling in from one AI agent run one AI agent output this node was executed successfully so now when I go back to the research database as you can see it follows that exact structure with our title and then the body of that research maybe I try gut microbiome what I want you to notice here is it's just consistent I know exactly what I'm going to get each and every time I run this because
I set up instructions in order to take the stress off my back so now when I go to the research database it's the exact same right and the research body is the exact same with the with the abstract the main topic and then the three different subcategories as well one 2 2 three so this has been how to craft instructions for your AI agent I'm telling you if you stuck it out to the end actually good job not many people want to stick through a video where you're just literally writing instructions and words down for
your AI agent but for those who do you're a special breed and you understand the importance of explaining things to a large language model now if you can't accurately describe the job that the large language model is filling then the large language model isn't going to accurately give you a response you would like so if you want to keep learning in depth again I highly recommend joining the AI foundations Community we are growing by the minute and this group will not be open at the price it is now forever so I highly recommend you get
an early to lock in your price that's not fomo that's just the truth and we truly care about each and every one of you in here make sure when you do join you introduce yourself as you can see that that little three dots means somebody's introducing themselves right now uh somebody actually just introduced themselves so we love welcoming new members and we hope to see you in there if you like learning and improving your AI agent building skills or just your AI knowledge in general then we'd love to have you so that being said that's
all I have for this video If you enjoyed please like And subscribe I would highly highly appreciate it now go out there craft your instructions for your AI agents and tell your friends and with that being said I'll see you in the next video
