How to Create and Use Perplexity Personal AI Chatbot Agents! #95

Summary

In this tutorial, Josh Evilsizer demonstrates the utility of Perplexity's AI agents, which are essentially pre-programmed chatbots designed for quick and efficient responses. He guides viewers through creating, editing, and using these agents to save time and enhance prompt efficiency. Examples include a Spanish proficiency agent for language practice and a data analysis agent for generating statistics. The video emphasizes the benefits of reusing and refining prompts for superior results, encouraging viewers to streamline their interactions for increased productivity.

Takeaways

• 😀 Perplexity agents are reusable, refinable, quick access prompts saved as collections, which can be thought of as pre-prompted chatbots designed for specific tasks.
• 🕒 They save time by allowing users to perfect their prompts and streamline chatbot interactions through repeated use.
• 📚 Users can create collections by clicking on the 'plus' sign in the library and filling out a title, icon, and description to make them easily accessible and understandable.
• 🔧 Editing a collection is straightforward and can be done by accessing the collection and using the 'edit' option, allowing for continuous refinement of the prompts.
• 🗣️ The video provides a step-by-step guide on creating a 'Spanish proficiency agent' to help improve Spanish language skills through a structured conversational prompt.
• 🌐 Examples are given to demonstrate the versatility of perplexity agents, including creating a weather forecast agent and a video brainstorming agent.
• 📊 For more complex tasks, such as data analysis, agents can be created to perform initial descriptive statistics on datasets with specific focuses, streamlining the analysis process.
• 🔄 The 'rewrite' function in Pro mode allows users to get different answers from various chatbots, enhancing the versatility and effectiveness of the agents.
• 🔄 Users can reuse the same prompt multiple times without worrying about token limits or context windows, ensuring consistent and efficient interactions.
• 🏆 The overall benefit of using perplexity collections is the time saved and the improved results achieved by reusing and refining prompts for various tasks and scenarios.

Q & A

What are perplexity agents and how do they function?

Perplexity agents are reusable, refinable, quick access prompts saved as collections. They function as pre-prompted chatbots that are springloaded to respond to specific needs or tasks repeatedly, allowing users to perfect their prompts and save time through streamlined interactions.

How can one create a perplexity agent?

To create a perplexity agent, one must navigate to the 'Collections' section in the library, click on the plus sign to create a new collection, provide a meaningful title, an icon, and a description that includes steps for future reference. Then, write the prompt that defines the agent's function and save it.

What is the purpose of the description field when creating a perplexity agent?

The description field serves as a helpful guide for future use. It allows users to document steps and instructions for themselves, ensuring they remember how to use the agent correctly, even after some time has passed.

How does the Spanish proficiency agent work as demonstrated in the script?

The Spanish proficiency agent is designed to engage in beginner-level Spanish conversation. When prompted with 'Ola', it responds with a question in Spanish. The user is expected to answer in Spanish, and the agent then critiques the response in English, providing feedback to help improve the user's Spanish proficiency.

What is the significance of the 'Al Roker' example in the script?

The 'Al Roker' example illustrates a simple perplexity agent that provides a whimsical weather forecast when prompted with the word 'high'. It demonstrates how agents can be tailored to deliver specific information in a personalized manner.

How can one edit a perplexity agent after it has been created?

To edit a perplexity agent, one returns to the 'Collections' section, selects the agent for editing, and uses the 'edit collection' option. This allows modification of all fields initially filled out, including the title, icon, description, and prompt.

What is the benefit of sharing perplexity agents?

Sharing perplexity agents allows others to access and utilize the same prompts, potentially streamlining their interactions and saving time. It also enables collaboration and the sharing of effective prompt strategies.

How does the 'video brainstorming' agent assist in content creation?

The 'video brainstorming' agent aids in content creation by generating video ideas based on a given topic. It takes on a specific persona and follows a structured approach to produce an outline, which includes the 'what is it', 'how does it work', and 'so what' aspects of the video.

What is the 'rewrite function' mentioned in the script and how does it work?

The 'rewrite function' allows users to request a different response from a different chatbot model without having to rephrase the original prompt. It's particularly useful in Pro mode, where users can choose from various chatbot models like GP4, Claude, or Gemini Pro to generate a new response.

Why should one consider using perplexity collections?

Perplexity collections should be used to save time by streamlining chatbot interactions and reusing, refining, and perfecting the most often-used prompts. This approach is akin to building a bridge once and benefiting from it indefinitely.

Outlines

00:00

🧑‍🏫 Introduction to Perplexity Agents

Josh Evilsizer introduces the concept of Perplexity agents, which are reusable, refinable, and quick-access prompts saved as collections. These can be thought of as pre-prompted chatbots designed to streamline interactions and save time. He guides viewers through creating a collection, providing a title, icon, and description to make future use intuitive. The example given is a 'Spanish proficiency agent' aimed at improving Spanish language skills through a structured conversation with the AI.

05:01

🔧 Editing and Using Perplexity Agents

The video demonstrates how to edit existing Perplexity agents by accessing the library and collection, then using the edit options. Josh also shows how to delete collections if needed. He provides examples of different agents, such as an 'Al Roker' agent that provides a whimsical weather forecast when prompted with the word 'high'. The detailed prompt for this agent includes a request for a fun and whimsical delivery style along with a famous quote.

10:03

🎯 Advanced Prompts and Rewriting Features

Josh explores more advanced uses of Perplexity agents, such as 'video brainstorming', where the AI generates video ideas based on a given topic. He emphasizes the importance of refining prompts over time to achieve the desired results. The video also highlights the 'rewrite' function, available in Pro mode, which allows users to generate different responses using various chatbot models. This feature is particularly useful for obtaining multiple perspectives on a given prompt.

15:03

📊 Data Analysis with Perplexity Agents

In this section, Josh applies Perplexity agents to data analysis tasks, showing how to use them for initial descriptive statistics on variables from a CSV file. He uploads a dataset and prompts the AI to focus on specific aspects, such as political leanings, to quickly parse and analyze data. The example illustrates how agents can save time and streamline the process of dealing with datasets by automating repetitive analytical tasks.

🏆 The Benefits of Using Perplexity Collections

Josh concludes by emphasizing the benefits of using Perplexity collections, such as saving time through streamlined chatbot interactions and the ability to reuse and refine prompts for improved results. He likens it to building a bridge that provides ongoing benefits. The video encourages viewers to try using Perplexity collections and to share their experiences in the comments. Josh also reminds viewers to like, subscribe, and share the video, and assures that he will respond to any questions left in the comments.

Keywords

💡Perplexity

Perplexity in the context of the video refers to a tool or system that helps in creating and managing AI chatbot interactions. It is used to streamline and refine prompts, making it easier for users to interact with AI agents. The video's theme revolves around teaching viewers how to use Perplexity to save time and improve their AI interactions, as exemplified by the creation of 'Spanish proficiency agent' and other AI chatbot agents.

💡Agents

Agents in this video are AI chatbot entities that are pre-programmed with specific prompts to perform tasks or answer questions. They are likened to 'pre-prompted chatbots' that can be quickly accessed and reused. The video demonstrates creating an 'agent' for Spanish proficiency, which is designed to engage in beginner-level Spanish conversation and provide feedback.

💡Collections

Collections are a way to organize and save reusable prompts in Perplexity. They act as a library of pre-set AI interactions that users can access and modify as needed. The video emphasizes the efficiency of using collections to save time and improve prompt quality, showcasing the creation and editing of a 'Spanish proficiency' collection.

💡Prompts

Prompts are the inputs or questions provided to the AI agents that guide their responses. The video focuses on the importance of crafting effective prompts and saving them as agents within collections for future use. An example from the script is the prompt for the Spanish proficiency agent, which instructs the AI to respond with a beginner-level Spanish question.

💡Refinable

Refinable in the video script refers to the ability to improve and adjust the AI prompts over time. This is showcased through the iterative process of creating the Spanish proficiency agent, where the presenter refines the prompt until it works as intended, highlighting the iterative nature of prompt development.

💡Quick Access

Quick access in the context of the video means the ability to rapidly retrieve and use saved AI prompts or agents. This is a key feature of using Perplexity collections, as it allows users to efficiently reuse prompts without having to recreate them each time, as demonstrated by the ease of accessing and using the 'Al Roker' weather agent.

💡Iterate

Iterate in the video refers to the process of revising and testing AI prompts until they produce the desired outcome. This is a critical part of creating effective agents, as seen when the presenter goes through multiple iterations of the Spanish proficiency agent's prompt before achieving a satisfactory result.

💡Sharable

Sharable in the video indicates the ability to share the created AI agents with others. This feature is mentioned when the presenter chooses to make the 'Spanish proficiency AI chatbot agent' sharable, suggesting that users can collaborate or distribute their agent creations.

💡Rewrite Function

The rewrite function mentioned in the video allows users to generate alternative responses from different AI models. It is a feature of the Pro mode in Perplexity, used to refine and diversify the outputs of AI agents, as illustrated when the presenter rewrites the 'video brainstorming' agent's response using different chatbot models.

💡Context Window

Context window in the video refers to the limitations in the amount of context an AI agent can process at one time. The presenter mentions that using Perplexity agents avoids concerns related to tokens or context window, implying that agents can operate independently of such constraints, allowing for more flexible and extensive interactions.

Highlights

Introduction to Perplexity agents as reusable, refinable quick access prompts saved as collections.

How to create a Perplexity agent with a step-by-step guide.

The importance of giving meaningful titles and icons to Perplexity collections for quick identification.

Utilizing the description area to provide helpful instructions for future use of the agent.

Creating a 'Spanish proficiency agent' to improve Spanish language skills through interactive prompts.

Demonstration of how to use the created 'Spanish proficiency agent' with a live example.

Editing and refining a Perplexity agent to improve its functionality over time.

The option to make Perplexity collections sharable or keep them secret.

An example of creating an 'Al Roker' agent for fun and whimsical weather forecasts.

How to use the 'Al Roker' agent to get a personalized weather forecast with a famous quote.

Advanced use of Perplexity agents for video brainstorming with a detailed prompt structure.

The ability to use the rewrite function in Pro mode to get different answers from various chatbots.

Saving threads as new agents for quick access to previous queries and their results.

Using Perplexity agents for data analysis by setting up a prompt for initial descriptive statistics.

The efficiency of reusing the same prompt multiple times without concerns about tokens or context windows.

The practical application of Perplexity agents in a business or data analysis context.

Why using Perplexity collections can save time and improve the quality of prompts over time.

Encouragement to share thoughts and experiences with Perplexity agents in the comments section.

Transcripts

hi my name is Josh evilsizer today I'm

going to show you how to use perplexity

collections as agents are you watching

the Right video well do you use

perplexity would you like to save time

would you like to improve your prompts

then yes questions answered in today's

video perplexity agents or collections

what are they how do you use them how do

you create them how do you edit them

I'll provide some examples and then

we'll end with the most important

question why should you care let's jump

right into perplexity doai here we are

what are perplexity agents they are

reusable refinable quick access prompts

saved as collections think of them as

pre- prompted chatbots springloaded to

respond to whatever it is you need or do

over and over again they allow you to

perfect your prompts and save time

through streamlined chatbot interactions

but enough talking

let's let's make one shall we so I click

on library to move to move to the place

where I need to create a collection and

here we are and you'll see up at the top

Collections and so I'm going to click on

the plus sign to create our first

collection and it's a bit of experiment

so we're going to see what we get today

on live YouTube recording it's not live

you're you're watching it recorded

whatever so here's the title uh we're

going to title it Spanish proficiency

agent so give it a meaningful title that

helps you when you're looking at it what

is this thing more meaningful is the

icon give it a emoji that resonates with

you because again the point here is to

be able to create or use these very

quickly and so visuals that resonate

with you that make sense hey this is

what I use this for right make sure your

visual makes sense is all I'm trying to

say there

description all right this is where we

can get very helpful for future us so

sometimes we'll create tools and a

couple days will go by and we'll forget

how to use them or exactly how to use

them so there's a lot of space in the

Des this description area so I could

have just left it at Spanish

conversationalist right here but I went

ahead and added steps telling future me

hey these are the steps that you should

follow to use this agent and you can see

that says type Ola hit enter so I'm just

helping out future me hey this hey hey

dummy this is how this chatbot works all

right so the prompt in this instance

we're going to try and create a

conversationalist that helps us um

improve or increase or adhere to our

current level of Spanish speaking

proficiency and so what have I written

here my name is Josh when I type Ola you

respond with a question in Spanish that

someone proficient in Spanish at the

inter agency language round table or ilr

level one could understand so speak to

me in beginner

Spanish I will then attempt to answer

your question in Spanish you must then

respond by critiquing my response using

English this time and providing feedback

in English to help me improve my Spanish

proficiency I think that's it so this is

our prompt so we're going to type Ola

hit enter and and then it's going to do

this if we've done this correctly or if

if what we've done will actually work

I'm going to leave it sharable you do

have the option to leave it secret right

here if you wanted to I share these a

lot so I just often leave them sharable

going to go ahead and crit crit hit the

create

button and you can see right over here

our Spanish proficiency AI chatbot agent

has been created so we're going to click

on it just to kind of show you how it

works here it is and I'm calling this

The Prompt prompt because we already

have a prompt and I'll show you where

it's at so over here on the right hand

side we have this helpful description

Spanish conversationalist and again the

steps hey Josh type Ola hit enter and

then the AI prompt that has been applied

is what I just wrote out and it's right

here and we're we're going to go ahead

and execute

so hola enter

let's see what

happens all right Kostas all right so I

need to

respond in good

Spanish I'm getting the feedback that I

requested it worked uh I'm just going to

be fully transparent with you iterations

one two and three did not work this is

iteration five uh so four and five have

worked uh in any event uh it just goes

to prove that um if iterating on your

prompts which is what I did did to

arrive at this this prompt that

works it proves the point of why I'm

creating these things anyhow enough

padding myself on the back we're going

to go ahead and move into how we edit

these and some other examples uh so use

I've kind of showed how to create and

use use at the same time

here create went over that and then edit

so if I wanted to

edit this

agent I just again library and then the

collection and at the very top these

three dots right here edit collection

and then all the fields that I filled

out earlier are right here uh ready to

be edited so it's it's pretty

straightforward and if I wanted to

delete it same thing delete collection

so there you go there's the Soup To Nuts

and now I'm going to go over some

examples so if you got what you need

have a nice day if you want to see some

examples hang out all right so the first

example we're going to use here is

another simple one

and I call it the Al Roker and if you're

old like me you'll this will resonate

and the prompt prompt is the word high

and so I type the word high in here and

Al Roker is going to give me the weather

hello there Josh today in Columbus the

skies have decided to put on a bit of a

show with clouds performing their best

dance moves across the Horizon expect a

cozy range blah blah blah blah pause

your screen and read it if you'd like to

uh ended with a great quote by Henry

Ward beeer why did this happen well

here's what my prompt looks like so we

click on

library and we click on the aloker

prompt and then we can read the AI

prompt applied which is my name is Josh

evilsizer I live in Columbus Ohio when I

enter hello or high respond with the

weather forecast for today the high and

low temperatures and the next day that

I'm likely to see Sunshine because I

live in Columbus Ohio and it's

wintertime provide this information in a

fun and Whimsical manner and with a

famous quote that is related to all the

preceding information so that's why you

saw what you saw and again just like all

the other agents that I've created or

showed you thus far uh I gave future

Josh some instructions to REM to remind

me how to use this so I call it weather

and wisdom and then type High hit enter

to make it work and that's what we did

all right so now we're going to take a

step forward into a little bit more of

an advanced prompt sorry

agent

and this one and I keep moving I'll just

stay over here for a minute uh video

brainstorming so you may have seen me

use an example like this before with

Bard now Gemini uh similar but it works

just a little bit better here and I'll

show you why video brainstorming okay so

the description generate a video

generate video ideas based on a given

topic and the instructions from past me

to Future me here is enter topic hit

enter everything else you see there is a

cheat sheet for this videos so just

ignore that all right so we're going to

enter a topic and then hit enter and

I've not shown you the AI prompt that

would be helpful um so this is rather

lengthy um the gist here is take on the

Persona and I'll just read it quickly

and I will let you know all the prompts

I've used here they're in the Google Doc

you saw at the beginning of the video

which is linked in the video description

below take on the Persona of an

efficiency focused productivity expert

enus and Enthusiast who through the

creation of YouTube videos enjoys

teaching others how to employ Simple

Technology processes and habits to

maximize efficiency and Effectiveness

and make life a little or a lot easier

whenever possible your videos focus on

helping people use technology and

processes to complete knowledge work

related tasks most efficiently and

effectively so it's a little bit about

me and my Approach then the structure of

the video which is the what is it how

does it work the so what with

descriptions about all of those and then

your role me telling the AI a agent or

or chatbot rather Your Role so your role

is to produce an outline using the

script formula from above using the app

and any Associated processes techniques

or habits that I and we can't read it

because the screen is zoomed into to

much provide to you so there you go it's

a big

prompt and the whole point here is I can

come back and refine this over time to

get it dialed in to give me exactly the

results that I need and I can continue

to reuse it over and over again and

we'll show you how that works so we've

got this thing teed up I've got a prompt

for

it here it is the first prompt

is perplexity collections as quick

access prompts so kind of meta of course

and we hit enter and see what it gives

us I need help give me some ideas what

do you

got all right so here's the what is

it

here's the how does it work and I'm

often most interested in the examples

that it's going to give

me some real word examples these can

often be very helpful sometimes they're

not at all and what's really cool here

is when all this is said and

done we can use the rewrite function to

get a different answer from a different

chat box

unfortunately this is only available in

Pro

mode and so as soon as this is done

providing this example I'll show you how

the rewrite function works for

us and since this is taking man

this really comprehensive this time all

right I think we're almost at the end

there we go if I didn't like this I

could just jump in here and move to the

rewrite function so the default is

co-pilot I'm going to go smaller here

and with Pro I have access to

experimental gp4 Claude and Gemini Pro

so I could ask it to rewrite hopefully

you get the idea I just select one of

those and it's going to rewrite this

answer uh using one of those different

chatbots which is really cool or or I

guess what I want to also make sure I

have shared with you the whole point one

of the whole points of this is that we

can continue to reuse the same prompt

over and over again without bumping into

any concerns related to tokens or

context window so you see here I clicked

library and then I come over here back

to video brainstorming and I can start a

new and then if we scroll down we can

see the last thread that I was on so I

can either start a new here or I can

click any of the last threads and I can

simply select the edit query function

and if I didn't like the way this answer

or I didn't like the answer that was

provided here so I asked it perplexity

collections as quick access prompts I

could tweak that question to say

perplexity collections as prompts to

reuse and refine like it didn't pick up

on what I was trying to trying to do and

then I hit enter and of course I'm going

to get a a Rewritten thread uh based on

this new

prompt I'm not going to wait for this to

go through its whole Spiel there um what

I also want to highlight when I click on

the

agent or collection and I click on one

of these threads down here up here I

have the opportunity to rename that so

if I wanted to save it I could call this

um maybe the one I

liked uh click anywhere or enter and it

saves it and so now it's over here

accessible in the library and you can

move right to that answer each time all

right so I hit the rewrite function and

the edit query function and I talked

about just starting over as it relates

to reusing the same prompt over and over

again I want to give you one more

example let's say you are a business or

data analyst and you're dealing

with

Statistics metrics you kind of let's say

you're you're generally you have a lot

of data sets and you approach them in a

similar way over and over again this is

this is this is this is this example all

right so I'm just going to read this

stuff because it gets a little bit meaty

um but here it is so initial descriptive

statistics on the variables and here is

the

description initial descriptive

statistics on the variables and then in

parenthesis mean median min max

quantiles and a visualization of the

distributions of key variables uh and

then so the steps reminding me are

upload the CSV hit enter uh and then

we're going to use the experimental

model and I'm going to focus on

political iCal leanings so here we go

we're going to upload a CSV CSV file

it's got to be a CSV here it is

attach there it is so this is Brewery

data by state and their political

leanings and so like I said I want to

focus

on if I didn't have something for it to

focus on I could just call it Brewery

data and because we have to type

something right it's the prompt because

we've already prompted it down

here which I didn't cover so we'll do

that now provide the following initial

descriptive statistics on the variables

including mean median min max quantiles

and any other statistics that would

assist to better understand this data

set and the second piece a visual

visualization of the distributions of

key

variables the so what here is

again I analyze data all all day long

I'm making this up and I do the same

three step steps when I get a data set

and so by plugging these in as agents

all I have to do is upload this CSV file

enter what I want it to focus on if I

want it to focus at all or just give it

a name and then hit enter now it by

default it's going to give

me GP a result from

gp4 and so as soon this as soon as this

is

done we're going to pretend um that the

experimental model gives us results that

we like better so you can see here gp4

is the default and so in the rewrite

function I just come to the

experimental and you can see

here it would be a little tedious to

read through the stuff that it's giving

us but this has tended to be provide

better

results and so there you go dealing with

data hopefully you can see how a agent

could save you time to start parsing the

data in ways that maybe you are used to

chopping it up very quickly when you

begin with a data set in any event three

hopefully different enough examples to

help you see how a I almost said B agent

a perplexity agent or collection uh

helps you use and reuse quick access

prompts and save time uh reuse and reuse

and refine to get Superior results all

right as I I promised before uh why

should you care or the why should you

use a perplexity

collection I lost my my spot on my

keyboard sorry save time by streamlining

your chatbot interactions and reusing

refining and perfecting your most

often-used prompts to get better and

better results so anytime you find

yourself rewriting a prompt you've

written at least once before or taking

the time to craft a really great prompt

pause copy it and save it as a

collection why well it's just like a

bridge build it once and benefit

forever thank you for watching if I have

inspired you to try this please let me

know in the comments down below of

course please don't forget link goodness

in the description down below please

like subscribe share this with somebody

else that might find it useful or or

enjoy it as always if you leave

questions I will absolutely leave

answers now please go and be

productive