Building a Client's $3,500 AI Chatbot LIVE

Name: Building a Client's $3,500 AI Chatbot LIVE
Uploaded: 2023-06-24T15:00:08.000Z
Duration: 1 h 12 min 54 s

# Building an AI Chatbot for Clients

In this video, Liam Otley builds an AI chatbot for a client's SAS platform that works with martial arts businesses. The chatbot is designed to collect information from potential customers and guide them along a personalized conversation to find the correct products and services.

Introduction

Liam Otley introduces himself as the owner of Morningside AI, an AI development and automation company.

Morningside AI helps businesses build custom-coded AI applications and work as automation partners to reduce friction, decrease expenses, and increase profits.

Liam Otley is in Amsterdam with his business partner before heading out later this week.

Client's SAS Platform

The client has a SAS platform that works with martial arts businesses to handle their leads, systems, and products.

The chatbot is meant to be placed on all of the different sites of the SAS customers so that he could have some kind of AI chat bot element that's going to take caption information from the users since you figure out what they're looking to find on the site help direct them to the kind of products and services and booking Pages.

The main functionality is to collect information from potential customers who visit the site and guide them along a personalized conversation so that they're able to find the correct products and services.

Chatbot Features

The chatbot was built using Botpress and StackAI.

It collects lead information such as email addresses, names, intent of user interaction with business etc., which can be used later for marketing purposes.

Zapier was used to send captured lead data directly into Google Sheets.

Conclusion

Liam Otley demonstrates how he built an AI chatbot for one of his clients' SAS platforms. He explains how it works together with Botpress and StackAI. He also shows how it guides potential customers through a personalized conversation to find the correct products and services. The chatbot also collects lead information, which can be used later for marketing purposes.

# Introduction to Building a Chatbot with Botpress and Stack AI

In this section, the speaker introduces the topic of building a chatbot using Botpress and Stack AI. They explain that they will be showing how to connect the two platforms using API calls, and that viewers will be able to download the bot template at the end of the video.

Building a Chatbot with Botpress and Stack AI

The speaker explains that they will be building a chatbot using Botpress and Stack AI.

The tech stack being used is explained - Botpress as the base application for handling logic, deployment, and UI of the bot; custom code blocks within Botpress are used to make API calls to Stack AI functionality.

The speaker notes that there are many different applications and platforms available for building chatbots with complex AI functionality.

An important export feature has been released by Botpress which allows users to share their bots with others.

# Creating a Chatbot in Botpress

In this section, the speaker walks through creating a chatbot in Botpress from scratch. They cover adding nodes, text blocks, information capture elements, routing, custom code blocks, API calls to Stack AI, and pushing leads to Google Sheets using Zapier.

Adding Nodes and Text Blocks

To create a new chatbot in Botpress from scratch:

Click "Create a Chatbot"

Choose "Completely from Scratch"

Add your first node by dragging out a standard node and renaming it.

Add text blocks to the node by copying and pasting.

Information Capture Elements

Add an information capture element in the form of a single choice button.

Ask users a question and create options for them to select.

Routing

Create routing between nodes based on user input.

Custom Code Blocks

Use custom code blocks within Botpress to make API calls to Stack AI functionality.

Pushing Leads to Google Sheets using Zapier

Use Zapier to push leads captured by the chatbot into Google Sheets.

# Building a Chatbot with Botpress

In this section, the speaker demonstrates how to build a chatbot using Botpress. They show how to test the chatbot and add elements such as messages and single-choice questions.

Testing the Chatbot

The speaker shows how to test the chatbot by starting a conversation with "hi" and demonstrating how each element displays.

They demonstrate capturing user information and selecting options for martial arts.

The speaker adds a new node for asking why users are looking to learn martial arts.

Custom Code Blocks

The speaker introduces custom code blocks in Botpress, which allow access to variables set within the program.

They demonstrate using an if/else block to create personalized messages based on user input.

Simplifying Nodes

By setting variables, the speaker is able to simplify nodes and reduce clutter in the program.

They demonstrate how personalized messages can be created using variables set earlier in the program.

# Large Conditional Block

In this section, the speaker explains how to use a switch case statement to generate personalized answers for different combinations of variables. They also introduce an execute code block that simplifies the process and saves time.

Using Switch Case Statement

A switch case statement can be used to generate personalized answers for different combinations of variables.

Provide Chat GPT with a breakdown of the structure of your chatbot.

The execute code block simplifies the process and saves time.

Providing Chat GPT with Structure

Provide Chat GPT with a breakdown of the structure of your chatbot.

This allows it to generate all possible combinations of variables.

Use a switch case statement or an FL statement to give personalized answers.

Execute Code Block

The execute code block is used to save time and simplify the process.

It sets the next message based on user input up to that point.

Combines inputs from previous messages into a personalized method.

Personalized Message

The personalized message makes users feel like everything they've said previously is leading them towards their desired outcome.

Sets the next message as this message and sends it directly afterwards.

# Information Capture

In this section, we learn how to capture more user information using single-choice blocks. We also see how text cards can be used in messages sent after users request more information.

Single Choice Block

Add a single choice block asking if users want more information on getting started.

Options are "Sure, tell me more" or "I have a question."

Text Card

After users select "Sure, tell me more," send a text card with an offer.

The offer can be customized using variables.

Information Capture

Add another information capture block asking if users have any questions regarding getting started.

Options are "Yes," "No," or "Email me the info."

Connect the yes route to a check knowledge node and the no route to the end.

# Knowledge Query Function

In this section, we learn about the AI aspect of chatbots. We see how Botpress knowledge query and Stack AI knowledge query can be used together to provide quick and accurate responses.

Botpress Knowledge Query

Use Botpress knowledge query as the initial knowledge query.

If it cannot answer a question, send an API call to Stack AI for a response.

Stack AI Knowledge Query

Stack AI uses GPT4 or 3.5 Turbo for its knowledge queries.

It has a better model and embedding system than Botpress.

Layering Botpress and Stack AI ensures quick and accurate responses.

Setting up a Knowledge Base in Botpress

In this section, the speaker explains how to set up a knowledge base within Botpress and enable the knowledge agent.

Creating a Knowledge Base

To create a knowledge base, click on "Add Knowledge Base" in the top left corner of Botpress.

Upload a text file with structured question-answer pairs for your FAQs.

Add a brief description and rename the knowledge base.

Enabling the Knowledge Agent

Go to the "Agent" section and enable the knowledge agent.

Turn off "Answer on Start Load".

Click on "Check Knowledge Node" and enable knowledge answering.

Configuring Routing

Set up routing depending on whether or not there is an answer from the knowledge base.

Add an expression with transition information labeled "No Answer from Knowledge".

Create another transition that is always true if there is no response from the knowledge agent.

Querying Botpress's Knowledge Base

In this section, we learn how to query Botpress's knowledge base using single choice questions and raw input.

Single Choice Questions & Raw Input

Set up a single choice question asking what they would like to know.

Change it from single choice to raw input to capture their text as value.

Catch this value as a new variable called "question".

Routing Based on Response

If there is no answer from the knowledge agent, route down to stack AI query node.

Create further cues by adding another node with single choice options for yes or no.

If yes, send them back up to ask more questions.

If no, take them to end of conversation.

Testing the Botpress Knowledge Base

In this section, we learn how to test the knowledge query by asking a question to the knowledge base.

Testing the Knowledge Query

Add "No Answer from WordPress Knowledge" statement to Stack AI queries.

Ask a question to the knowledge base by saying "Hi, I want to stay fit kickboxing. I have a question. What would you like to know?"

# Setting up Stack AI for Botpress

In this section, the speaker explains how to set up Stack AI for Botpress and how it can be used to answer more difficult questions.

Setting Up Stack AI

To set up Stack AI, you need a free account.

Once you have an account, go to the document and database templates and select the document database Q&A.

Upload your document into the system.

The main advantage of using a Stack AI query over the native Botpress one is personalization of answers.

Using Stack AI with Botpress

Copy and paste the code provided in the description into an execute code block in Botpress.

Add your API key to the code block.

Use workflow.question as input and capture output in API response variable.

# Routing and Lead Capture

In this section, the speaker explains how to set up routing for a chatbot and how to capture leads.

Setting Up Routing

To set up routing, click on the red line and take it to the further questions node.

Create an infinite loop where users can ask questions to the knowledge base.

If the chatbot is unable to answer a question, it will go to Stack AI.

At the end of the chatbot, if there are no more questions, users will be sent off with a custom knowledge base query.

Capturing Leads

Start lead capture with a single choice section asking if users have any questions regarding getting started.

One option is "email me the info."

Ask for their email address and save it as an email variable.

Ask for their full name and save it as a name variable.

Use custom code block to execute Zapier query.

Set zapier success value as false initially.

Take all lead data saved in variables (name, email, etc.) and add as payload for API call.

Make post request using URL provided by Zapier account setup.

When successful, set new variable as true.

Add transition after successful post request that sends user text saying "Perfect! We'll be in touch with you shortly" followed by an offer (e.g., 50% off).

Ask if they have any further questions before sending them down to further queue.

# Martial Arts Program

In this section, the speaker briefly discusses setting up routing for martial arts program inquiries.

Quickly pass through martial arts program inquiry section without going into detail.

# Building a Chatbot for Martial Arts Business

In this section, the speaker discusses how to build a chatbot that helps customers get information about the business and find products that are best for them. The chatbot is personalized to help customers feel like they are getting a customized experience.

Creating Options for Customers

Give customers options such as building confidence, building strength, learning discipline or all of the above.

Save customer's response to a variable and turn off custom routes.

Add an always transition to move on to the next node.

Adding Custom Code Block

Add a custom code block based on customer's response in previous node.

Personalize message depending on what they've answered in the previous node.

Add next message response and next card with single choice question.

Single Choice Question

Drag in a single choice question asking if they have a martial art in mind.

Save their response to martial art type variable created earlier.

Options include kickboxing, MMA, karate.

Finalizing Chatbot

Turn off connectors and add always transition at bottom set to true.

Finish building chatbot that helps customers find right class or martial art for their needs.

# Publishing Chatbot

In this section, the speaker discusses how to publish the chatbot so it can be used by clients.

High-Level Overview of Chatbot

Built chatbot using Bot Press and Stack AI with lead capture that sends data to Google sheet via Zapier integration.

Chatbot helps customers find right class or martial art for their needs.

Can be customized to provide different services to clients.

Publishing Chatbot

Click publish in top right corner of Bot Press.

Refresh page and click on Integrations to see options for handing over chatbot to client.

Copy pre-configured link and send it to clients so they can chat with the bot on a hosted web page.

Alternatively, take script tag and send it to clients so they can put it on their website.

# Interacting with Chatbot

In this section, the speaker demonstrates how to interact with the chatbot and provides an example conversation.

Example Conversation

Ask chatbot about kids program and why someone would want their child to start martial arts training.

Chatbot responds that discipline is a foundational aspect of all martial arts and asks what type of martial art they have in mind.

Options include karate, kickboxing, MMA.

Chatbot offers more information on classes with discipline coach John and answers questions about fitness level needed for classes.

# Customizing UI

In this section, the speaker discusses how to customize the UI of the chatbot.

Adding UI

Take script tag from Bot Press and add it to client's website.

Customize UI using Bot Press features such as AI side of Bot Press.

Experiment with different deployment options available from Bot Press.

# Conclusion

In this section, the speaker concludes by encouraging viewers to keep experimenting with building different types of chatbots using Bot Press and Stack AI.

Final Thoughts

Join the speaker's community Discord for workshops and Q&A sessions.

Keep experimenting with different features of Bot Press and Stack AI to build different types of chatbots.

# Join the Discord and Watch Best Niches for AI Automation Agency Videos

In this section, the speaker encourages viewers to join his Discord community and watch his "Best Niches for AI Automation Agency" videos. He explains that these resources are valuable for those looking to start their own AI automation agency.

Join the Discord Community

The speaker encourages viewers to join his Discord community.

He emphasizes that the community provides value and answers questions.

Viewers are encouraged to extract value from the free resources provided in the community.

Watch "Best Niches for AI Automation Agency" Videos

The speaker recommends watching his "Best Niches for AI Automation Agency" videos.

These videos provide useful information on how to approach starting an AI automation agency.

The speaker shares a framework and criteria that he uses and recommends for picking a niche.

By following this advice, viewers can set themselves up for success.

# Subscribe to Channel and Sign Up for Newsletter

In this section, the speaker encourages viewers to subscribe to his channel and sign up for his newsletter. He explains that he puts out videos two to three times a week on AI automation topics, as well as offering tips and tricks via email.

Subscribe to Channel

Viewers are encouraged to subscribe to the speaker's channel.

The channel puts out videos two to three times a week on AI automation topics.

The content is designed to be as valuable as possible.

Viewers can sign up for the speaker's AI automation newsletter via links in the description or pin comment.

The newsletter offers tips and tricks sent directly to subscribers' inboxes multiple times per week.