Building your own personal chatbot can seem like a daunting task, but with the right tools and knowledge, it’s not as complicated as you might think. In this blog post, we’ll be discussing the steps involved in building a chatbot on any topic of your choosing, just like customGPT did in 20 mins with the Tom Brady chatbot.
Why Build Your Own Personal Chatbot On A Topic?
Building your own chatbot on a specific topic has many benefits. For starters, it allows you to dive deep into a subject that interests you and gain a deeper understanding of it. This can be especially useful if you’re looking to become an expert in a particular field or learn more about a particular person or brand. Gone are the days of spending hours on Google rummaging through 100 websites.
Moreover, building a chatbot on a topic also allows you to engage with others who share your interests. By building a chatbot, you’ll be able to connect with like-minded individuals and share your knowledge and insights. This can lead to new friendships and opportunities for collaboration. Imagine building a chatbot on your favorite topic and then sharing it with friends to enjoy or research.
Additionally, building a chatbot can also be a great learning opportunity. You’ll have the chance to work with new technologies and tools and learn how to use them effectively. This can be especially beneficial for those looking to develop their technical skills or for students/researchers looking to deep dive into a particular field or topic.
Here are some of the main benefits of building your own sitemap and chatbot:
Conduct Research on a Topic
By creating a custom sitemap and chatbot, you can conduct research on a particular topic or industry. For example, if you are interested in learning more about a specific industry or market, you can use a tool like ahrefs to collect a list of relevant web pages and create a custom sitemap.
You can then use a generative AI chatbot like CustomGPT to analyze the content of these web pages and provide insights and predictions about the market. This can help you make more informed decisions about your business or investments.
For example, here is a CustomGPT chatbot built on the topic of “Abraham Lincoln”. It was built in 15 mins by using 1000 webpages on the topic of Lincoln.
When asked the question “How was Lincoln’s relationship with his son?”, the bot responded with:
The available evidence indicates that Abraham Lincoln had a close and respectful relationship with his son. Biographer Jason Emerson wrote that Robert was "intelligent, witty, kind, gentlemanly, proper, and generous," and that Abraham showed his gratitude to his stepmother by being "like an own son to her." Furthermore, the soldiers remarked on the "complete and sacred" understanding between the man and the boy when they marched hand in hand through the streets of Richmond.
To build a ChatGPT chatbot like this in 10 mins:
- Use this free tool and give it a list of keywords.
- Input the generated sitemap into CustomGPT.
- Few mins later, you are done! Start asking away..
Gather Competitor Intelligence
Creating a custom sitemap and chatbot can also help you gather intelligence about your competitors. By collecting a list of web pages from your competitors’ websites, you can use a generative AI chatbot to analyze their content and provide insights about their strategies, products, and marketing. This can help you identify areas where you can improve and stay ahead of the competition.
Ninja Tip: Using this technique, you can create content based on your competitor’s content. It’s almost like you have read all your competitors’ pages and are now ready to analyze and generate content based on that.
Learn from Your Private Data
Creating a custom sitemap and chatbot can also be used to learn from your own private data. For example, if you have a large knowledgebase or repository of documents, you can use a generative AI chatbot to analyze the content and provide insights and recommendations based on the data.
UPDATE (new) : You can now also easily upload your documents and build a chatbot in seconds. No more painfully searching through PDF documents to get the answers you want.
This can help you identify patterns and trends that may not be apparent by manual analysis.
And best of all : You can do this in minutes using a secure no-code cloud-hosted platform like CustomGPT.
Automate Customer Support and Engagement
Finally, creating a custom sitemap and chatbot can also be used to automate customer support and engagement. By using a generative AI chatbot, you can provide instant and personalized responses to customer inquiries and support tickets. This can help reduce response times and improve customer satisfaction.
Overall, creating a custom sitemap and chatbot can provide many benefits for businesses and individuals. By leveraging the power of generative AI, you can gain insights, automate tasks, and improve decision-making.
How Does This Work?
A chatbot built using OpenAI works by using a big computer program called a model to understand and respond to people’s questions. Think of it like a really smart robot that you can talk to.
When you ask the chatbot a question, it takes your words and tries to understand what you’re asking. Then, it looks through all the information it has been trained on to find an answer. This information will be the collection of articles or pages that you will give it + the general information from the Internet.
Finally, the chatbot uses all the information it has found to come up with an answer to your question. It’s like a detective trying to solve a mystery. It collects all the clues and pieces them together to find the answer.
The more information the chatbot has been trained on, the better it will be at answering questions.
Example
For the impatient, if you want to see demo of the final “Tom Brady” chatbot, see the Live Demo here. This was built using this exact step-by-step guide below in about 20 mins.
Step-By-Step Guide
So without further ado, let’s get rolling.
Step 1: Choose Your Topic
The first step in building your own chatbot is to choose a topic. This can be anything you’re passionate about or knowledgeable about. It can be a celebrity, a sports team, a historical figure, a specific brand or keyphrase. Once you’ve chosen your topic, you’ll need to gather information about it. This information will form the basis of your chatbot’s knowledge base.
UPDATE: If you are unsure about how to build your collection, see this library of free tools we have built to help you create your collection of knowledgebases.
For example, let’s say your topic is “Tom Brady”
Step 2: Collect Your URLs
Next, you’ll need to collect a list of URLs that contain information about your chosen topic. There are a few different ways to do this. One of the simplest methods is to use Google to search for relevant websites and then use ahrefs to export the list.
To build my Tom Brady list, I simply searched Google for a bunch of Tom Brady related keywords and then used ahrefs to download the list.
Step 3: Create the Sitemap
Once you’ve cleaned your URLs, it’s time to create your sitemap. You can use this free online tool to create your sitemap. Use the Copy option to copy the link to the sitemap.
Build Your CustomGPT Chatbot
And now, the final step: To build your personal custom ChatGPT-style chatbot based on the webpages you collected, do these simple steps:
Step 1: Setup Project in CustomGPT
Once you sign up for CustomGPT, select “Create Project” and enter your validated sitemap URL into your Project.
Step 2 : Enjoy!
That’s it – you are done. In a few minutes, click on the “Ask Me Anything” button and start questioning your ChatBOT just like you would with ChatGPT.
Here is the final “Tom Brady” chatbot. Give it a try!
Frequently Asked Questions
What Webpages Can I Use?
You can use any web pages that you want. You can literally research any topic, make a collection based on that and make a chatbot from it.
Fun example: Our CEO researched his own name and built a chatbot from it (See the AldenBot). Now he can create responses in his own tone, voice, and writing style. Imagine writing emails, social media posts, essays, etc, using your own voice.
Can We Use Your Spreadsheet?
Absolutely. Here is my Google Sheets . Just “Make A Copy” and replace the URL list.
Help – I’m Not Technical
This procedure is super simple. If you are non-technical, just use this free tool (or any of the other free tools from our library of free tools) and it will give you the sitemap. It took me 2 mins to run through this entire setup process.
Can I build a chatbot just using keywords?
We hear you – you have a list of keywords based on your niche and want to build a chatbot from it. Now, you can in minutes. Just use this free tool and give it your list of keywords. Then using the generated sitemap, build your chatbot in CustomGPT.
Ninja Tip: Feeling like having a little fun? Search for your own name using the free tool and build a ChatGPT chatbot based on keywords with all the public information about you.
Conclusion
Building your own chatbot is a fun and rewarding experience. And it might (just might!) save you a ton of time in your daily life. The possibilities are endless, whether it is researching a topic, writing emails, or writing responses in your voice, tone, and style.
By following the steps outlined in this blog post, you’ll be able to build a chatbot on any topic of your choosing. Whether you’re a sports fan, a history buff, or just interested in a particular keyphrase or topic, building a chatbot is a great way to expand your knowledge or dig deeper into a topic. So, why not give it a try today?
Got Questions? Ask them in the comments below. What will you build today?
12 Comments
My company is considering building a chatbot that will need to reference 100s of 1,000s of URLs (if not more). So we’d be looking at either the top plan or a custom plan. I have a few questions about building the sitemap:
1. Should we include public URLs that may have already been scanned/ingested by ChatGPT?
2. By specifying our desired URLs does that give the data on those pages more “weight” than other pages scanned by ChatGPT?
3. New information pages/URLs are added weekly and we’d like to add them to our customer chatbot. Also some pages may change and we’d need them updated. Do we create additional sitemap files with the new/updated URLs or just add them to a main sitemap file (the one used to start building chatbot)?
4. Can the custom chatbot handle spreadsheets of information and reply correctly using information on the sheets? If so, do we post those spreadsheets online and provide the URL to the sitemap file?
5. We have some password protected pages that we’d want to include in the custom chatbot, is there a way to ingest those or will we have to make “public” versions/URLs for the sitemap file?
Thank you. We look forward to moving ahead on this project.
David Berky
CIO, innoviHealth inc
> 1. Should we include public URLs that may have already been scanned/ingested by ChatGPT?
ChatGPT has ingested only upto June 2021. I would recommend creating the collection based on exactly what you want the bot to extract content from.
> 2. By specifying our desired URLs does that give the data on those pages more “weight” than other pages scanned by ChatGPT?
Yes. We use the “prompt” from the user to extract out CONTEXT from your content. We then pass that to OpenAI 3.5 to ask it to give the final answer. So the OpenAI completion puts more weight on the context we are passing it.
> 3. New information pages/URLs are added weekly and we’d like to add them to our customer chatbot. Also some pages may change and we’d need them updated. Do we create additional sitemap files with the new/updated URLs or just add them to a main sitemap file (the one used to start building chatbot)?
For now, we just ingest in all the pages and build the bot. Our plan is that in the near future, we will be able to capture CDC (change data capture). This way, as you add/delete/update pages, the bot will adapt (just like Google does with its index)
> 4. Can the custom chatbot handle spreadsheets of information and reply correctly using information on the sheets? If so, do we post those spreadsheets online and provide the URL to the sitemap file?
Not yet – but that is coming soon (we have it almost working in dev). Just to be clear: It will extract TEXT from the spreadsheet but not the full structure. So the days of being able to ask BI questions like “Give me the average of the Northeast sales from people aged over 55” are a few months away for most AI technologies.
> 5. We have some password protected pages that we’d want to include in the custom chatbot, is there a way to ingest those or will we have to make “public” versions/URLs for the sitemap file?
We are working on an “Upload” feature which is coming soon. This will let you upload all sorts of docs (PDFs/DOC,PPT,etc) to build the chatbot.
Will I be able to train my own customGPT with PDFs documents or is it only available to train it through URL sitemaps?
The “Upload” feature is now live. You can train with PDFs and 55+ other formats (including docs, zip files, ppts, xls, mp3 podcasts, etc)
Having followed the instructions I keep getting a 404 from Github for my sitemap and I cannot tell what I am doing differently/wrong.
The TXT files are confirmed to be configured with my domain provider.
This is the link I am trying to publish to: https://envdig.github.io/mixpanelenvoy.github.io/
What am I doing wrong / what do I need to look out for on the Github side to make this work??
John – unfortunately, I cannot provide support for the hosting side. This is something you will need to take up with Github (or use a different file hosting system) – you could use S3 or any other file hosting system. I see that your top-level site seems to be working fine though : https://envdig.github.io/ — so you should probably debug from there.
Good info
Many of these platforms support zero-code, which means you can create a chatbot without writing a single line of code.
Yes — the main goal is the ease of use and the level of technical expertise needed. If 100 non-technical people start on this goal of building their own personal bot, a large percentage should be able to do so using just their browser. Ease of use is a feature!
With zero-code, which is supported by many of these platforms, you may build a chatbot completely hands-free.
With the help of many of these platforms, you may build a chatbot without writing any code at all.
Creating your own personal chatbot on a topic of interest can be a rewarding and insightful experience. It provides an opportunity to delve deep into a subject, gain expertise, and connect with others who share similar interests.