Are Books Dead? I Turned My 50,000 Word Draft Book Into a Functional Chatbot

I wrote a book that sits in next-to-last-draft form on my hard drive, or the cloud, or wherever. I'm two weeks from finishing it at any given point, but I haven't gotten there yet.

Last week, I decided to turn it into a chatbot using a Custom GPT. Custom GPTs are one of Open.ai's new ai tools that let users build their own chatbots based on their own data.

GPT means generative pre-trained transformer. It's an AI that understands and generates human-like text by learning from large amounts of data.

• "Generative": It creates new content, like text, based on patterns it learned from the data it was trained on.
• "Pre-trained": The AI was previously trained on a large dataset before it's used for specific tasks or questions.
• "Transformer": A type of AI model that's really good at understanding the context and relationships in language.

Why did I do that? I was guilty about this almost-done book sitting around, for one. For two, I wanted to see what it would be like. Third, I knew the chatbot would fix any small errors in its output: correct spelling and grammar and adjust awkward sentences I've yet to fix, so it wouldn't embarrass me.

I also did it because I recognized that my knowledge was doing nothing sitting on a hard drive.

Even Draft Books are Data

It's 50,000 words in twelve chapters based on my four years of experience as a grievance analyst and grievance team manager. What does that mean? It means that when you ask my Custom GPT a question, it comes back with an answer based on my data. The advantage is that it doesn't have to search the web or the billions of words it was trained on to give you a general or bland answer.

Instead, it gives you a job-specific answer that is immediately practical based on the words, or "knowledge" as it's known, that the author loaded into it.

The other advantage? Using a Custom GPT-based chatbot loaded with specific reduces the chance of the hallucinations AI is known for. AI, given bad instructions by an unskilled user, tends to act like a loyal dog. It wants to appease its master and to do that will sometimes pretend it knows what it's talking about.

Giving users direct access to the specific knowledge they want, significantly reduces this effect, while allowing the AI to act like the loyal pup it wants to be.

And it only took ten minutes to build - on top of the several hundred hours the book took to write.

What Can You Do With A Custom GPT Based on a Book?

What can't you do, is the better question. You can ask the GPT book questions and drill down further. You can give it scenarios and use it to help walk you through them. You can ask for summaries on complex issues. You can ask it to develop personalized training to build your knowledge base.

A Custom GPT chatbot puts the author and their expert knowledge in your pocket. And not just the author but the author-as-coach. A coach who will never make you feel stupid for asking a question, or speak down to you. A coach whose only function is to help you improve.

In my book's case, it is ready to discuss the grievance process, the players, functions, the organizations, the pitfalls, and the key concepts. Concepts like bias, procedural fairness, decision-making, and the administrative law it's based on.

What's My Book About?

Complaint management, my experience, and most importantly, what I learned as a grievance analyst and complaints team manager.

Without question, my four years of work at the Canadian Forces Grievance Authority were some of the most rewarding years I spent with the Canadian Forces. It not only shaped my understanding of complaint management and soldier grievances it gave me a solid foundation in administrative decision-making.

After retiring from full-time service in February 2023, I started writing my experiences down to create a knowledge base. I retired to found a complaint management company, Barker HR Solutions, with Caitlin Barker , and won our first contract as a dedicated grievance analyst for the Canadian Force's Military Personnel Group's Training Group.

Why This Book?

My onboarding and initial training at the grievance authority were useless. In those first months as a grievance analyst, I committed never to let anyone else experience what I experienced. There was no training other than a couple of days of death by PowerPoint online course, without context, followed by some bazaar practical exercises that focused more on what the grievance products should look like, rather than how you need to think to address a complaint.

My onboarding was shameful, and considering the organization's challenges, it was completely irresponsible. My idea in writing the book was to collect what I learned and pass it on to those who followed me.

One of the most shocking examples I managed to address was that of a retired member. She waited over six years for $10,000 she was clearly owed. Six years. All she needed was someone to look at it, write a letter and get a signature and the $10K would be hers.

How I Learned

My first year was likely among the poorest examples of onboarding and leadership I've ever witnessed. The pedagogy, if it could be called that, was "figure it out for yourself, and I'll let you know how you did."

Essentially, without direction, I was expected to produce a completed file, and only once I thought it was done would I be told if it was right, or more likely, what I got wrong.

I called this approach - guess the answer.

While I don't thrive on negative feedback, it was all that was on offer, so I doubled down. If I had to learn the job through criticism and negative feedback, the more files I did, the more criticism I would get, and the faster I would learn. So, based on that, I started attacking files and moved far more files in my first year than was expected of a new analyst.

Admittedly, not everyone works like that. By turning my book into a Custom GPT, someone can still learn in that environment, but they won't be alone. For the record, I eventually built the onboarding training package for new folks I needed.

More on Custom GPTs

The Evolving Resource: Beyond a Static Document

The beauty of a custom GPT is its dynamic nature. I could continuously add new modules and chapters, ensuring the resource stays relevant and ever-growing. This isn't just a book; it's a living document that evolves and adapts, providing an ongoing, interactive learning journey.

It Turns English to French Instantly

What's more, this GPT isn't limited to English. It can respond in French with a simple request, making the resource accessible to a broader audience. This feature breaks down language barriers, enhancing the learning experience for diverse individuals.

And it does that without French foundational documents. I add knowledge in English, it accesses that knowledge and responds in French. And the same could be done with French knowledge.

Empowerment Through Knowledge

This journey from a draft book to an interactive, custom GPT tool represents more than just a technological advancement; it's a new paradigm in learning, onboarding, and ongoing job-specific coaching.

It embodies flexibility, personalization, and the breaking of traditional educational boundaries. As I continue to expand this living document, my goal remains clear: to empower learners by providing them with an informative, engaging, and responsive resource.

Blog Post Generator

I also turned my draft book into an endless blog post generator using a second Custom GPT. When I'm at a loss for a blog post, I can ask it to give me ten ideas. If I have an idea, I can ask it to give me ten variations on that idea. After I pick one, It produces a draft and recommends a picture to go with it.

What I can also do with this is produce a response for someone who asks a question. In as long as it takes me to write their question down, I can give them a 500-word custom answer less than a minute later. That is particularly useful for answering repeated questions.

Takeaway

I don't think books are dead. There's something about holding a book in your hand that won't go away anytime soon. But my book? It's a technical "how-to". You don't need to read the whole thing to get the answer you're looking for. You can flip to the page. And if a book can be used like that, it's a perfect chatbot.

Ask it a natural language question and get a natural language answer in less time than it would take to open the book and find the page.

Again, books aren't dead, but you'll be left behind if you aren't learning AI's value and using it daily. If you have a draft book or a collection of "how-to" notes, put it to work to the benefit of others. The written word is data; with AI tools coming out daily, you have no excuse not to take advantage of it.

Interested in learning more about the process or want to check out my Custom GPT? Send me a DM.