structuring knowledge for a small business

TL;DR: I visited a small manufacturing company I had never heard of, learned what they do, and built them a Notion-based knowledge system in a weekend using AI.

I thought it would be cool to see how things get made for some of the regional industry around West Virginia. So I was browsing through a Charleston area business directory and came across a company called E&H. I sent them an email and told them I would be interested in seeing their manufacturing facility.

Getting familiar with E&H

I met up with the owner of E&H later that week and we spent a couple hours talking about their business and what they do. I had no intentions of doing anything “tech” or “business” related for this company. I was really just hoping for a tour of their facility so I could see some sparks fly and learn how they build their oil and gas products. That was all I had in mind.

Enter Notion

But that night, back at home, I finally got around to downloading an application that I had been wanting to check out, called Notion. I had heard this software mentioned quite a few times from various sources but had never actually used it. I didn’t even really know what it was used for. So that night, after I had gotten home from the E&H meeting, I created a free account and started tinkering around with features.

To build out a workspace and actually learn how to use the application, I realized I would need some mock material. That’s when I thought of E&H. Their company website was loaded with product information and pictures and general content—so little by little I began pulling that content over into the Notion project I had just created. I didn’t really like the application at first. It seemed too bloated and redundant, but then I kind of started liking it. And then I really loved it.

When I sold my last business, I spent a ton of time creating and organizing documentation for the buyer, but it was all done in Google Drive, and without the assistance of AI. As I was building this Notion workspace, I kept thinking how great this would have been for that process…

Workflow

I started tinkering on Friday night with no understanding or familiarity with the application and by Monday evening I had built a fully functioning dashboard prototype for a company and product (timber bridges) I had just learned about. There is no way I could have done that without AI.

I had ChatGPT and the Notion tutorial webpage open on one monitor and the Notion project itself open on a second monitor. I copied and pasted (and uploaded) all the information I could gather from the E&H website into ChatGPT so it would have a good contextual understanding of the company and its products. This ChatGPT window then became the product expert who I consulted for specific content that I wanted for various workspace sections in Notion.

Why this is effective

Interesting to note that because the model is already trained on so much data, it doesn’t need a whole lot of context about the business before it can start filling in the blanks and creating an accurate picture of the company. In other words…you have to assume the model is already familiar with every market, and every niche within every market, and so a small amount of specific information about a business goes a long way in helping it create an accurate picture.

I designated a separate ChatGPT window for application ‘tech support’ which I consulted, in conjunction with the company tutorials, for anything related to the Notion software functionality. This was especially helpful for speeding up the application learning curve, and general flow, because I could ask unlimited questions without having to worry if my line of questioning is too fragmented and scattered for the ‘person’ I’m talking to. I could just ask any question in any direction and piece together the responses myself to create an understanding of how things work.

ChatGPT flattens learning curve

Another big timesaver was being able to simply ask my ‘tech support’ agent how a professional would hypothetically go about setting up the Notion project, and why they would do it that way. This gave me a really useful reference point for how the software is meant to be used from a high-level, which gave me instant confidence to start building my own version.

If I had time I could go into a lot more detail about some of the interesting nuanced things I learned from this little project. One of them being the Scenarios section of the project.

Find your customers anywhere

I thought this was really neat. I prompted ChatGPT to generate hypothetical scenarios describing customers who might need the specified product offered by E&H. It seems like you could go much deeper into this process and basically map-out the entire customer base for any product. Essentially reverse engineering your sales strategy from those specific end points. It’s literally telling you exactly where your customers are, who they are, and what they’re doing. So cool!

SCENARIOS

Looking at the provided documents for E&H, in several concise sentences, can you describe 10 hypothetical scenarios for customers who need a bridge like the ones they manufacture? Be specific about the person’s circumstances. What are the details of their particular situation? Where are they located (all should be in states located east of South dakota and texas).

COMPONENTS

I’m uploading more documents that contain information from this company. Create a list of inputs likely used by the business. I want to create a comprehensive list. Intellectual property, customer relationships, knowledge, expertise, distribution, etc. (don’t make anything up though just for the sake of it.)