AI Usage Required

Happy Tuesday. It’s been a fun few days to look at headlines and your financial portfolios. 

In the sporting world, it is Masters week which means Thursday and Friday will be completely unproductive (sorry boss), as I will be enjoying the peace from perfectly Jim Nantz’s traditional “hello friends” remarks and the hope that Rory McIlroy can finally bring home the green jacket. 

A Quick Note on the Markets

What a crazy start to the week. In the matter of minutes Monday morning, the S&P went from -4.5% to +4.5%. A $3T move in less than 30 minutes due to a fake tweet from a news account. We ended the day mixed, but we will still have volatility in the near future. 

When times like this occur, as in the market taking a pretty significant dip and assets falling 20%+ off their highs, it has traditionally been a great time to buy. Asset prices are discounted and investors are uncertain. But with time, uncertainty fades, and as we know the stock market is a long term game. Hopefully in a few years these few days will not even show up on a 10yr chart. Or I’m super wrong and this last sentence will look horrible in a few months.

Chances are I’m not and this is a great time to be buying into the market. On to regularly scheduled programming. 

Disclaimer: This is not financial advice. Please do your own research and/or consult a financial professional before making any investment decisions. 

Enabled by AI

Alright, don’t worry we won’t be spending that much time on AI today. Just kidding, we will but it’s a novel concept so stick around. Even if it sucks, the bright side is that I may take your eyes away from your portfolio for 6 minutes or so. To start, I want to share a tweet from the CEO of Shopify today.

Shopify is now making it a baseline expectation that AI is used by all employees and usage will be a part of performance reviews. I imagine a lot of users will simply use it to write emails, take notes, or help write documents. This is like buying a jetpack to hop over puddles. You are only scratching the service on the abilities of the technology.

As a showcase, here is a list of everything I have built with AI with little coding experience:

1. Get Daily Bank Files - Took 2 days to build.

   • For my job building a cash management solution, we need daily bank transaction data in order to demo our tool effectively to our clients. 

   • I built a tool that generates unique bank data on a daily basis through a python script and then stored on a front-end website. 

2. Bank Fee Analyzer - Took 6 hours to build.

   • Again for my job, we needed to see if AI could help analyze PDF account analysis statements. 

   • Using a demo account analysis statement, I build a PDF analyzer of bank fees along with a front-end tool that is easy to use and interact with. 

3. Sales Tracker - Took 6 hours to build.

   • I wanted to track the sales calls that I have demoed our product on and keep track of where the clients are in the pipeline, what I’ve heard from clients, if they have closed, etc. 

   • This is then going to plug into AI and identify trends in my calls and what clients say. 

4. Golf Scorecard Roaster - Took 30 minutes to build.

   • Upload a picture of your golf scorecard, and you get roasted. Built it during one sitting at a bar. Fun project. 

   • Roastmygolfscore.com

5. Stealth Product - Projected timeline is 4 weeks.

   • Currently building out a pretty advanced web application with the help of Cursor. For this one, I do have an engineer friend helping because there is some pretty advanced stuff.

Building these may be a little impressive, but I assure you I still make basic mistakes like storing API keys in client-side code (meaning anyone can find my API keys). Point is, some of these are pretty complex products that I was able to build by myself with no prior coding experience. 

So that brings me to the main point here, products are easier than ever to build. The best ones will be the ones that you build for yourself or the ones that are built for an extremely niche audience.

The Opportunity

This leads to an incredible opportunity in applied AI studios. Applied AI is what it sounds like, taking AI and building the application layer for consumers, companies, or individual clients on a contract basis. These studios have been starting to pop up for broader applications, like helping large companies incorporate AI, but I am more interested in the niche application. Since operating funds are low, you can build products in virtually any market and make meaningful revenue.

In order to explore that, this is how I would start an applied AI studio in the golf market. I’m choosing golf because I used to work in the industry and it’s a consumer I know well.

Applied AI Golf Studio

With the golf market, I would focus on consumers & the B2B applications that can be used by courses. 

First, I would build out two or three consumer applications. They would be a very early V1 of the product, but I would use a group of around 20 to measure the viable success of them.

One would be a series of local golf newsletters. A lot of the information I get for this newsletter would be scrapped from user reviews, social media posts, and other forms of data that is publicly available. Second would be a tee time agent. The worst part of booking golf in a busy city is getting a good tee time Saturday morning. You either have to be up at 6am a week before, or catch a cancellation. Consumers would easily pay $20/month for a tee time agent that automatically books times for them. 

Both of these would take around a week to spin up, but can be built with AI and have a low cost to run. Next, I would focus on the B2B side and more specifically on course management. This is where things change a little bit.

Niche markets, such as golf already have products in place to cover most of their needs. The issue is, these products are a square peg in a round hole. Since there are not many products on the market, everyone has to use the same ones, yet all courses have different needs. 

Here I would focus on an outbound sales motion where I talk to courses, resorts, and clubs about where their technology is lacking. From here, I can work with select clients to build their own custom system that works best for them. These systems could be built in anywhere from 4-12 weeks. 

Alright 300 words is enough for a hypothetical.

Focusing Niche

From that hypothetical, I hope two things were very clear. First, there is not a limit to what you can build or do with AI, you just have to know how to use it. Second, you can get extremely niche because of the low cost of operations. By spending less on engineering and other areas, you can earn more net revenue. You also earn more when you focus on a hyper-specific market instead of spreading yourself over a broad market.

Thank you, I actually did put a lot of effort into these graphs today.

Even though there is a smaller revenue opportunity in the niche game, there are less players. It turns into simple math that I could walk through right now, but I have written too much. 

You will start seeing applied AI studios popping up in bunches. They are the new consulting firms, except they will actually provide meaningful value.

Song of the Day

Took me long enough to do this, but finally have to share a song from my favorite band. This 16 minute rendition of #41 by Dave Matthews Band at Wrigley Field in 2010 is incredible. The energy from stadium shows like this are palpable and the solos from Jeff Coffin (the sax guy) and Tim Reynolds (electric guitar) are some of the best I’ve heard. Expect about 3 minutes and 45 seconds of song, and the rest just ol

https://open.spotify.com/track/4JokKmIbiyDvkXRJlXPd6z?si=9309db5e987841b3