So today was a pretty good day. I had a meeting with an investor (not going to lie, it was paid) but I needed real world advice from a third party. I think I have talked about it in the past but getting feedback from friends and family can simply skew the true feeling as they don’t want to let you down. In the developer meeting we had a little discussion of implementing AI, it’s probably going to happen.
So I had a great meeting with Lance Cottrell from Feel the Boot. Nice guy who was able to give honest feedback about our business and where we are in the stage of growing. It was good to hear that he liked the idea and felt we were onto a good thing and knows it is a huge market. In fact, while doing my research today about the TAM of home industry, I found a report from Service Titan who pulled from Angi’s. In 2022 the home services market was about $650B which is a big number but what was more important to me was that its estimated that there are 650M appointments per year. That includes home improvement but appointments none the less. That blows my 250M appointment assumption out of the water. Will need to update my deck.
Anyways, Lance gave some solid advice and really stated we need to get more homeowners on the platform and some more appointments set before we would really be ready to get some investment. He said that when we do, to keep him in the loop. He seems like he may be interested in joining us. We talked about network effects and how it can help us create a moat. I told him we had our patent happening and he advised not to talk much about it until we had something at least provisionally pending. Well, emailed the lawyers after the meeting and they got it filed today. So my good news is we are patent pending.
So in the developer meeting we discussed possibly using AI in our system. I couldn’t seem to really justify it yesterday when I was first thinking about it but after some thought this morning and discussion it does seem fit. I want to use AI in 2 ways. First, letting users speak or type in their problem. Instead of selecting cards on the app to find their problem, we can clean up the interface and make it simpler on users by just letting them tell us what the problem is. “I took a shower this morning and had not hot water” would go to plumbing > water heater unsure > broken. My wife told me a while back she didnt understand what the acronym “HVAC” is. I never thought of it that way and if someone has a broken air conditioner, they may not know it is under HVAC. So with AI, someone can simply say “my air conditioning isnt blowing cold air” and we can understand what the ultimate goal is, which is to get someone to look at the a/c.
So one more thing I want to add, not sure if it will be AI or not. I want to add some basic troubleshooting. No hot water, lets look at the water heater. Is it on? Is there an error code? Is the circuit breaker tripped? If we can solve some simple problems that don’t need a professional, homeowners will love the app that gets them back up and running and doesn’t cost them any money. Now the AI part will need a trained model and need some finesse. We don’t need it hallucinating telling people to stick a fork into the electrical outlet to test it. But some basic troubleshooting to see if the appliance is dead vs outlet, if its a GFCI that is tripped or maybe a breaker. Nothing crazy (can probably do this with just some simple programming, but AI makes it COOL). We will see if its worth it. Token costs is the concern.
