🛠️ Build it, buy it, or wait? Where to invest your AI effort
November 22, 2025
By Ted SteinmannThe hardest thing to know right now in the AI utilization space is whether to build it, buy it, or sometimes, maybe just wait...
The application of language models as a somewhat nacent technology is accellerating and a nearly unfathomable pace. Each time I feel like I've learned something new and novel that I could latch on to, build with, and make a fortune, it seems that the world around me has already had the same idea and was able to move with it a lightening speed within an already mature ecosystem. Part of the reason for that is because the companies building AI are building with AI and therefore have developed a core competency in using the tools they build for others to build their own. Dogfooding at it's best.
Some of these are pleasantly surprising. One example in particular was with my service-financials project which I'm using it to track finances for my Rotary Club. As I'm learning is true with many service organizations, budgeting is not necessarily top of mind unless you have an accountancy or other financial expertise in your club. Having built a spreadsheet (using AI of course) with robust formulas and automatic calculations minimizing data entry effort, I was pleasantly surprised this morning to see Gemini suggest I ask it to summarize the document I'd recently been working in.
Here is a quick video demonstrating the type of response you can get from service-financials using spreadsheet ai in Google Drive:
Pretty impressive. And honestly, MUCH better summary than the one I just gave at the Rotary meeting last week.
In a space that moves this fast, leverage beats novelty. Start with what already exists, reserve your scarce build energy for the pieces that differentiate you, and maybe be willing to wait when the ground is still shifting. That spreadsheet summary reminded me: outcomes matter more than scaffolding. Use AI to shorten the distance to value, then invest deeper where it truly compounds.
Categories: blog
Tags: technology