Let's put profits aside for now and focus on how to optimize workflows with AI. Recently, I've been delving into the industry to understand the pain points faced by various sectors and exploring ways to implement AI solutions. This includes accounting firms, law firms, staffing agencies, and more. I'm using these experiences to improve my ability to digest AI iterations and to completely upgrade my workflow. Many people are doing this through OpenClaw, but surprisingly, engineers around me aren't very interested. I'm also curious about the reasons behind this. I believe the best thing OpenClaw does is forcing you to build a complete information transmission flow. It compels you to write good soul md files, helping you manage memory effectively, unify information entry points, and fully liberate permissions. OpenClaw is like a hammer pushing us to find nails. My current approach is to categorize each required function into tools like pliers, screwdrivers, scissors, etc. Once the entire process becomes proficient, I will train my ultimate worker. Besides saving token consumption, my problem-solving will also become more perfect. In this era, it's rare to see handwritten text, so cherish it.