Using AI in Practice: Small Repetitions over Grand Innovation

Talk about AI for long and it turns grand quickly. Automation, innovation, replacement… On site, what matters first is what to do today. So I use AI less as a "clever assistant" and more as a tool that reduces repetitive work. The name of a tool may be flashy, but its use has to stay plain to last.

Three things it attaches to well

1. Drafting — removing the blank screen, producing material for me to correct quickly
2. Summarizing and organizing — turning long notes, meeting content and material into structure
3. Polishing expression — fixing rough sentences to read better (careful not to change the facts)

These three share one thing: the quality of the input is the quality of the output. A note thrown in carelessly is organized carelessly. In return, a note that separates fact, utterance and interpretation is easy for the AI to structure and easy for me to verify.

Ways not to use it

• Writing unverified numbers as they are
• Letting the AI rewrite what a person on site said into a sentence it invented
• Avoiding responsibility with "the AI said so"
• Pasting sensitive personal information in as it is

A tool raises your speed, but judgment and responsibility stay with the person. The more you use AI, the more carefully you read and the more briefly you verify. In public posts especially, ask once more at the end: "is the basis of this sentence in my notes?"

Start with small experiments

One thing a day: structure a meeting note and look at it, draft an email, pull out tag candidates. Practice changes when the record of small experiments piles up, not with a grand adoption plan. Each time you experiment, leave one line on "what did this save / what did I nearly miss". Those lines become the team's AI usage guide.

AI does not replace the site. It helps leave it more clearly. That, at least, suits the way I work.