Unless you’ve been living under a rock, you’ll have noticed the hype around AI has been going at a frenzied pace. To be candid, I worry sometimes about the future of my career and my role in the world. If I’m not riding the frontiers of AI in accounting, will I be made redundant?

It’s times like these that I remind myself that while AI is indeed a transformational technology, I also shouldn’t go looking for an answer in search of a question. I do deploy AI in my practice where it provides value, but I’m not trying to automate (or “agent-ify”) every single thing I do.

I still believe that leaving some useful friction for myself in my work is a good thing. Yes, there is useless friction that I continue to seek ways to remove, but the good friction (reviewing a client’s general ledger for the past week with my own two eyes, for example) is what leads to insights that help me better serve my clients.

Ultimately, I ask myself this question: Will assigning this task to AI help me better serve my clients? If yes, then I adopt it. If no, then I leave it alone. I ask this question because I run a client-first practice, not an AI-first practice. Taking this approach, I believe, will keep me from getting excited and then disappointed by the usual hype cycle.