I usually ask 4 questions:
1. does it repeat often?
2. does it waste time or create friction?
3. does it follow a clear pattern?
4. would a simpler system actually get used?
If the answer is yes to most of those, it is probably worth automating.
If the process is still vague, chaotic, or constantly changing, it usually needs cleaning up before it needs automation.
That one mindset shift saves people a lot of wasted build time.