A simple way to choose is to start where the overlap between high volume, predictable steps and clear rules is greatest. This intersection is your automation sweet spot—processes that will deliver real value quickly without excessive complexity.
1. List candidate processes
For a typical practice this might include onboarding, standard working papers, deadline tracking, reminder runs, and management packs. Don't filter too early—you want a complete picture of where time is being spent before you start prioritizing.
Talk to your team. The people doing the work every day often have the clearest view of which tasks feel repetitive and which cause the most frustration.
2. Score each process
Give each candidate a simple 1–5 score for:
- Time spent per month — How many hours does your team spend on this across all clients?
- Frequency of errors or rework — Does this process often need corrections or cause downstream problems?
- How repeatable the steps are — Is it the same sequence every time, or does it vary significantly?
- How stable the rules are — Will this process stay consistent over the next 12–24 months?
Tip: Processes that score highly on the first two criteria (time and errors) but lower on repeatability might still be worth automating—they just need more careful design upfront.
3. Pick one low-risk, high-impact candidate
The first automation should be small enough to succeed quickly but important enough that the time savings are obvious. That early win builds trust with your team and creates momentum for bigger projects.
Resist the urge to tackle your most complex process first. Even if it's the biggest time sink, a failed or delayed first project can damage appetite for automation across the firm.
What happens next
Once you've identified your first candidate, the real work begins: documenting the current process, identifying edge cases, and designing the automated workflow. Each of these steps is an opportunity to simplify before you automate.
Remember: the goal isn't to automate everything. It's to free your team from the work that machines do better, so they can focus on the work that requires human judgment and expertise.