"We should automate more" is easy to say and hard to act on. The instinct is usually to either automate nothing (too overwhelming to start) or to try to automate everything (too big to finish). Both fail for the same reason: there's no way to tell which work actually deserves the effort.
After building automation into a lot of businesses, I've found the useful question isn't "what could we automate?" — almost anything can be. It's "what should we automate first?" Get that right and the first project pays for itself, builds trust, and funds the next one. Here's the framework I use.
Start by finding the work, not the tools
Forget the software for a minute. The best automation candidates are tasks your team already does by hand, the same way, over and over. They share a recognizable shape:
A task is a strong automation candidate if it's…
- Repetitive — it happens many times a week, the same way each time
- Rule-based — the steps are consistent and don't need much judgment
- Time-consuming — it adds up to real hours across the week
- Trigger-able — something clearly kicks it off (a form, an email, a booking)
- Error-prone by hand — copy-paste mistakes cost you when it's done manually
If you want a quick way to surface these: for one week, every time someone on your team does something that makes them think "a computer should do this," write it down. By Friday you'll have your list.
The usual suspects
Across most service and growing businesses, the same handful of tasks show up again and again:
- Lead intake and follow-up — capturing new inquiries, qualifying them, and sending timely follow-ups so none go cold from a slow reply.
- Answering the same questions — hours, pricing, availability, "do you do X" — handled by a chatbot that answers from your own content.
- Scheduling and confirmations — booking, reminders, and reschedules that currently eat phone and email time.
- Document and email triage — sorting and pulling key details out of invoices, contracts, claims, or inbound email.
- Re-entering data across tools — the same client info typed into three systems because they don't talk to each other.
Prioritize by ROI, not by excitement
Once you have a list, don't start with the most interesting one — start with the one that pays off most for the least effort. A simple way to score each candidate:
Value = (hours saved per week × how painful the task is) ÷ effort to build. The task at the top of that list — high hours, high pain, low build effort — is almost always where to start. It funds and de-risks everything after it.
The reason to lead with the highest-ROI, lowest-effort task isn't just math. It's momentum. The first automation that visibly gives someone back three hours a week builds the trust you need to tackle the bigger, scarier projects later.
What not to automate (yet)
Some work looks automatable but isn't worth it — at least not first:
- Tasks that need real human judgment — high-stakes decisions, sensitive client conversations, anything where being wrong is expensive.
- Things you only do occasionally — automating a once-a-quarter task rarely earns back the build cost.
- Broken processes — if a workflow is a mess by hand, automate it and you just get a faster mess. Fix the process first, then automate it.
The short version
- Don't ask what you could automate — ask what to automate first.
- Look for work that's repetitive, rule-based, time-consuming, and trigger-able.
- Score candidates by hours saved and pain, divided by effort to build.
- Start with the highest-ROI, lowest-effort win to build momentum and trust.
- Don't automate a broken process — fix it first.
Not sure which is your first win?
That's exactly what the free 30-minute call is for. You walk me through how your business runs, and I map the highest-ROI tasks worth automating, what it would take to build them, and what the realistic return looks like. You leave with a clear picture whether or not you hire me. More on the automation & AI service →