One of the most common questions we get when working with new clients is: "How do I know what can be automated?"
It's a great question, and the honest answer is: almost anything that follows consistent rules can be automated. The better question is: which things *should* be?
Here's the framework we use.
The Four Filters
Before automating anything, run it through these four filters.
Filter 1: Is it repetitive?
Does this task happen more than once? Does it follow largely the same steps every time? If yes, it's a candidate. One-off tasks don't benefit from automation, but recurring ones do, and the return compounds over time.
Filter 2: Is it rules-based?
Can you describe the decision logic clearly? "If X happens, do Y. If condition A is true, take action B." If you can write it down as a flowchart, it can be automated. If the task requires nuanced judgment, contextual reading of a situation, or creative problem-solving, it should stay with a human.
Filter 3: Does it involve data moving between systems?
A huge portion of manual work is simply moving information from one place to another. A form submission to a CRM. A completed task to an invoice. An email confirmation to a calendar event. If data transfer is at the core of the task, automation can handle it perfectly.
Filter 4: Is the cost of an error acceptable?
Some tasks are fine to automate even if they occasionally produce a minor error. Others, like anything involving financial decisions, sensitive client communications, or legal documents, need human review before anything goes out. Automate where errors are low-risk. Keep humans in the loop where they're not.
A Practical Exercise
Sit down and list the ten things you or your team do most frequently each week. For each one, ask: Is it repetitive? Is it rules-based? Does it move data? Is an error acceptable? If the answer to the first three is yes and the last is yes or low-risk, it's a strong automation candidate.
Most businesses find that 40–60% of their weekly recurring tasks fall into this category. That's a significant amount of time and mental energy that could be reclaimed.
What Happens After You Identify the Candidates
Knowing *what* to automate is step one. Knowing *how* to automate it, which tools to connect, what the trigger logic should be, how to handle exceptions, is where the real design work happens. That's what we do.
If you've done this exercise and have a list of candidates you're not sure how to tackle, we'd be glad to walk through it with you.