Automation is having a moment in healthcare—again. And as with most buzz cycles, it’s easy to fall into the trap of thinking automation is the universal answer to every problem. But here’s the inconvenient truth:
Not everything should be automated.
In fact, some workflows absolutely shouldn’t be—at least not yet. That’s not anti-technology; it’s just good sense. Because when automation is misapplied, it becomes expensive, brittle, hard to maintain, and—ironically—less efficient than the manual work it was meant to replace.
So let’s say the quiet part out loud: Knowing when not to automate is a critical part of any successful automation strategy.
Here’s how to spot the red flags before the bots roll in.
Let’s start with the obvious. If the process you’re targeting only happens occasionally—and “occasionally” means quarterly or less—it’s probably not worth automating. The development time, testing cycles, and long-tail maintenance overhead won’t pencil out.
Sanity check: Would a good old-fashioned SOP do the job? Then skip the bot.
Automating a broken process doesn’t make it better. It just makes the mess happen faster. If a process is poorly documented, inconsistently executed, or depends on tribal knowledge to work—stop. Automation can’t fix poor operational design. Fix the process first.
Helpful framing: Automate discipline, not dysfunction.
No bot in the world can read a sticky note, interpret an outdated spreadsheet with hidden macros, or guess which field is the “right” one when five different systems say five different things. If the process requires a human to decipher inconsistencies, it’s not ready for automation.
Litmus test: If you can’t write repeatable logic for it, don’t automate it.
What does “working” look like? If a team is chasing automation without a clear sense of ROI, improvement targets, or success criteria, the whole initiative is skating on thin ice. Measuring performance post-deployment shouldn’t be an afterthought—it should be the blueprint.
Rule of thumb: If there’s no baseline, there’s no business case.
It happens all the time: a problem gets scoped for automation, only to realize the solution is… turning on a feature in the EHR. Or editing a rule. Or redesigning a queue. If 90% of the value can be unlocked with a smarter build, that’s the move.
Smart strategy: Reach for automation when it’s the best option, not the first option.
Despite all the caveats above, automation is still one of the most powerful levers available—when used well. It excels in high-volume, repeatable workflows with consistent logic and available data.
Perfect use cases include:
Bottom line: Automation should reduce friction and amplify impact. But it should never be a substitute for thinking.
The rise (and fall) of high-profile automation startups has offered the industry a valuable reality check. It turns out that selling thousands of bots is easy—maintaining thousands of bespoke, non-standard bots is another matter entirely.
The biggest lesson? Standardization matters.
If every automation is a snowflake, the maintenance overhead becomes a blizzard. Eventually, even great technology can collapse under the weight of its own delivery model. In hindsight, the downfall wasn’t bad tech. It was the lack of repeatability. And it taught the entire industry that automation doesn’t scale unless it’s built to.
Healthcare automation isn’t about novelty—it’s about fit. It’s not about being first—it’s about being effective.
Every organization should be asking:
If the answer is yes across the board, automation may be the game-changer it’s promised to be. But if not—resist the urge. Do the dishes first. Then build the machine.
Because the smartest automation strategy starts with knowing when to wait.
