I got this question last week from a prospect: “Can we just automate it?”
They were drowning in manual work. Their team was overwhelmed. AI and automation tools are everywhere now, and they figured technology could solve their problems.
I get it. We look at automation for our own business too.
But here’s what I told them: automating a broken process just gives you a faster way to produce bad outcomes.
As AI continues to expand, businesses are rushing to automate everything they can to save money and improve efficiency. The pressure is real. The tools are accessible. The promise is compelling.
But most companies are automating in the wrong order, and making their problems worse.
So today, I’m going to show you how to assess what to automate, when to automate it, and the framework for doing it right.
Let’s dive in.
Why automating broken processes makes things worse.
Here’s the pattern we see constantly:
A company has a manual process that’s slow, inconsistent, and frustrating. They decide to automate it. They implement new technology. And the problems multiply.
Why? Because automation doesn’t fix bad processes. It scales them.
If your process has bottlenecks, automation makes those bottlenecks faster but doesn’t remove them.
If your data is inconsistent, automation spreads that inconsistency at scale.
If your handoffs are unclear, automation just creates automated confusion.
The result: you’ve spent money and time implementing technology that made the underlying problems worse, not better.
Here’s what should happen first: develop strong processes and clean data before looking to automate.
Get things running well manually. Then use technology to speed up what’s already working.
The framework for assessing what to automate.
Not everything should be automated. And nothing should be automated before the process is sound.
Here are the questions to ask before implementing automation:
Question #1: Is the process clearly defined and documented?
If people are doing the work differently every time, automation won’t help. It will just automate inconsistency.
First, standardize the process. Document it. Make sure everyone follows the same steps.
Question #2: Is the data clean and reliable?
Automation depends on good data. If your data has errors, duplicates, or inconsistencies, automation will amplify those problems.
First, clean your data. Establish standards. Make sure the inputs are reliable.
Question #3: Are the current bottlenecks process-related or volume-related?
If the bottleneck is a broken process (unclear handoffs, missing steps, poor communication) automation won’t fix it.
First, remove the process bottleneck. Then, if volume is still the constraint, automation can help.
Question #4: Does the process require judgment or is it rules-based?
Automation works best for repeatable, rules-based tasks. If the work requires judgment, context, or adaptation, humans should stay involved.
Automate the repetitive. Keep humans where judgment matters.
Question #5: What’s the cost of getting this wrong?
Some processes are low-risk. Automating them makes sense even if they’re not perfect. Other processes (like client communications or financial decisions) have high stakes.
Start with low-risk automation. Build confidence. Then expand to higher-risk areas.
The roadmap for piloting new processes and automation.
If you’ve answered those questions and determined that a process is ready for automation, here’s how to pilot it effectively:
Step 1: Define clear goals.
What are you trying to achieve? Faster turnaround? Reduced errors? Lower costs? Be specific.
Vague goals like “improve efficiency” don’t give you a way to measure success.
Step 2: Assess current processes and KPIs.
Document how the process works today. Measure current performance—time, cost, error rate, whatever matters.
You need a baseline to know if automation actually improved things.
Step 3: Refine the process and remove bottlenecks.
Before adding technology, fix what’s broken. Eliminate unnecessary steps. Clarify handoffs. Standardize the approach.
This is the step most companies skip—and why their automation fails.
Step 4: Put technology in place to speed up the process.
Now that the process is sound, implement the automation. Start small. Pilot with one team or one workflow.
Don’t try to automate everything at once.
Step 5: Review what’s working.
After the pilot, measure results against your baseline. Did you hit your goals? What broke? What worked better than expected?
Use this data to refine before expanding.
Step 6: Automate and scale.
Once the pilot proves the automation works, expand it. But keep monitoring. Technology doesn’t stay effective forever—processes evolve, and automation needs to evolve with them.
Here’s the lesson: automation is powerful. But only when applied to processes that are already working.
Fix the process first. Then use technology to scale what’s effective.
Most companies do it backward. They automate first and hope it fixes their problems. It doesn’t. It just creates expensive, automated chaos.
The companies that get automation right follow a disciplined approach: strong processes and clean data first, then technology to accelerate what’s already working.
That’s how you make automation pay off.
Whenever you’re ready, here are three ways we can help…
1. Strategy & Growth Blueprint: Market-grounded insights + an annual plan + a 90-day execution board your team owns.
2. Operations & Tech Reset: We map bottlenecks, design future-state processes, and build a phased tech roadmap ready to launch.
3. Manager+ Accelerator: We build core skills in delegation, feedback, goal-setting, and shape leaders who drive execution.
