Every organization has dozens of workflows that could be automated. The question isn't whether to automate—it's which workflow to automate first. Pick wrong, and you waste months on low-impact work. Pick right, and you generate momentum, prove value, and fund the next automation.
The mistake most organizations make: automating what's easy instead of what matters. They choose workflows based on technical feasibility, not business impact. The result: automated processes that save minimal time, generate little excitement, and don't justify further investment.
Here's a systematic framework for identifying your highest-ROI automation candidate.
The Five Factors That Predict ROI
Not all automation opportunities are created equal. Five factors determine whether a workflow will deliver exceptional ROI or waste resources:
1. Volume: How Often It Happens
A workflow that runs once per month has limited impact potential. A workflow that runs 50 times per day compounds value quickly. Volume is the multiplier for every other benefit.
Consider two workflows: Processing vendor invoices (200 per month) versus generating daily sales reports (20 per day). Even if invoice processing takes longer per instance, the report generation has higher volume and likely higher total impact.
How to measure: Count actual occurrences over the last 90 days. Don't rely on estimates—people consistently overestimate frequency. Check logs, databases, or email counts for real data.
2. Time per Occurrence: The Manual Burden
How long does one execution take from start to finish? Include all the hidden time—waiting for approvals, switching between systems, double-checking data, handling errors.
A financial services firm thought reconciliation took "about 15 minutes." Time tracking revealed: 8 minutes of active work, 20 minutes waiting for data to refresh, 12 minutes checking for discrepancies, and 15 minutes following up on exceptions. Real time: 55 minutes.
How to measure: Shadow someone doing the work for a full week. Time each step. Include wait time, context switching, and rework. The real number is usually 2-3x initial estimates.
3. Exception Rate: How Often It Breaks
Automation works best on consistent, predictable workflows. High exception rates mean constant human intervention, reducing ROI and creating frustration.
A company automated expense report approvals. Exception rate: 40%. Every exception required manual review. The automation created more work than it saved because humans had to monitor and fix edge cases constantly.
How to measure: Track how often the "happy path" actually happens. What percentage of cases require special handling, manual intervention, or escalation? Under 10% is excellent. Over 30% is risky.
4. Risk: What Happens When It's Wrong
Some workflows have low error tolerance. Getting them wrong creates compliance issues, financial losses, or customer churn. Start with lower-risk workflows to build confidence and capability.
High risk: Loan approvals, medical dosing calculations, financial reporting. Low risk: Meeting scheduling, report formatting, data entry for internal records.
How to measure: Ask "What's the worst that happens if this fails?" If the answer involves regulatory fines, customer harm, or material financial impact, it's high risk. Save those for later.
5. Change Resistance: Will People Actually Use It?
Technical success doesn't guarantee adoption. If the automated workflow disrupts people's habits, threatens jobs, or requires behavioral change, resistance will limit ROI.
An automated scheduling system worked perfectly but usage remained below 30%. Investigation revealed: it required managers to change how they thought about scheduling, and senior staff saw it as questioning their judgment. The automation was technically sound but organizationally failed.
How to measure: Assess who's affected, what changes for them, and whether they have veto power. Workflows with enthusiastic champions and minimal disruption have the highest adoption rates.
The Automation ROI Scoring Model
Score each potential workflow across these five dimensions. Use a simple 1-5 scale:
Volume Score:
Time per Occurrence Score:
Exception Rate Score (inverted):
Risk Score (inverted):
Adoption Score:
Calculate total score by multiplying: Volume × Time × Exception Rate × Risk × Adoption
Why multiply instead of add? Because these factors compound. High volume with low time per occurrence still has limited impact. High time per occurrence with high exception rate creates more problems than it solves.
Real-World Example
A mid-sized company evaluated five automation candidates:
A. Monthly financial close process:
B. Customer support ticket routing:
C. Expense report processing:
D. Sales contract generation:
E. Daily sales report generation:
The winner: Daily sales report generation. Not because it's the most complex or prestigious, but because it scores highest across all dimensions. It happens frequently, follows predictable patterns, has minimal risk, and people want it.
Second place: Expense report processing. High volume, enthusiastic adoption, and meaningful time savings.
The monthly financial close scored lowest despite taking the most time. Low frequency, high risk, and high exception rate make it a poor first candidate—though it might be worth automating later after building capability.
Common Mistakes to Avoid
Automating the loudest complaint: The squeakiest wheel isn't always the highest impact. Measure objectively.
Choosing based on technical interest: The most technically interesting workflow isn't always the most valuable. Follow the ROI.
Starting with the most complex: Build capability on simpler workflows first. Complexity increases failure risk.
Ignoring adoption: Perfect automation that nobody uses delivers zero value. Factor in people from the start.
Focusing only on time savings: Time savings × volume × adoption = actual value. All three matter.
After You Choose
Once you've identified your highest-scoring workflow:
Validate the scores: Shadow the process in detail. Confirm volume, time, and exception rate with real data.
Map the happy path first: Document the 80% case. Worry about exceptions later.
Identify data sources: What systems does the workflow touch? Who owns them?
Define success metrics: How will you measure ROI? Time saved? Error reduction? User satisfaction?
Start small: Build a proof-of-concept for one part of the workflow. Prove value before scaling.
The Bottom Line
The best workflow to automate first isn't the most complex, the most painful, or the most interesting. It's the one that delivers the highest ROI with the lowest risk. Use data to identify it. Score objectively. Start there.
Success with your first automation creates momentum, proves value, justifies investment, and builds capability for harder problems. Choose well.
