Automation Check: The Difference Between Theory and Practice
"We can automate that!" – This phrase comes up in meetings more often than ever. And it's often true. The problem: Not every automation makes sense. Some cost more than they save.
Why Half of All Automation Projects Fail
The statistics are sobering: Around 50% of automation initiatives don't achieve planned results. The most common reasons:
1. Wrong Process Selected
- Too complex, too many exceptions
- Executed too rarely
- Too many manual decisions required
2. Wrong Expectations
- Overestimated time savings
- Underestimated implementation costs
- Ignored maintenance efforts
3. Lack of Preparation
- Process not standardized
- No clear rules defined
- Poor data quality
The 5 Criteria for Automation Potential
Criterion 1: Repetition Frequency
Rule of thumb: The more often a process runs, the higher the potential.
| Frequency | Potential | Example |
|---|---|---|
| Multiple times daily | Very high | Invoice receipt verification |
| Daily | High | Daily reports creation |
| Weekly | Medium | Inventory updates |
| Monthly | Low | Month-end closing |
| Annually | Very low | Year-end closing |
But note: Even rare processes can be worthwhile if they're very time-consuming.
Criterion 2: Rule-Based Nature
Automation needs clear rules. Ask yourself:
- Are there clear if-then relationships?
- Can decision criteria be documented?
- How often are there exceptions?
Automatable:
- "If invoice amount > €1000, then approval by department head"
- "If inventory < minimum quantity, then trigger order"
Difficult to automate:
- "If the customer is important..." (What does "important" mean?)
- "Decide by feeling..." (Which feeling?)
Criterion 3: Digital Availability
The process must be digitally representable:
Easy to automate:
- Data exists in systems
- Interfaces (APIs) available
- Standard formats (CSV, XML, JSON)
Hard to automate:
- Handwritten documents
- Phone calls as input
- Decisions based on gut feeling
Pro Tip: Use our Digital Maturity Assessment to determine how digital your processes already are.
Criterion 4: Error Susceptibility
Manual processes with high error rates are ideal candidates:
Typical error sources:
- Typos in data entry
- Forgotten steps
- Inconsistent execution
- Copy-paste errors
Rule of thumb: With an error rate > 2%, automation almost always pays off.
Criterion 5: Time Effort and Costs
Calculate the true costs:
Annual process costs =
(Processing time × Hourly rate × Frequency) +
Error costs +
Opportunity costs
Calculate it specifically: Our Process Cost Analyzer helps you determine true costs including hidden costs.
The Automation Scoring Model
Rate each process on a scale of 1-5:
| Criterion | 1 (low) | 5 (high) |
|---|---|---|
| Frequency | Annually | Hourly |
| Rule-based | Many exceptions | Clear rules |
| Digitization level | Paper-based | Fully digital |
| Error susceptibility | < 1% | > 10% |
| Time effort | < 10 min/month | > 20 hrs/month |
Interpretation:
- Score 20-25: Automate immediately
- Score 15-19: Good candidate, examine in detail
- Score 10-14: Possible, but calculate ROI precisely
- Score < 10: Probably not worthwhile
Which Automation Technology Fits?
Different approaches suit different processes:
RPA (Robotic Process Automation)
Suitable for:
- Existing systems without APIs
- Data transfer between legacy systems
- Rule-based screen work
Examples: UiPath, Automation Anywhere, Power Automate
Workflow Automation
Suitable for:
- Multi-stage approval processes
- Task distribution and tracking
- Document workflows
Examples: n8n, Make (Integromat), Zapier
AI-Powered Automation
Suitable for:
- Document processing (OCR + NLP)
- Email classification
- Data extraction from unstructured sources
Examples: Claude/GPT APIs, Google Document AI, AWS Textract
Low-Code/No-Code Platforms
Suitable for:
- Quick prototypes
- Simple integrations
- Citizen development
Examples: Retool, Bubble, Airtable Automations
The Most Common Automation Mistakes
Mistake 1: Starting with the Most Complex Process
Don't start with the biggest, but with the simplest process. Quick wins create momentum.
Mistake 2: Aiming for 100% Automation
80% automated with human review is often better than 100% with error risk.
Mistake 3: Ignoring Maintenance
Every automation needs maintenance. Plan 15-20% of initial effort per year for upkeep.
Mistake 4: Not Involving Employees
The people who execute the process know it best. Involve them.
Your Automation Roadmap
Phase 1: Inventory (1-2 weeks)
- Create list of all repetitive processes
- Collect rough time estimates
- Identify pain points
Phase 2: Scoring (1 week)
- Evaluate each process against the 5 criteria
- Identify top 5 candidates
- Run Automation Check
Phase 3: Business Case (1-2 weeks)
- Calculate ROI for top candidates
- Estimate implementation effort
- Make Build-vs-Buy Decision
Phase 4: Pilot (4-8 weeks)
- Automate one process
- Measure, learn, optimize
- Document for scaling
Free Automation Check
Want to know if your process is suitable for automation?
Our tool checks:
- Automatability based on the 5 criteria
- Suitable technology recommendation
- Estimated savings potential
- Next steps
→ Check Automation Potential Now
Conclusion: Systematics Beats Gut Feeling
The question isn't "Can we automate this?" – usually yes. The question is: "Should we automate this?" And only systematic analysis answers that question.
With the right criteria and a clear scoring model, you avoid expensive mistakes and focus on automations that truly create value.
Identified processes with high automation potential and ready to get started? Our AI Adoption Audit helps you with strategic implementation – from tool selection to implementation.
