Execution Results

When an RPA robot completes its task, it’s essential to correctly interpret the results. Heptora will show you both the overall process status (whether it completed or failed) and the result of each individual action (OK, KO, or ERROR).

The Three Execution States

Within an RPA process, especially in tasks that handle lists of data (such as an Excel table or CSV file), each line or record the robot attempts to process has its own result:

✅ OK (Success)

Indicates that the data line was processed correctly. The robot completed all programmed actions for that record without encountering any problems.

Example: If the robot needed to enter data from a row into a web form, an OK result means the operation was performed successfully.

⚠️ KO (Controlled Error)

Indicates that the data line could not be processed correctly because it violates a business rule or already meets the objective. The robot knows why it failed, but marks it as KO since these cases require human intervention to verify the data and, if necessary, correct the information.

KO Characteristics:

The overall process can continue with subsequent lines
It’s an expected error controlled by business rules
Requires manual data review
Doesn’t indicate a problem with the robot, but with input data

❌ ERROR (Uncontrolled Error)

Indicates that the data line could not be processed correctly for some unknown reason. For some reason, the robot couldn’t complete its task for that specific record.

ERROR Characteristics:

The overall process can continue with subsequent lines
It’s an unexpected error not covered by business rules
May be caused by technical or connectivity issues
May require relaunching the process or technical intervention

Practical Example: Invoicing

Imagine a process that starts with an Excel file containing invoices and needs to check them on a website. If they appear as “pending,” then the data must be entered into the website’s form for invoicing.

Complete Scenario

Invoice	Web Status	Complete Data	Connectivity	Result	Reason
INV-001	Pending	✅ Yes	✅ OK	OK	Processed and sent correctly
INV-002	Paid	✅ Yes	✅ OK	KO	Already paid, no action required
INV-003	Pending	❌ No (missing Tax ID)	✅ OK	KO	Incomplete data or incorrect format
INV-004	Pending	✅ Yes	❌ Failure	ERROR	Website didn’t load at that moment

Result Breakdown

✅ OK - Invoice INV-001

Checked on the website
Appeared as “pending”
Form filled correctly
Sent for invoicing without errors
Action: None, process completed

⚠️ KO - Invoice INV-002

Checked on the website
Appeared as “paid” (doesn’t meet business rule)
Action: Manually review if it’s really paid or there’s a system error

⚠️ KO - Invoice INV-003

Checked on the website
Appeared as “pending”
When trying to fill the form, Tax ID is missing or has incorrect format
Action: Correct the data in Excel and relaunch only this invoice

❌ ERROR - Invoice INV-004

Attempted to check on the website
Website didn’t load due to momentary connectivity issue
Action: Relaunch the process or manually process this invoice

Why Can a Record Fail?

Common Causes of KO (Controlled Error)

Missing or incorrect data
- Empty required fields
- Incorrect format (dates, Tax IDs, emails)
- Values outside allowed range
- Disallowed special characters
Business rules not met
- Record already processed previously
- Doesn’t meet established conditions
- State incompatible with required action
- Duplicates detected
Failed validations
- Data that doesn’t pass verifications
- Inconsistencies between related fields
- Outdated information

Common Causes of ERROR (Uncontrolled Error)

Connectivity problems
- Momentary network drop
- Connection timeout
- High latency
System availability
- Target system temporarily unavailable
- Unscheduled maintenance
- Server overload
Interface problems
- Unexpected changes in website or application
- Elements not found
- Excessive loading times
Technical errors
- Permission or authentication issues
- Server errors (500, 503)
- Exhausted system resources

Where to Check Results

Heptora offers two ways to check your process results:

📧 Option 1: By Email

By default, Heptora sends an email to the user who started the task (and any other email defined in the process).

The email contains:

Quick summary of overall task status
Breakdown of individual actions
Number of OK, KO, and ERROR results
Information about the executed process
Execution date and time

Limitations:

Only provides an overview
Doesn’t include detail of each specific line
Summarized information for quick reference

🌐 Option 2: Executions Panel on the Web

For more thorough analysis, access the Executions Panel within the Heptora web platform.

Features:

Complete history of all executed tasks
Detailed view of each execution
Specific results per line/record
Precise identification of which lines gave KO or ERROR
Description of failure reason
AI-proposed solution
Ability to filter and search
Results export

How to access:

Log in to Heptora
Navigate to the Executions Panel
Select the execution you want to review
View detailed results of each action

Key Difference: KO vs ERROR

Understanding the difference between these two types of failures is fundamental to taking the right actions:

⚠️ KO - Controlled Error

Aspect	Description
Nature	Problem with input data
Predictability	Expected, covered by rules
Cause	Incorrect, incomplete data, or doesn’t meet rules
Solution	Manual data correction or human validation
Prevention	Improve input data quality
Responsible	User or business team

Typical example: An invoice with incorrect Tax ID, a duplicate record, a state that doesn’t allow the action.

❌ ERROR - Uncontrolled Error

Aspect	Description
Nature	Problem with automation or environment
Predictability	Unexpected, not covered
Cause	Technical failures, connectivity, availability
Solution	Relaunch process or technical intervention
Prevention	Monitoring, automatic retries, alerts
Responsible	IT team or technical support

Typical example: Website that doesn’t load, connection timeout, server error, interface change.

Human Intervention and Failure Review

Why Review Failures?

The differentiation between OK, KO, and ERROR exists so you know exactly what needs your attention:

Process completeness: Ensure 100% of records are processed
Data quality: Identify recurring problems in input data
Continuous improvement: Detect failure patterns for future prevention
Audit: Maintain record of what was processed and what wasn’t
Compliance: Ensure no vital information is lost

Recommended Review Flow

1. Review the Summary

Check the notification email
Verify the number of OK, KO, and ERROR
Identify if there’s a pattern

2. Analyze KOs (Controlled Errors)

Steps:

Access the Executions Panel on the web
Identify each record marked as KO
Read the failure reason description
Classify by problem type

Actions:

Incorrect data: Correct the data at the source (Excel, CSV, etc.)
Business rule: Validate if the record should really be processed
Manual process: Complete the task manually for that record
Report: If it’s a special case, document for future reference

3. Analyze ERRORs (Uncontrolled Errors)

Steps:

Identify the time of errors
Review if there are patterns (same time, same website, etc.)
Verify the technical cause

Actions according to cause:

Isolated failure (momentary connectivity):

Relaunch the complete process or only failed records
ERROR records may process correctly on retry

Systematic failure (website down, interface change):

Contact IT team or support
Don’t relaunch until underlying problem is resolved
Document the incident

Intermittent failure:

Configure automatic retries if available
Monitor if it repeats
Consider adjusting timeouts or wait times

4. Complete the Process

For records that weren’t processed automatically:

Option A - Correction and Relaunch:

Correct problematic data
Create a new file with only failed records
Relaunch the process with this data

Option B - Manual Process:

For each KO/ERROR record that can’t be relaunched
Complete the task manually
Document in your system that it was manually processed

Option C - Discard:

If the record shouldn’t be processed (valid KO)
Mark as reviewed and discarded
Document the reason

Best Practices

Daily review
- Set a time of day to review executions
- Don’t let unreviewed records accumulate
Documentation
- Keep a log of recurring problems
- Create a knowledge base of solutions
Communication
- Inform data owners about recurring KOs
- Escalate systematic ERRORs to technical team
Continuous improvement
- Analyze failure patterns
- Propose improvements in source data quality
- Suggest additional validations in the process
Metrics
- Keep track of OK vs KO vs ERROR percentage
- Set improvement goals
- Celebrate achieved improvements

Reprocessing Failed Records

When to Relaunch the Process

✅ Relaunching is useful when:

ERRORs were due to isolated connectivity failures
Data causing KOs has been corrected
Target system is now available
Only a few hours have passed since failure

❌ Don’t relaunch if:

KOs are due to valid business rules
Data remains uncorrected
Technical problem persists
Days have passed (may cause duplicates)

Reprocessing Strategies

Option 1: Complete relaunch

Execute the process again with all data
Robot will process those that were already OK (verify if this causes duplicates)
Useful when there are many isolated failures

Option 2: Selective relaunch

Create a file with only KO/ERROR records
Execute the process with only this data
More efficient and avoids duplicates

Option 3: Hybrid process

Automatically relaunch ERRORs
Manually process complex KOs
Combine the best of both worlds

Frequently Asked Questions

Is a process with KOs and ERRORs considered “failed”?

Not necessarily. The overall process status can be “Completed” even with individual records with KO or ERROR. The process did its job of attempting to process all lines. What’s important is reviewing and managing the failures.

Will KOs and ERRORs be automatically reprocessed?

No. Failed records won’t be automatically processed in the next execution. You must manually intervene by correcting data (KOs) or relaunching the process (isolated ERRORs).

How long are detailed results kept?

Results are maintained in the Executions Panel according to your plan’s retention policy. Generally, you have access to complete history for several months.

Can I export records with KO/ERROR?

Yes, from the Executions Panel you can export detailed results, including only failed records, to facilitate their correction and reprocessing.

How can I reduce the number of KOs?

KOs are reduced by improving input data quality:

Implement prior validations in source files
Train users on correct format
Automate data generation when possible
Create templates with built-in validations

How can I reduce the number of ERRORs?

ERRORs are reduced with technical improvements:

Configure automatic retries for isolated failures
Improve process robustness with dynamic waits
Schedule executions during low-load times
Monitor availability of target systems

Quick Summary

Status	Meaning	Typical Cause	Required Action
✅ OK	Processed correctly	-	None
⚠️ KO	Controlled error	Incorrect data or rule not met	Correct data and review
❌ ERROR	Uncontrolled error	Technical failure or connectivity	Relaunch or contact IT

Need more help?

If this guide didn’t solve your problem or you found an error in the documentation:

Technical support: help@heptora.com
Clearly describe the problem you encountered
Include screenshots if possible
Indicate which documentation steps you followed

Our support team will help you resolve any issue.

Secrets Management - Protect credentials to avoid authentication errors
Robot Installation - Correctly configure your environment
Executions Panel (coming soon) - Complete guide to the web panel