AI Testing
AI Testing
Event
Test Automation with
AI Agents on
Real Transaction Data
Test Automation with
AI Agents on
Real Transaction Data
Test Automation with
AI Agents on
Real Transaction Data
Test Automation with
AI Agents on
Real Transaction Data
PerfecTwin validates using real transactions from your production system.
AI Agents handle the repetitive work.
PerfecTwin validates using real transactions from your production system. AI Agents handle the repetitive work.
PerfecTwin validates using real transactions from your production system. AI Agents handle the repetitive work.
PerfecTwin validates using real transactions from your production system. AI Agents handle the repetitive work.
One validation process —
from collecting real transactions to analyzing results
One validation process —
from collecting real transactions to analyzing results
One validation process —
from collecting real transactions to analyzing results
Turn real transactions into test cases — execution to analysis, fully automated.
Turn real transactions into test cases — execution to analysis, fully automated.
AI Agents join every stage
AI Agents join every stage
to expand coverage and pinpoint root causes fast.
to expand coverage and pinpoint root causes fast.
Data Agent
Data Agent
Data Collection
Data Collection
1.Data Collection
1.Data Collection
Collect transactions from
the production system
Collect transactions fromthe production system
Collect transactions fromthe production system
Collect transactions fromthe production system
Scenario Agent
Scenario Agent
Test Case Creation
Test Case Creation
2.Test Case Creation
2.Test Case Creation
AI auto-generates
scenarios & cases
AI auto-generatesscenarios & cases
AI auto-generatesscenarios & cases
AI auto-generatesscenarios & cases
Transaction
Mapping Agent
Transaction Mapping Agent
Test Execution
Test Execution
3.Test Execution
3.Test Execution
Run large-scale
repeated tests
Run large-scalerepeated tests
Run large-scalerepeated tests
Run large-scalerepeated tests
Report Agent
Report Agent
Result Analysis
Result Analysis
4.Result Analysis
4.Result Analysis
AI compares &
analyzes results
AI compares &analyzes results
AI compares &analyzes results
AI compares &analyzes results
ERP Edition
Standard Edition
AI Agents for SAP ERP Migration and Operations
* AI Agent features are planned for the first half of 2027
Scenario Agent
Scenario Generation
Type a prompt, get a runnable test scenario — PerfecTwin automates the SAP ERP
screens for you. Scenarios are easy to edit and reuse, meaning less manual work
and faster response to change.
Start generating a test
scenario?
Create a test scenario for a sales order of 10 units of material. Sales organization 1000, distribution channel 10, division 100. Enter CUST-1000 as the sold-to party.
Launching the SAP sales
order screen.
Auto-Execute
Screen
Data Entry
Sales order scenario creation completed
Transaction Mapping Agent
Agentic Self-Healing Scenarios
When a screen change breaks a scenario, PerfecTwin determines whether
it's a business error or a UI change. If it's a UI change, Agentic Self-Healing corrects
just that step and re-runs it — no rebuilding from scratch.
Review the corrections below
- Failure triage Not a business error.
- Target element 'Save' button not found.
- Probable cause A screen-change failure.
Correction btn_save → btn_save_hdr
Rationale Same save action and tooltip; the label changed to an icon and the position moved.
Report Agent
AI-Analyzed Reporting and Remediation Guidance
Get an official report — cover, summary, results, evidence — generated automatically.
Multilingual output covers global customers and audits, cutting documentation work.
Tests Passed
26K 86.7% pass rateTests Failed
4K 13.3% fail rateNo pricing condition
record
Tax code not set
View fixPosting failed: period
closed
Currency decimals
missing
Master data maintenance required.
[FI owner]
Tax code (country: JP, tax code: A2)
[SD owner]
Tax condition (transaction: VK11)
Report drafted and annotated. I wrote the summary, narrative, and captions for this evidence report.
Rewrite Section
Auto-classifies failure types — groups 4,000 failed results into 6 root-
cause categories.
Adjust tone: Audit, Executive, Technical.
Explain a Step
Auto-clusters similar failure patterns, reorganized by business
workflow.
Plain-language rationale for any action row.
Regenerate Captions
“Found 1,248 failures caused by a missing tax code”
AI captions for all 28 screen evidences.
Authoring Insights
• 14 step summaries drafted
• 28 screen descriptions generated
Authoring Insights
• 14 step summaries drafted
• 0 deviations detected
• 28 captions generated
Translation Support
KR / EN / JP multilingual output.
What Our Clients Say
What Our
Customers Say
“PerfecTwin's AI-powered result analysis and reporting significantly cut the time to consolidate a high volume of test results under a tight schedule. Consistent, fast results helped us review progress and make decisions with confidence.”
— Consulting Firm E
© 2026 LG CNS. All rights reserved.
Follow
© 2026 LG CNS. All rights reserved.
Follow
© 2026 LG CNS. All rights reserved.
Follow
© 2026 LG CNS. All rights reserved.
Follow