Visa Deutsch

The Gatekeeper Protocol

Assumption

Round1

ConfidenceLow

Updated2026-01-24

Current Assumption

"Exam-focused feedback reduces telc A1 failure rates for first-time applicants."

Target User

Visa applicants in Turkey (TR) and Brazil (BR) risking high exam fees with low confidence.

Phase

In Development

Primary Signal

Decision Framework

○ Intent

Define the specific behavior change: Users treat the simulator as a "binding" prerequisite, refusing to book the real telc exam until they achieve a "Visa Deutsch PASS."

○ Assumption

Exam-focused feedback reduces telc A1 failure rates for first-time applicants.

○ Experiment Design

Implement hard-coded telc A1 scoring algorithm and test with 20 beta users against manual grading results.

○ Decision Gates

If Pass-Rate Correlation > 90% -> Scale
If L1 Feedback is ignored -> Pivot
If Refund Claims > 5% -> Stop

○ Current Decision

Proceed with Development

○ Confidence LevelLow

Current Position

Phase: Development / Implementation of Grading Logic.
Testing: Evaluating if LLM-based "Social Appropriateness" checks (e.g., BR001/TR004) accurately mirror telc examiner point deductions.
Decision Gates: If Free-to-Paid conversion for Schreiben/Sprechen > X% → Scale.

Decision History

2026-01-24

Refine (Implementation)

Hard-coding the **telc A1 Scoring Algorithm** (1.66× multiplier) and the "Compensatory Rule" (allowing a pass with 0 points in one module if others are strong).

2026-01-18

Signal Creation

Finalizing the **L1 Overlay** requirements for the Turkish and Brazilian corridors to handle "negative transfer" errors like V2 violations (BR002/TR001).