Systems We've Built
Working Pilot / Prototype

Frontline English

A bilingual frontline-learning prototype that teaches one operational boundary at a time: what the learner can handle, when to get help, and what reliable English phrase to use when help is needed.

The first micro-lesson was designed for a Polish Nail Lounge & Co. pilot. The goal is not to grade how someone sounds. It is to make a real workplace decision easier to recognize and act on.

Try the Micro-Lesson Discuss a Custom Training System
One construct at a timeThe first lesson isolates boundary recognition instead of mixing policy, vocabulary, and multiple-choice trivia.
Spanish as supportEnglish remains the taught language while Spanish appears underneath to reduce unnecessary friction.
Human-defined authorityThe business decides where the learner's authority stops and what complete phrase should be used to escalate.
No pronunciation scoringThe pilot does not score accent, nationality, “native” sound, or speech pronunciation.
The Operating Problem

Frontline language training can fail when it teaches words without teaching the boundary.

A learner can know basic English and still face a harder workplace question: “Am I allowed to answer this myself?” The first lesson deliberately separates that operating decision from everything else.

The lesson asks one question

Can the learner recognize a situation that is inside their authority?
Can they recognize a situation that should be escalated?
When escalation is needed, can they produce one reliable English phrase?
Can the product teach this without requiring English typing, pronunciation scoring, or a complicated dashboard?
What O2S Built

A small learning loop around a real operational decision

The pilot micro-lesson keeps the interaction intentionally narrow. The learner sees a service scenario, chooses whether they can handle it or need help, and practices the escalation phrase only when help is the correct decision.

Introduce the skill
Teach one phrase
Show a worked example
Present a client scenario
Handle or get help?
Build the phrase if needed
Record the session
01

Binary decision

Every practice scenario uses the same two choices: “I can handle this” or “I need help.” That isolates boundary recognition from unrelated distractors.

02

Separate error types

Handling something that should be escalated is different from over-asking for help on something the learner can handle. The session log keeps those mistakes separate.

03

Phrase reconstruction

When help is correct, the learner rebuilds a short phrase by tapping chunks in order. That practices recognition and production without making A1-A2 learners type English.

Interactive Public Demo

Try the learning loop yourself

This sanitized demo uses fictional service situations and a generic manager escalation phrase. It preserves the pilot's core interaction: recognize the boundary, then practice the phrase when help is needed.

Public prototype · bilingual support · no employee dataThe demo does not record audio, score pronunciation, grade accent, or send session data anywhere. Reloading starts a fresh session.
Learning Controls

The design is intentionally constrained

Complete phrase per location

The escalation language is set as a complete approved phrase. The system does not assume a person's name and a role noun can be swapped into the same grammar.

Help and handle stay balanced

The pilot set includes both escalation-boundary situations and ordinary in-scope requests so “always ask for help” cannot masquerade as mastery.

Irregular scenario order

Help and handle situations are mixed rather than alternating predictably, reducing the chance that a learner can answer from the sequence instead of the situation.

English taught, Spanish supports

The English phrase remains primary. Spanish appears underneath as support rather than replacing the taught language.

In-memory session log

The prototype can record first-choice accuracy, missed-help errors, over-asks, retries, phrase builds, time, and voluntary restarts for the current session only.

No learner-facing numeric grade

The learner receives simple corrective feedback. Evaluation is about the learning loop and operating boundary, not creating a high-stakes score.

Pilot protocol
3 learners.
Same lesson.
Same test.
The protocol is designed to test the prototype consistently, not to make broad claims from a tiny sample.
How the Pilot Is Evaluated

Test the product, not the person.

Run the lesson one learner at a time without teaching them how the interface works first.
Use fixed pre-test, post-test, and next-day prompts so all pilot learners face the same protocol.
Measure the boundary decision separately from exact operational knowledge such as hours or payment policy.
On get-help cases, ask what the learner would say to the client and note exact, approximate, or forgotten next-day recall.
Treat three learners as an early usability and learning signal—not proof of broad effectiveness.
Human in the Loop

The business defines the boundary. The prototype helps a learner practice it.

1

Manager / operator

Defines which situations frontline staff may handle, which require help, and the complete escalation phrase approved for that location.

2

Learner

Practices the boundary decision and the phrase in short bilingual service scenarios without being asked to diagnose, improvise policy, or prove pronunciation.

3

Facilitator

Observes usability and learning behavior, then uses the session log and standardized pre/post/next-day prompts to evaluate the prototype.

What It Intentionally Does Not Do

The pilot proves the learning loop before adding complexity.

The source curriculum explicitly holds back features that could distract from the first question: can the prototype teach the operational boundary?

No speech recording or pronunciation scoring
No accent, nationality, or “native-sounding” scoring
No AI roleplay in the current pilot
No learner accounts or subscriptions
No manager dashboard
No high-stakes numeric grade
No full menu or policy personalization yet
No persistent analytics beyond the current session
What This Demonstrates

The lesson is specific. The training-system capability is broader.

Frontline English shows how O2S can convert a workplace judgment call into a focused digital learning loop: define the boundary, reduce unnecessary language load, practice the correct action, capture useful evidence, and refine from real learner behavior.

Define the task Set the boundary Teach one behavior Practice Observe Refine
Training That Fits the Work

What frontline behavior does your team need to recognize and perform consistently?

O2S can help turn a real operating requirement into a focused training workflow, prototype, or custom internal learning system.

Discuss a Custom Training System Explore Systems We've Built