Nousable
M

Maria

B1 · Business English

Session just completed

Lessons

47

Skill dimensions

23 × 2 modes

Multimodal input

Signal extraction from a tutoring session.

Collection

Signal extraction from a tutoring session.

100total signals
32grammar
14vocabulary
32pronunciation
12communication
10interests & context

Transcript

Extracted Signals

These signals update the learner's skill graph.

Learner Graph

9 skill nodes across 4 categories. Each tracked in two modes: receptive and productive.

grammar
vocabulary
pronunciation
communication
Interest
|R = Receptive · P = Productive

From the updated graph, the engine selects the next activity.

Action

Autonomous content selection. The engine scores urgency for every skill, selects a candidate, and prescribes the next activity.

Urgency Scoring
Weighted Selection
Variety Constraints
Prescription

Urgency Scores

Conditionals

1.92

weakness

0.63

×

staleness

1.2

×

trend

1.08

×

preference

1.0

+

exploration

0.02

Selection Pool

Weighted random selection from top 3 candidates

52%
35%
13%
Conditionals
Verb Tenses
Coherence
Selected

Conditionals

52% probability

Prescription

Skill

Conditionals

Expressing real and hypothetical conditions, consequences, and imagined situations

Phenomenon

GRAM-B2-213

Difficulty

B1

Interaction Type

ai_avatar

Context

business_negotiation

TO TALK ABOUT AN IMPOSSIBLE SITUATION — 'IF' + PAST PERFECT SIMPLE and 'WOULD HAVE' + '-ED' in the main clause to imagine a different past

Available interaction types

sentence_completionsentence_buildinglistening_clipreading_passageai_avatar

When Maria completes this, new signals feed back in.

The Loop

Collection

100 signals

22-skill extraction pipeline

Learner Graph

23 × 2 dimensions

Event-sourced, time-decaying

Action

1 prescription

Autonomous content selection