Liya Engine

Customise the AI assistant's name, tone, formality, encouragement level, disclaimers, and language.

personaConfig controls how the AI presents itself and communicates. It affects all conversational intents — coaching sessions, mock interviews, general chat, and career guidance.

Schema

{
  "personaConfig": {
    "name": "Liya",
    "role": "AI Career Coach",
    "company": "Acme Corp",
    "tone": "professional",
    "formality": "medium",
    "encouragement_level": "moderate",
    "language": "en",
    "locale": "en-US",
    "disclaimers": {
      "general": "This is AI-generated guidance and should not replace professional advice.",
      "legal": null,
      "salary": "Salary figures are market estimates and may vary by location."
    }
  }
}

Fields

`name`

The AI assistant's name as presented to users. Default: "Liya".

White-label deployments typically set this to match their product's assistant name.

`role`

The AI's stated role. Default: "AI Career Coach". Shown in introductions and when users ask what the assistant does.

`company`

Optional company name injected into the AI's context. When set, the AI can reference it to feel more embedded in the platform.

`tone`

Value	Behaviour
`professional`	Concise, direct, formal-leaning (default)
`friendly`	Warm, approachable, slightly informal
`casual`	Conversational, relaxed
`formal`	Strictly formal, document-style language

`formality`

Value	Behaviour
`low`	Contractions, short sentences, first-name basis
`medium`	Balanced (default)
`high`	Full sentences, no contractions, polished

`encouragement_level`

Value	Behaviour
`minimal`	Neutral, assessment-focused
`moderate`	Balanced encouragement with honest feedback (default)
`high`	Actively supportive tone, emphasises growth

`language`

BCP 47 language code for response language. Default: "en". Pass "es", "fr", "de", etc. to receive responses in that language.

`locale`

Locale for formatting (dates, currency). Default: "en-US".

`disclaimers`

Optional per-category disclaimer text injected at the end of relevant responses. null disables the disclaimer for that category.

Example — white-label deployment

{
  "personaConfig": {
    "name": "Aria",
    "role": "Your career guide",
    "company": "TalentHub",
    "tone": "friendly",
    "formality": "low",
    "encouragement_level": "high",
    "language": "en"
  }
}

Persona Configuration