# Auto-responding

A key benefit of the persona cognitive models is the entities now have goals, beliefs and attitudes. This means **when you ask a persona to respond, it's not only *****communicating***** in character but it's *****behaving***** in character**.

Conducttr uses a multi-stage reasoning pipeline that simulates the internal cognitive and emotional life of a persona. This is why it's important to have those fields on the persona completed.

## Behavioural Simulation Engine

When the persona is asked to compose a message in the Comms Manager, either in the compose mode or auto-responding mode, it's uses a three-stage reasoning pipeline which we call our 4Rs model. The persona asks itself:

* is the request RELEVANT?
* does it RESONATE?
* how should I REACT?
* what's my RESPONSE?

These stages are explained below.

<details>

<summary>A detailed diagram is here</summary>

```mermaid
graph TD
    Start((External Stimulus)) --> ELM{1. ELM Cognitive Routing}

    subgraph Information Processing
    ELM -- High Relevance / Goal Alignment --> Central[Central Route - Structured Reasoning]
    ELM -- Stress / Identity Threat --> Peripheral[Peripheral Route - Heuristic Response]
    end

    Central --> OCC[2. OCC Affective Appraisal]
    Peripheral --> OCC

    subgraph Internal State Update
    OCC --> Goals[Goal Desirability Check]
    OCC --> Norms[Approbation and Standards]
    OCC --> Taste[Attraction and Disposition]
    Goals --> Emotion[Derive Resultant Emotion and Mind State]
    Norms --> Emotion
    Taste --> Emotion
    end

    Emotion --> Coping{3 Coping and Intent Selection}

    subgraph Intent Formation
    Coping --> Epistemic[Epistemic Coping]
    Coping --> Social[Social Coping]
    Coping --> Action[Action Coping]
    Epistemic --> Intent[Select Dominant Intent]
    Social --> Intent
    Action --> Intent
    Intent --> Primary[Select Primary Action]
    end

    Primary --> COMB{4. COM-B Execution Gate}

    subgraph Behavioral Output
    COMB --> Cap[Capability - Social and Intellectual Power]
    COMB --> Opp[Opportunity - Environmental Constraints]
    COMB --> Mot[Motivation - Internal Driver]
    Cap --> Final[Persona Authentic Action]
    Opp --> Final
    Mot --> Final
    end

    Final --> Loop((Update Persistent Memory))
    Loop -.-> Start
```

</details>

#### 1. Relevance - Adaptive Information Processing (The ELM Framework)

To manage cognitive load and persona authenticity, the engine utilises the [Elaboration Likelihood Model (ELM)](https://en.wikipedia.org/wiki/Elaboration_likelihood_model). This allows the system to determine the most realistic "route" for processing information:

* Central Route: Engaged when a stimulus aligns with the persona’s core goals or professional expertise. This results in high-effort, reasoned, and structured responses.
* Peripheral Route: Triggered by environmental stress, identity threats, or low-relevance stimuli. The persona relies on mental shortcuts, heuristics, or emotional reactivity.

#### 2. Resonance - Affective Appraisal & Emotional State (The OCC Model)

We utilize the [OCC (Ortony, Clore, and Collins) Model](https://www.igi-global.com/dictionary/occ-model-of-emotion/20746) to derive balanced reactions. Rather than assigning arbitrary "moods," the engine evaluates the stimulus against three cognitive variables:

1. Goal-Desirability: Does this event help or hinder the persona’s objectives?
2. Approbation: Does the action align with the persona’s internal standards or social norms?
3. Attraction: Is the stimulus inherently aligned with the persona's disposition?

This appraisal generates a specific Resultant Emotion, which is then scaled by situational factors like morale and vigilance to update the persona's "Mind State."

#### 3. Reaction - Coping and Action Selection

This stage takes the emotional OCC output and selects a coping orientation that reflects the persona’s role, motivations, and situational context.

Three orientations are evaluated:

* **Epistemic** – reducing uncertainty by seeking information or verification.
* **Social** – managing relationships or emotional alignment with others.
* **Action** – intervening directly to influence the situation.

The strongest orientation becomes the **dominant intent**, which is translated into a **primary action** such as asking a question, warning others, coordinating activity, or offering help.

This step converts internal appraisal into a concrete behavioural goal that is then evaluated by the COM-B stage to determine whether the behaviour can occur.

#### 4. Response- Behavioral Capability & Execution (The COM-B System)

To ensure the persona "acts" realistically, we apply the COM-B model ([Capability, Opportunity, Motivation, and Behavior](https://thedecisionlab.com/reference-guide/organizational-behavior/the-com-b-model-for-behavior-change)). This acts as a logical gatekeeper for the simulation:

<table data-header-hidden><thead><tr><th width="152.666748046875"></th><th></th></tr></thead><tbody><tr><td><strong>Component</strong></td><td>Function in the Engine</td></tr><tr><td><strong>Capability</strong></td><td>Does the persona have the knowledge or social standing to respond this way?</td></tr><tr><td><strong>Opportunity</strong></td><td>Does the current channel or environment allow for this specific reaction?</td></tr><tr><td><strong>Motivation</strong></td><td>Do the internal drivers (derived from OCC and ELM) outweigh the social cost?</td></tr></tbody></table>

If the engine detects a conflict—such as a persona wanting to speak out but lacking the institutional "Opportunity", it triggers a pivot, forcing the persona to react through a "social mask" or via a more cautious channel.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://helpdocs.conducttr.com/feature-documentation/designing-and-running/personas/working-with-personas/auto-responding.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.