> ## Documentation Index
> Fetch the complete documentation index at: https://docs.neuroncx.in/llms.txt
> Use this file to discover all available pages before exploring further.

# Setup

> Configure skills, forms, schema, agents, and workflow to get a new assessment running end to end.

## Overview

Assessments in NeuronCx let a call get automatically evaluated for quality — the AI listens to (or reads) an interaction and scores it against a defined set of skills and parameters.

The evaluation chain works like this:

**Skill Dimension → Schema → Parameters → Rubric + Prompt → evaluated by AI**

* A **Skill Dimension** is a category of ability being measured (e.g. Product Knowledge, Brand Standards).
* A **Schema** groups related **Parameters** under a form.
* Each **Parameter** carries a **Rubric** (the scoring rule — a score, a flag, or a sentiment) and a **Prompt** (the instruction the AI uses to judge that parameter against the call).
* The AI runs through every parameter and produces a scored, flagged, and summarized result for the call.

> [!NOTE]
>
> This setup touches three areas of the NeuronCx dashboard:
>
> * **Apps** — where forms, schema, parameters, rubrics, and the workflow are configured
> * **Cx Ops** — where agents are added, and where results and audits are viewed

***

## Prerequisites

Before starting a new assessment setup, gather the following from the client:

* **Agent list** — the customer executives who will be assessed, along with their branch/location.
* **Assessment parameters and agent scoring** — the specific skills, questions, and scoring criteria the client wants each call judged against.

> [!NOTE]
>
> These details are unique to every client engagement and shape everything built in the steps below.

***

## 1. Skill Mapping Setup

Skill Mapping is the foundation — it defines *what* abilities are being measured before any form or workflow is built. The entire flow — creating skills and mapping them to parameters — happens on one screen: **Apps → Forms → \[specific form] → Skills tab**.

**1. Open the form's Skills tab**

Go to **Apps → Forms** (third icon in the toolbar) → open the relevant form (e.g. `Conditional Assessment Form`) → select the **Skills** tab, alongside Schema, Parameters, Rubrics, and Adhoc Runs.

**2. Filter by schema**

Use the radio buttons at the top — **All**, or a specific schema (`Flagging`, `call_handling_soft_skills`, `Appointment_service_assessment`, `Telephone etiquette`, `Brand_SOPs_Common`, `Upselling_Cross`) — to narrow the list to one schema's skills.

**3. Review the Skill → Parameter table**

Each row pairs a **Skill** (with a unique bracketed ID, e.g. `[6491] Courtesy_Empathy`) to its **Parameter**, prefixed by the schema it belongs to (e.g. `call_handling_soft_skills- courtesy_empathy`).

A single skill can map to more than one parameter — for example, `[6489] Upselling_Offers` maps to both `Upselling_Cross- Salon_Upselling_Offers` and `Upselling_Cross- Clinic_Upselling_Offers`.

**4. Add a new skill-to-parameter mapping**

Click **+ MAPPING** (top-right) to create a new skill and map it to a parameter under the appropriate schema.

![Setup Skills](https://mintcdn.com/neuroncx/SVPDD1QciZJOXSpN/images/setup_skills.png?fit=max\&auto=format\&n=SVPDD1QciZJOXSpN\&q=85\&s=810f8e42369ce36d0eed755d01612786)

> [!NOTE]
>
> Each schema's parameters are named with the schema as a prefix (e.g. `Appointment_service_assessment- objection_handling`), which makes it easy to trace a skill back to the schema — and therefore the score bucket — it contributes to.

### 1.2 Skill Definition

Once the skill attribute is onboarded, define its structure and the actual skills within it.

**1. Assign Property Types**

Open the attribute's edit screen. Under **Property Types**, add subskill slots (e.g. `Subskill_1` through `Subskill_5`) using **+ Add**. These are the placeholders that will later hold real values.

**2. Define Skill Dimensions**

These property-type slots get filled in as actual **Skill Dimensions** — the categories being measured. Example dimensions: `Product Knowledge`, `Brand Standards`, `Communications Skills`, `Conversion Skills`, `Booking Process`, `Specialized Skills`.

**3. Attach payloads (subskills) to each dimension**

Open a specific Skill Dimension (e.g. Product Knowledge) and, in its payload table, attach the real values for each subskill slot.

Example — under **Product Knowledge**:

* Subskill\_1 → Rates And Packages
* Subskill\_2 → Membership Benefits
* Subskill\_3 → Offers and Deals
* Subskill\_4 → Hotel Details

***

## 2. Forms, Schema & Assessment Definitions

With skill dimensions defined, the next step is building the **form** the assessment will run on, and attaching the **schema, parameters, and rubrics** that drive scoring — all done under **Apps**.

### 2.1 Forms

**1. Go to Apps → Forms icon**

In the **Apps** section, use the icon toolbar at the top: the Forms icon is the **third icon** in the row (next to Workflow, and before Settings and Reports).

**2. Select a Form Type tab**

Forms are organized under tabs — **All Forms**, **MultiPurpose**, **WAF**, **Order**, **Insights**, **Assessments**, **OCR**, **Transcription**. For assessment setup, use the **Assessments** tab.

**3. Review the forms list**

Each row shows **Status**, **Form** name and ID, the **Workflow** it's tied to, and its attached **Schemas**.

**4. Create a new form**

Use **+ New Form** to create a form of the current type.

![Form Creation](https://mintcdn.com/neuroncx/5-ZojVOnTxkbUIJR/images/form_creation.png?fit=max\&auto=format\&n=5-ZojVOnTxkbUIJR\&q=85\&s=c2a3c0327c8941766c66c773529e758d)

> [!NOTE]
>
> A form's **Modality** (text or audio) and **Bot Channel** (WhatsApp or Phone) are shown alongside its name in this list.

### 2.2 Schema

Click into a specific form (e.g. `Conditional Assessment Form`) to open its detail view. This view has its own tab bar: **Schema**, **Parameters**, **Skills**, **Rubrics**, **Adhoc Runs**.

**1. Open the Schema tab**

This is the default tab when you open a form. Toggle **Active Schemas** to see currently active schemas.

**2. Read the schema table**

Each row shows an **Id - Status**, the **Schema** name, its **Weightage** (e.g. 20, editable via the pencil icon), and its linked **Attributes**. Some schemas are marked **Conditionally Compiled** — meaning they only apply under certain conditions (e.g. `Brand_SOPs_Common` only compiles for `Brand_SOPs`, `Upselling_Cross` only compiles for `Upsell_cross_sell_assess`).

**3. Add rules, conditions, or a new schema**

Use **+ RULE**, **+ CONDITIONS**, or **+ SCHEMA** as needed.

![Schema Tab](https://mintcdn.com/neuroncx/5B6mz3pCRh61cLt7/images/schema_tab.png?fit=max\&auto=format\&n=5B6mz3pCRh61cLt7\&q=85\&s=a8fd5bfce9a2e4bbc9afb027f3894de2)

![Schema tab inside a form showing active schemas with weightage](https://mintlify.s3.us-west-1.amazonaws.com/neuroncx/images/assessment/09-schema-tab.png)

### 2.3 Parameters

Still inside the same form detail view, the **Parameters** tab (next to Schema) lists the individual things being scored.

**1. Open the Parameters tab**

Filter by schema (All, or a specific schema) to see its parameters.

**2. Review each parameter's Rubric and Prompt**

Every parameter row shows:

* **Rubric** — the scoring rule, e.g. "Maximum score of 20" or a flag like "Flagged as rudeness\_coercion"
* **Prompt** — the instruction text the AI uses to judge that parameter on the call

**3. Cross-check against the Rubrics tab**

The **Rubrics** tab (also in this form's tab bar, alongside Skills and Adhoc Runs) shows the rubric definitions available to attach to parameters.

![Parameters](https://mintcdn.com/neuroncx/5B6mz3pCRh61cLt7/images/parameters.png?fit=max\&auto=format\&n=5B6mz3pCRh61cLt7\&q=85\&s=37887bb4ab1c164b686dca69afa51b1f)

> [!TIP]
>
> Rubric types generally fall into three kinds:
>
> * **Score** — true/false, or a graded numeric score
> * **Flag** — strictly yes/no
> * **Sentiment** — a sentiment tag
>
> Pay attention to how a parameter should behave when it doesn't apply to a given call (`not_applicable`), so it doesn't unfairly penalize the score.

### 2.4 Assessment Definitions

Under **Apps → gear/Settings icon**, three additional definitions support scoring:

* **Assessment Typology And Rubrics** — the types of rubrics available
* **Assessment Flag Definitions** — definitions for flag-type parameters
* **Assessment Sentiments** — definitions for sentiment-type parameters

***

## 3. Attribute Mapping

Skill Dimensions on their own don't drive scoring until they're mapped to a form's schema.

**1. Go to Apps → gear/Settings icon → Attribute Mapping**

Only forms with attached rubrics appear in the left-hand form list.

**2. Select the form**

Choose the relevant form (e.g. `Ovum_Agent_Skill_Assessment`) to see its existing mappings.

**3. Map each Skill Dimension to a Schema Key**

Each row pairs a **Skill Dimension** (e.g. "Greeting", "Maternity Package Details") to a **Schema Key** (e.g. `Soft_Skills_Audit`, `Filtered_Maternity_Calls`). Use **+ MAPPING** to add a new pairing, or the pencil icon to edit an existing one.

![Image](https://mintcdn.com/neuroncx/TgBRzoS1hYf55o0K/images/image-2.png?fit=max\&auto=format\&n=TgBRzoS1hYf55o0K\&q=85\&s=1781fe3bf603968ff2cb2898cd5398ad)

***

## 4. Adding Agents

Agents are the customer executives who handle the calls being assessed.

**1. Go to Cx Ops → Supervisor Console → Agent List**

This shows every agent with their current status (Online, On Break, Offline), External ID, and Queue.

**2. Add a new agent**

Click **+ Add Agent** and fill in the agent's details.

**3. Allocate the agent to a branch**

Assign the agent to their branch/queue so their calls and scores are tracked under the right location.

![Agent List](https://mintcdn.com/neuroncx/TgBRzoS1hYf55o0K/images/agent_list.png?fit=max\&auto=format\&n=TgBRzoS1hYf55o0K\&q=85\&s=29850b5a3e69cdd5865496939f41d1c3)

***

## 5. Creating the Workflow

The workflow is the engine that actually runs the assessment — transcribing the call, applying the schema, and scoring it against the rubric.

### 5.1 Workflow Blocks

**1. Go to Apps → gear icon (Resources and Settings)**

This panel lists every configurable resource: Prompts, Response Schema, LLM Functions, Files, Workflow Blocks, and more.

**2. Open Workflow Blocks**

Select **Workflow Blocks** ("Create and Edit Workflow") to see **My Blocks** and **All Blocks**, listing every block with its Id, Block Name, Description, and Type.

![Screenshot 2026 07 20 115508](https://mintcdn.com/neuroncx/TgBRzoS1hYf55o0K/images/Screenshot-2026-07-20-115508.png?fit=max\&auto=format\&n=TgBRzoS1hYf55o0K\&q=85\&s=40165a9a0ace51cb320a2ab2c33abd09)

### 5.2 Creating a New Block

**1. Click + BLOCK**

This opens a modal to create a new workflow block.

**2. Fill in the Block Label and Attribution Type**

Give the block a label, then pick its category from the Attribution Type list (e.g. Cart, Support, Booking, Form Inputs, Cloud Telephony, Transcription Workflow, Analytical).

**3. Attach the block**

Click **ATTACH** to finalize (enabled once required fields are filled).

![Image](https://mintcdn.com/neuroncx/TgBRzoS1hYf55o0K/images/image.png?fit=max\&auto=format\&n=TgBRzoS1hYf55o0K\&q=85\&s=d3a1f98a2f1ffc8cc249dd0d9ff22cb5)

### 5.3 The Workflow Canvas

Opening a workflow directly (via the arrow next to it in the Assessments list) shows its visual node pipeline — the actual sequence of steps a call goes through.

Example pipeline for a call quality assessment workflow:

1. **UserDetailsAugmentation** — enriches caller/user details
2. **Landing number condition check** — checks the number/routing
3. **Transcribe and qualify for assessment** — transcribes the call and qualifies it for scoring
4. **LLM\_Schema\_Generation** — generates the assessment output per schema
5. Branches into **CallQualityIterativeForms**, **VOX Form Schema Extraction**, and **AssessmentRubricsFunction** — which applies the final rubric scoring logic

Each block can be tested individually (▶), and the whole workflow is finalized with **VALIDATE & SAVE**.

![Image](https://mintcdn.com/neuroncx/TgBRzoS1hYf55o0K/images/image-1.png?fit=max\&auto=format\&n=TgBRzoS1hYf55o0K\&q=85\&s=0b43e8fa78cd69450c7285a811a99df0)

***

## Next Step

Once this setup is complete, results begin to appear automatically. Continue to [Monitor](https://docs.neuroncx.in/assessment/monitor.md) to see how live scores, flags, and coverage show up on the dashboard.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.