Capture Requirements (BRD Agent)
The BRD Agent conducts an AI-driven interview to capture business requirements. It reads your uploaded documents, understands your Schema Agent context, and asks targeted questions to generate a formal Business Requirements Document.
Complete the Create Model step first to set up your Semantic Model.
Starting the Interview
Navigate to the BUSINESS REQUIREMENTS tab to begin:
The AI Agent Chat shows:
"Welcome, this is the BRD Agent here to help you clarify requirements, define metrics, and ensure your dashboard or data model aligns perfectly with your business needs."
"Determining confidence in available information..."
The Q&A Process
The BRD Agent asks targeted questions based on:
- Your model requirements description
- Uploaded context documents
- Schema Agent ERD and table information
Question Format
Each question includes:
| Element | Description |
|---|---|
| Question | AI-generated question about requirements |
| Answer Box | Your response text |
| Status | Approved/Pending indicator |
| Actions | Reject / Approve / Add Comments |
Example Q&A Session
Question 1:
"For the customer card activation tracker, should we compute the activation speed as the difference (e.g., in days) between the card's issuance (using the CREATED_DATE in Worldline Cards) and its ACTIVATION_DATE, and then calculate the activation rate as the percentage of issued cards that get activated within a specified timeframe?"
Your Answer:
"Yes that's correct"
Question 2:
"For segmenting the results, can we use the CUSTOMER_SEGMENT field from Temenos Customers to define customer groups and the CARD_TYPE field from Worldline Cards for card categorization? Also, should the analysis apply to all issued cards or would you prefer to restrict it to a recent issuance period?"
Your Answer:
"Yes, use CUSTOMER_SEGMENT from Temenos Customers and CARD_TYPE from Worldline Cards. Include all issued cards across the full date range to show trends over time."
Question 3:
"Regarding data quality, are there any specific rules to apply such as ensuring that both the activation and issuance dates are properly formatted and non-null (with null activation dates considered as not activated), or any other conditions we should enforce?"
Your Answer:
"Null ACTIVATION_DATE means not activated. Exclude cards with null CREATED_DATE. Cards where ACTIVATION_DATE is before CREATED_DATE are data errors - flag but exclude from metrics."
Review Actions
For each Q&A pair:
| Action | When to Use |
|---|---|
| Approve | Answer is complete and accurate |
| Reject | Need to revise the answer |
| Add Comments | Provide additional context |
After approving all questions, click Send To AI to generate the BRD.
Generated BRD Document
BRD Structure
The generated Business Requirements Document includes:
1. Executive Summary
- Business Problem — What issue the model addresses
- Proposed Solution — How the model solves it
- Objective — Measurable goals
2. Business Context
- Business Drivers — Why this matters
- KPIs & Functional Requirements — What to calculate
3. Detailed Requirements
- Metrics Definitions — Formulas and logic
- Data Quality Rules — Validation requirements
- Security Requirements — Access controls
Example BRD Output
# Business Requirements Document: Customer Card Activation Tracker
## 1. Executive Summary
**Business Problem:**
The business currently lacks visibility into the friction customers
experience immediately after receiving a new card. There is no
consolidated view to track how quickly cards are being activated
after issuance.
**Proposed Solution:**
Develop a "Customer Card Activation Tracker" using dbt. This data
product will analyze the time lag between card issuance and
activation, enabling the product team to identify bottlenecks.
**Objective:**
Compute activation speeds and rates, segmented by customer profile
and card type, to improve the overall customer onboarding experience.
## 2. Business Context
### 2.1 Business Drivers
- **Onboarding Friction Analysis:** Identify customer segments or
card types that experience delays in activation
- **Trend Analysis:** Monitor activation performance over time
### 2.2 Key Performance Indicators (KPIs)
- **Activation Speed (Lag):** ACTIVATION_DATE - CREATED_DATE (days)
- **Activation Rate:** (Cards with ACTIVATION_DATE / Total Cards) × 100
Team Collaboration
Multiple team members can participate in the BRD process:
- Multiple users can answer questions
- All contributions are attributed
- Use Share button to add collaborators
Tips for Better BRDs
| Tip | Result |
|---|---|
| Be specific | Clear metrics definitions |
| Include edge cases | Better data quality rules |
| Reference source columns | Accurate transformations |
| Define business terms | Complete glossary |
Next Steps
- AI Modeling — View generated artifacts
- Execute DBT — Build the data model