Dynamic Dataset Columns: Enrich Rows with AI-Generated Data

Use Dynamic Columns to add AI-generated summaries, sentiment labels, extracted entities, vector-retrieved context, parsed JSON fields, and conditional routing to any dataset — no code required.

📝
TL;DR

Dynamic Columns let you enrich any dataset with AI-generated data — summaries, sentiment labels, entities, vector-retrieved context, parsed JSON fields, and conditional routing — directly from the FutureAGI dashboard, no code required.

TimeDifficultyPackage
15 minBeginnerDashboard only

By the end of this guide you will have a support ticket dataset enriched with AI-generated summaries, sentiment labels, extracted entities, vector-retrieved context, parsed JSON fields, and conditional routing: all populated automatically across every row.

Prerequisites
  • FutureAGI account → app.futureagi.com
  • An LLM API key configured in the platform (GPT-4o, Gemini, etc.)
  • For the Retrieval column: a vector database (Pinecone, Qdrant, or Weaviate) with data already indexed

Starter dataset

Save as support_tickets.csv and upload via Dataset → Add Dataset → Upload a file (JSON, CSV).

ticket_id,customer_message,agent_response,response_metadata,priority
T001,My laptop battery drains in under 2 hours even after a factory reset.,I have opened a replacement request. You will receive a prepaid shipping label within 24 hours.,"{""resolution_time_hours"": 24, ""replacement_approved"": true, ""ticket_owner"": ""support_team""}",high
T002,I never received my order from three weeks ago. Tracking says it was delivered.,I see the delivery was marked complete. Let me file a lost shipment claim with the carrier right now.,"{""resolution_time_hours"": 48, ""replacement_approved"": false, ""ticket_owner"": ""logistics_team""}",high
T003,The app crashes every time I try to open my account settings.,This is a known issue in version 3.2. Please update to version 3.3 using the link below.,"{""resolution_time_hours"": 2, ""replacement_approved"": false, ""ticket_owner"": ""engineering_team""}",medium
T004,Can I change my subscription plan before the billing date?,Absolutely! You can change your plan at any time from Account > Billing. The new rate applies from the next billing cycle.,"{""resolution_time_hours"": 0, ""replacement_approved"": false, ""ticket_owner"": ""billing_team""}",low
T005,I was charged twice for the same order last month.,I have confirmed the duplicate charge and initiated a full refund. It will appear in 3-5 business days.,"{""resolution_time_hours"": 72, ""replacement_approved"": true, ""ticket_owner"": ""billing_team""}",high

Tutorial

Run Prompt — generate summaries

  1. Add Column → Dynamic Columns → Run Prompt
  2. Column Name: summary
  3. Model Type: LLM, Select Model: gpt-4o-mini
  4. System: You are a support assistant that writes short, factual summaries.
  5. User: Summarize this customer support message in one sentence: {{customer_message}}
  6. Concurrency: 5
  7. Test → Run

Classification — label sentiment

  1. Add Column → Dynamic Columns → Classification
  2. Column Name: sentiment, Column (source): customer_message
  3. Add labels: Positive, Neutral, Negative
  4. Choose a model, Concurrency: 5
  5. Test → Run

Extract Entities — pull structured data

  1. Add Column → Dynamic Columns → Extract Entities
  2. Name: entities, Column: customer_message
  3. Instructions:
Extract: issue type, product, urgency level, and location (if present).
Return concise entity values.
  1. Choose a model, Concurrency: 5
  2. Test → Run

Retrieval — fetch context from a vector database

Note

Requires an external vector database (Pinecone, Qdrant, or Weaviate) with data already indexed; this is separate from FutureAGI’s Knowledge Base.

  1. Add Column → Dynamic Columns → Retrieval
  2. Select your Vector Database type

Pinecone:

FieldValue
Columnagent_response
Pinecone API KeyYour API key
Index NameYour index
Number of chunks to fetch3
Key to extracttext
Concurrency5

Qdrant:

FieldValue
Columnagent_response
Qdrant API KeyYour API key
Qdrant URLYour instance URL
Collection NameYour collection
Number of chunks to fetch3
Key to extracttext
Concurrency5

Weaviate:

FieldValue
Columnagent_response
Weaviate API KeyYour API key
Weaviate Cluster URLYour cluster URL
Collection NameYour collection
Number of chunks to fetch3
Search TypeSemantic Search or Hybrid
Key to extractcontent
Concurrency5
  1. Test → Run

Extract JSON Key — parse a field from JSON

  1. Add Column → Dynamic Columns → Extract a JSON Key
  2. Column Name: resolution_time, Column: response_metadata
  3. JSON Key: $.resolution_time_hours
  4. Concurrency: 5
  5. Run

Note

The source column must have data type JSON. If it’s stored as text, change its type to JSON using Edit Column first.

Conditional Node — route by priority

  1. Add Column → Dynamic Columns → Conditional Node
  2. Column Name: triage_output

Branch 1 (if):

  • Condition: {{priority}} == "high"
  • Column Type: Run Prompt
  • System: You are a senior support analyst.
  • User: Write a detailed triage summary and next action for this high-priority ticket: {{customer_message}}

Branch 2 (else):

  • Column Type: Run Prompt
  • System: You are a support assistant.
  • User: Write a one-line summary for this ticket: {{customer_message}}
  1. Concurrency: 5
  2. Test → Run

Tip

You can add elif branches between if and else for more granular routing; each branch supports all column types: Run Prompt, Retrieval, Extract Entities, Extract JSON Key, Classification, and API Calls.

Run evals on the enriched dataset

  1. Click Evaluate → Add Evaluations → select groundedness
  2. Map: input → customer_message, output → agent_response
  3. Add & Run

For all six dynamic column types in detail, see Create Dynamic Column.

What you built

You can now enrich any dataset with AI-generated columns, vector-retrieved context, parsed JSON fields, conditional routing, and inline evaluations — all from the FutureAGI dashboard.

  • Generated summaries with Run Prompt
  • Labeled sentiment with Classification
  • Extracted entities with Extract Entities
  • Fetched vector DB context with Retrieval
  • Parsed JSON fields with Extract JSON Key
  • Routed by priority with Conditional Node

Next steps

Dataset Management

Create and manage datasets

Synthetic Data Generation

Generate synthetic rows

Experimentation

A/B test prompts

Dynamic Column Docs

Full column type reference
Was this page helpful?

Questions & Discussion