Perplexity Computer adds Snowflake and Databricks data integration

🔑 Enhanced Key Takeaways

•Perplexity Computer's new integration enables non-technical teams across business, product, sales, finance, and operations to conduct complex data analyses, such as pipeline analysis, product usage reviews, customer segmentation, and revenue trend summaries, directly from live enterprise data without needing to write SQL.
•The system automatically generates a 'Data Map' of the connected Snowflake or Databricks environment, which captures key information about the data model, including tables, columns, query patterns, and relationships. This Data Map translates natural language questions into accurate SQL queries and continuously improves through user feedback and administrative edits.
•This advanced data integration is available to Perplexity's Pro, Max, Enterprise Pro, and Enterprise Max subscribers, with robust organization-level controls managed by administrators to ensure data governance and security.
•Perplexity Computer is designed as a cloud-based 'digital worker' that can orchestrate tasks across up to 19 different AI models simultaneously, allowing multi-step workflows to run in the background without requiring continuous user interaction.
•The integration supports structured data types like CSV, JSON, and Parquet-backed tables within Snowflake, and Unity Catalog tables, views, Delta Lake tables, schemas, and external tables in Databricks, but currently does not support unstructured data such as images, audio, or video.

📊 Competitor Analysis▸ Show

markdown

Feature/Platform	Perplexity Computer (with Snowflake/Databricks)	Snowflake Cortex Analyst	Databricks (Mosaic AI, Agent Bricks)	ThoughtSpot
Core Functionality	AI-driven data agent for natural language queries, dashboards, analysis, and automations on live enterprise data. Orchestrates multiple AI models.	Text-to-SQL conversion with strong governance, exclusively for Snowflake data.	Comprehensive AI-first platform for ML lifecycle, deep learning, custom model development, and AI agent deployment.	Self-service analytics with natural language querying and proactive insights.
Data Integration	Direct connectors for Snowflake and Databricks (structured data).	Exclusively with Snowflake data.	Lakehouse architecture, natively supports structured, semi-structured, and unstructured data.	Connects to various data sources, including cloud data warehouses.
Target User	Non-technical business users (sales, finance, operations) for self-service analytics.	Data analysts and developers within Snowflake ecosystem.	Data engineers, data scientists, and ML practitioners for advanced AI/ML workloads.	Business users for self-service BI and insights.
Key Differentiator	'Data Map' for semantic understanding, multi-model orchestration, cloud-based background task execution.	In-database AI functions, strong governance within Snowflake.	Unified platform for data engineering and ML, open model ecosystem, full-stack LLM development.	Natural language search for data, AI-driven insights, and automated root cause analysis.
Visualization	Can produce dashboards and analysis.	Lacks built-in visualization tools.	Integrated AI/BI with Databricks SQL.	Excels at creating polished visualizations.
Pricing Model	Subscription-based (Pro, Max, Enterprise Pro: $40/user/month or $400/year; Enterprise Max: $325/user/month).	Credit-based system (overall Snowflake pricing).	DBU-based ($0.07–$0.55/DBU) plus cloud infrastructure costs.	Per-user/month ($25–$50/user/month).
Benchmarks	N/A	N/A	N/A	N/A

🛠️ Technical Deep Dive

Perplexity Computer operates as a cloud-based AI system, functioning as a 'general-purpose digital worker' capable of deep research, connected tool usage, and persistent memory across tasks.
Its core functionality involves breaking down complex goals into smaller, manageable steps and intelligently routing these sub-tasks to different AI models (orchestrating across up to 19 models simultaneously) based on their specialized capabilities.
The system employs a Retrieval-Augmented Generation (RAG) pipeline that performs real-time web crawling, cleans and ranks reputable sources, and synthesizes responses using a combination of large language models like Claude 3.5, GPT-4o, and Perplexity's proprietary models, always providing inline citations.
The underlying architecture is transformer-based, enhanced with a sophisticated attention mechanism and a scaled feed-forward network, and includes a knowledge integration framework that fuses input context with external memory and cached knowledge representations.
For Snowflake and Databricks integration, Perplexity Computer generates a 'Data Map' by exploring the connected account's schemas, tables, views, and historical usage patterns. This Data Map is securely stored per organization, is versioned, and improves over time through user feedback and direct admin edits.
A partnership with 1Password allows Computer to act on behalf of a user without the underlying credentials ever touching the AI agent or model, ensuring actions are authorized, governed, and auditable.

🔮 Future ImplicationsAI analysis grounded in cited sources

Perplexity Computer will significantly reduce the bottleneck of data science teams for routine analytical requests.

By enabling non-technical users to directly query and analyze live enterprise data via natural language, Perplexity Computer empowers business teams to gain insights without waiting for SQL experts.

Perplexity will expand its 'Computer' capabilities to integrate with a broader array of enterprise applications beyond data warehouses.

Perplexity Computer is already compatible with a range of business applications like Gmail, Outlook, GitHub, Slack, Notion, Salesforce, and HubSpot, indicating a strategic direction to become a more comprehensive 'work layer' across enterprise tools.

The 'Data Map' technology will evolve into a more sophisticated, self-optimizing semantic layer for enterprise data.

The Data Map already learns from user feedback and admin edits, suggesting a path towards increasingly intelligent and autonomous data understanding and query optimization, potentially reducing manual data modeling efforts.

⏳ Timeline

2022-08

Perplexity AI, Inc. founded.

2022-12

Public launch of Perplexity's main search engine.

2023-03

Perplexity Pro launched and raised $25.6 million Series A funding.

2024-01

Raised $73.6 million in Series B, valuing the company at $520 million.

2024-05

Launched Perplexity Pages and Perplexity Enterprise Pro.

2025-01

Completed $500M Series C funding round at a $9 billion valuation.

2025-09

Company valuation reached $20 billion.

2026-02

Perplexity transitioned to a subscription-first business model.

2026-03

Launched Computer for Enterprise and the AI-native browser Comet Enterprise.

2026-05

Perplexity Computer adds direct integration with Snowflake and Databricks.

Perplexity Computer adds Snowflake and Databricks data integration

⚡ 30-Second TL;DR

Key Points

🧠 Deep Insight

🔑 Enhanced Key Takeaways

🛠️ Technical Deep Dive

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline

📎 Sources (13)

👉Related Updates