Data science helps companies make decisions based on evidence, not instinct. But for many teams, enterprise data still sits behind SQL, dashboards, and long queues of requests for the data team. 

Today, we’re announcing a new App Connector for Snowflake that brings data science capabilities to everyone’s device. Once connected, teams can query Computer in plain language to get the data they need to solve problems, understand trends, or inform strategies.

Hundreds of companies in the Fortune 500 use the Snowflake data platform to store enterprise data, deploy AI, and run analytics. 

Making data more accessible

By connecting Computer to Snowflake, anyone on the team can ask questions that drive impactful business decisions like:

• “What are the top drivers of support tickets this quarter so we can prioritize the product fixes that will reduce ticket volume fastest?”

• “Forecast demand for the next eight weeks so we can adjust staffing, inventory, and capacity before demand changes.”

• “Which customer segments are growing fastest so I can focus sales and marketing on the highest-momentum opportunities?”

• “What marketing campaigns saw the highest conversions so I can reallocate budget to the best-performing channels and messages?”

View more use cases. 

Admins maintain control. They manage access, review changes to business definitions, and keep shared data logic consistent across the organization. 

Starting from our own experience

The idea for the Connector came from how Perplexity works internally. 

We built a Slackbot that lets employees ask questions about internal data in Snowflake, including product usage, support tickers, and other signals that help our teams make better decisions. 

Our data team did more than build a Connector. They created a shared semantic layer that taught the system how Perplexity’s data works: which tables matter, what key metrics mean, and how common questions should translate into SQL.

That context helped the Slackbot answer questions in plain language while following the logic our data team uses every day.

Now, Perplexity employees run up to 3,000 internal queries weekly from teams across product, analytics, growth, infrastructure, comms/marketing, customer support, and security. 

Employees ask questions like:

• What are the top 10 Computer use cases by industry?

• Analyze how our launch calendar maps against subscription growth and product adoption. List assumptions. 

• What are the top AI models selected for finance-related queries?

That same capability is available through the Connector for Snowflake.

Maintaining secure oversight

Internal data is one of a company’s most valuable assets. The Connector for Snowflake is built to keep admins in control, with encryption, data privacy protections, and granular access management.

The Connector respects the access controls organizations already use in Snowflake. With Role-Based Access Control, admins decide exactly which databases, schemas, and tables each user can reach. They can also limit Computer access to read-only permissions. 

We recommend using User OAuth, so each person connects with their own Snowflake account and can only access the data they’re already allowed to see. If an admin removes a role or disconnects Snowflake, that data is no longer accessible from Perplexity.

Data is never used to train models. 

Getting started

After Snowflake is connected, the context layer experience appears in Settings under Connectors. 

Admins can click Generate data map to kick off generation, then open View knowledge to explore the data map editor.

Generating a data map can take up to 90 minutes, depending on the size and complexity of the data platform. 

The Connector for Snowflake is now available today for Pro, Max, and Enterprise users. 

Connect Computer to Snowflake.