Get Faster Business Answers from Snowflake with Supper
Snowflake gives companies a powerful place to centralize and govern business data. But when a question falls outside an existing dashboard, that data can still feel out of reach to the people who need it. They have to send the request to the data team and wait, even when the data needed to answer it is already in Snowflake.
Supper helps Snowflake customers get more business use from the work they've already done. Employees can ask questions in ordinary business language, investigate the answer, and save useful analyses for continued use. The same definitions the data team trusts remain involved throughout the process.
Supper runs on top of Snowflake. Snowflake remains the governed data platform, while Supper gives more employees a practical way to use it. When part of an answer lives in another business system, Supper can include that approved source in the same analysis.
This brings Snowflake data into more daily decisions. Models, Semantic Views, and policies already maintained in Snowflake can support questions across finance, operations, customer teams, and the rest of the business.
Add Supper without changing your Snowflake environment
Supper connects through a customer-provisioned system account and queries the existing Snowflake environment directly. Customers don't have to create another warehouse or copy their Snowflake models into Supper.
The 2 products remain focused on different parts of the experience:
| What Snowflake provides | What Supper adds |
|---|---|
| A governed platform for storing and processing warehouse data | A dedicated analytics product designed for everyday use by business teams |
| Semantic Views that define metrics, dimensions, facts, and relationships | A way to apply those definitions across conversational analysis, reusable Skills, charts, and continued reporting |
| Business context and metadata through Horizon Context | Additional company-specific instructions, known exceptions, and analytical guidance used during execution |
| Native roles, masking, and data-access policies | User-level controls across the data available from Snowflake and connected operational systems |
| Native AI capabilities through Cortex Agents and Snowflake CoWork | A cross-source business analytics experience with Supper Spaces, evaluations, reusable Skills, and centralized administration |
| Data-ingestion options such as Openflow | The ability to use selected operational sources in an analysis before every field has been ingested and modeled in Snowflake |
Every question Supper answers from Snowflake puts the company's governed data and models to work for another business user. The data team continues managing sources and warehouse policies in Snowflake, while Supper handles the employee-facing analytical experience.
Put Snowflake's semantic work into everyday use
Snowflake customers may already have approved metrics and relationships defined through Semantic Views. Supper can ingest those views into its own semantic model and treat approved definitions as canonical.
The data team doesn't have to rebuild logic it already trusts. Supper applies that logic when employees ask questions, continue an analysis through follow-ups, or run recurring Supper Skills.
When more context is required, Supper can add company-specific instructions and known exceptions. That context is reused across the product instead of being reconstructed inside each conversation.
Snowflake explains how Semantic Views define metrics, facts, dimensions, and relationships. Horizon Context organizes governed context for people, BI tools, and AI agents.
A definition maintained in Snowflake can therefore guide more than a warehouse query. Through Supper, it can shape the answers business teams receive, the analyses they save, and the recurring processes they run.
Answer questions that span Snowflake and other systems
Business questions often span several systems. A company may keep product activity and customer history in Snowflake while current operational information still lives elsewhere.
Consider this question:
Which customers with unresolved support escalations also showed a decline in product usage this month?
Product usage may already be modeled in Snowflake. Current escalation status may live in another application. Supper can use both sources and apply the same customer definitions throughout the analysis.
Employees don't need to know where each field is stored. Supper selects the approved sources required for the question and returns a combined answer using that person's permissions.
The data team can continue bringing important sources into Snowflake through existing pipelines or tools such as Snowflake Openflow. Supper lets employees work with selected operational information during that process.
As more data is modeled in Snowflake, Supper can draw more of the answer from the warehouse without changing how employees ask questions. Snowflake stays central to the company's data strategy, and business teams don't have to wait for every integration project to finish.
Give more employees a practical way to use Snowflake data
New questions rarely arrive on a reporting schedule. A customer team may need to investigate an account before a renewal call, while finance may want to understand a variance that appeared that morning.
Supper lets those teams begin directly. It interprets the question using the company's semantic model, queries Snowflake and any other approved sources, and returns an answer they can examine further.
Follow-up questions retain the analytical context. Employees can change the time period, compare groups, inspect the records behind a result, or create a chart. Useful work can be saved in Supper Spaces.
When an analysis becomes a recurring process, Supper Skills preserve the reviewed instructions behind it. Other authorized employees or agents can run the same process without rebuilding the work from a blank conversation.
This expands the audience for Snowflake data without requiring every employee to work directly in the warehouse. More questions are answered using governed Snowflake models, and the data team receives fewer routine requests.
Supper also tests its work against company-specific evaluations. When a question reveals missing context, the semantic model can be corrected so future answers use the updated logic.
Keep Snowflake governance at the center
Snowflake continues to enforce the policies assigned to the Supper system account and role. Those policies define which warehouse data Supper can query.
Supper then applies user-level allow and deny rules within that boundary. The same controls cover approved data from connected operational systems.
An employee may have access to selected customer and product information while remaining restricted from finance or HR data. Supper's Admin Control Center lets data teams review questions, investigate the sources involved, and manage access.
Snowflake remains authoritative for warehouse governance. Supper carries that controlled access into a business-facing experience that can also include approved information from outside the warehouse.
Spend less on company-wide AI analytics
AI analytics can get expensive once usage spreads beyond a small group. Consumption-priced products charge for additional questions and model work. Internal builds require continued engineering and data-team maintenance.
Supper uses flat pricing with unlimited questions. Employees can investigate follow-ups without monitoring credits or worrying that another question will create another charge.
For companies comparing Supper with a consumption-priced AI data product or internal build, Supper and Snowflake can save tens of thousands of dollars per year. The exact amount depends on user count, question volume, model choices, and the work required to maintain the alternative.
Supper keeps costs down by using the semantic work already maintained in Snowflake and sharing analytical logic across the product. Companies don't have to recreate the same business rules for each workflow.
Flat pricing also makes it easier to expand access across departments. As more employees use Supper, more analytical work can run against the governed data the company already maintains in Snowflake.
Provisioning the Snowflake connection generally takes about five minutes of customer work. Supper can usually prepare the initial semantic model and evaluations within 1 or 2 days.
Frequently asked questions
How does Supper connect to Snowflake?
Supper connects directly to Snowflake through a customer-provisioned system account and queries the existing warehouse. Snowflake continues to store, process, and govern the data. Supper gives business teams a dedicated product for analyzing it.
Does Supper replace Snowflake?
No. Snowflake remains the customer's warehouse and governed data platform. Supper runs on top of Snowflake and gives more employees a direct way to use its data.
How does Supper use Snowflake's semantic layer?
Snowflake's semantic layer defines business meaning inside Snowflake. Supper carries those definitions into employee questions, connected operational data, recurring Skills, evaluations, and Supper Spaces.
Can Supper use data outside Snowflake?
Yes. Supper can combine Snowflake data with approved information from connected operational systems. This lets teams answer questions that require both warehouse and current application data.
Does Supper replace Cortex Agents or Snowflake CoWork?
No. Cortex Agents and Snowflake CoWork provide native AI capabilities within Snowflake. Supper sits alongside them as a dedicated product for cross-source business analysis, reusable work, and centralized administration. Snowflake documents how Cortex Agents work with CoWork.
Can Supper and Snowflake reduce AI analytics costs?
They can. Supper's flat pricing, reuse of existing Snowflake work, and lower maintenance requirements can save tens of thousands of dollars per year compared with some consumption-priced products or internal builds. Actual savings depend on the deployment and alternative.
How long does it take to connect Supper to Snowflake?
Provisioning the Snowflake connection generally takes about five minutes of customer work. Supper can usually prepare the initial semantic model and evaluations within 1 or 2 days.