Use Case Intelligence RSS Search use cases Search
Mistral Software / SaaS global

Snowflake

Enterprise Knowledge Search

Snowflake needed Cortex Analyst to generate highly accurate SQL against complex enterprise datasets while also explaining answers conversationally for non-technical users and detecting ambiguous questions before bad query results could affect business decisions.

Read the original story Explore all use cases
14% higher accuracy than other competing solutions
2X higher accuracy compared to state-of-the-art LLMs
90% or higher accuracy consistently across customer evaluations

Reported impact

Business outcome

Snowflake used Mistral Large and Codestral to raise Cortex Analyst's text-to-SQL accuracy well beyond competing approaches, making self-service analytics faster and more reliable for non-technical users while setting a higher bar for user experience in generative SQL products.

  • Customer Experience
  • Productivity Improvement

Solution

Snowflake built Cortex Analyst as a structured-data assistant that turns natural-language questions into trustworthy SQL and then explains the results conversationally to business users. The implementation pairs two Mistral models with a multi-agent generation pattern: Mistral Large and a fine-tuned Codestral create candidate SQL for a user's request, then Codestral synthesizes the final query from those expert outputs. Cortex Analyst serves the final SQL together with explanations and reference material so users can understand how the answer was produced. The workflow is designed to run against complex enterprise datasets, detect ambiguous or unanswerable questions, and keep execution accuracy high enough for decision-critical analytics without forcing non-technical teams to hand-author SQL.

Data flow

A business user asks a natural-language question in Cortex Analyst; Mistral Large and a fine-tuned Codestral generate candidate SQL over structured enterprise data, Codestral synthesizes the final query, and Cortex Analyst returns the SQL with conversational explanation, references, and warnings when a question is ambiguous or unanswerable.

Solution architecture

3 components · 1 layer
  1. Serving
    • Cortex Analyst Provides the structured-data assistant interface that returns SQL, context, and conversational explanations to users.
    • Mistral Large Handles reasoning and natural-language generation for candidate SQL creation and conversational response quality.
    • Codestral Generates SQL and synthesizes the final SQL response from multiple agent outputs after fine-tuning for the use case.

Architecture clues

  • Cortex Analyst is designed to identify ambiguous or unanswerable questions and suggest alternatives.
  • Snowflake uses Mistral Large and a finetuned Codestral for initial SQL generation and Codestral again for final synthesis.
  • Snowflake uses a two-step approach in which multiple agents generate SQL and another agent synthesizes the final SQL.

Evidence from the source

"14% higher accuracy than other competing solutions"
"2X higher accuracy compared to state-of-the-art LLMs"
"90% or higher accuracy consistently across customer evaluations"
"Snowflake uses Mistral Large and a finetuned Codestral for the initial SQL generation and uses Codestral for synthesizing the final SQL response."
"Snowflake uses a two-step approach: first, asking multiple agents to generate SQL and then asking another agent to synthesize the final SQL based on the expert responses."
Transparency Evidence & methodology 100/100 reference value · 89% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

01

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value100/100
  • Problem clarity20/20
    Measures whether the customer's operational problem is specific enough to compare.
  • Implementation depth25/25
    Rewards named components, data flow, integrations, and architecture detail.
  • Outcome specificity20/20
    Measures whether the reported benefit is concrete rather than a generic claim.
  • Quantified impact25/25
    3 metrics describe measurable impact.
  • Deployment scale10/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence89%
  • Source authority20/25
    The main evidence is an official vendor-published customer story.
  • Claim traceability21/25
    10 of 12 key fields link to a named source.
  • Direct evidence20/20
    5 direct source quotes retained.
  • Date provenance10/10
    The publication date comes from the source page or publisher.
  • Source transparency8/10
    1 source reference document publisher, URL, authority, retrieval, and supported fields.
  • Implementation specificity10/10
    Rewards source-backed components, data flow, quantified impact, and a specific outcome.
02

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

  1. Mistral

    Snowflake customer story

    First-party
    https://mistral.ai/customers/snowflake/
    Source type
    Vendor-published customer story
    Published
    Apr 24, 2026
    Supports
    • Customer
    • Business problem
    • Solution
    • Products
    • Solution components
    • Data flow
    • Integrations
    • technical area
    • use case category
    • Business outcome
    • outcome category
    • Metrics
    • Architecture
    • evidence quotes
    • Publication date

Date evidence

  • Published on source page Apr 24, 2026 · Mistral · First-party
    Open date source
Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.

Related

Similar use cases

See all like this
Mistral Apr 24, 2026 SAP Software / SaaS global SAP needed to reduce repetitive support demand, break down knowledge silos, and automate business processes for regulated, multilingual customers without sacrificing European data localization or trust in AI-generated… 100% European, locally hosted AI, maintained in SAP-operated infrastructure. Customer Service AutomationEnterprise Knowledge Search Microsoft Teams · Mistral Large · SAP Build Work Zone MongoDB Sep 9, 2026 Squary AI Software / SaaS United States Squary AI needed to eliminate slow, outage-prone legal search infrastructure that combined MongoDB with OpenSearch and hurt document review speed, cost, and trust. At Quatrini Law Group, Squary AI analyzed more than 460,000 pages and saved over 400 attorney and paralegal hours. Document ProcessingEnterprise Knowledge Search Amazon Bedrock · Claude · MongoDB Atlas In-house Sep 1, 2026 Thomson Reuters Software / SaaS global General-purpose LLMs can produce subtle but consequential errors in legal and compliance work, where professionals need trustworthy, citation-backed analysis across very large document sets. As of Q1 2026, CoCounsel has crossed one million users Document ProcessingEnterprise Knowledge Search CoCounsel · Practical Law · Tabular Analysis