The Library

Product Introduction

1. Definition: The Library by Querri is a semantic data modeling and governance layer within an agentic AI data analytics platform. It is a centralized knowledge base that automatically organizes and contextualizes raw data from disparate sources (like CRMs, ERPs, and ad platforms) around business-specific questions, definitions, and rules.
2. Core Value Proposition: It exists to eliminate data ambiguity and inconsistency in AI-driven analytics. The Library creates a single, trusted source of truth by learning a company's unique business context, ensuring that every AI-generated answer, dashboard, and report is accurate, consistent, and grounded in how the business actually operates, thereby preventing AI hallucinations in data analysis.

Main Features

1. Semantic Data Modeling: The Library ingests data from over 130+ read-only connectors (e.g., Salesforce, Shopify, Google Ads, PostgreSQL) and allows users to define business logic in plain English. Users can establish "Facts & Rules," such as defining what constitutes an "active customer" or how "lifetime value" is calculated. This semantic layer translates technical data schemas into business-friendly terms, which the AI then uses to generate correct SQL queries and analyses.
2. Context-Aware AI Grounding: This is the core technical mechanism. When a user asks a question in plain English, Querri's AI doesn't query raw data directly. Instead, it first consults The Library's learned context—definitions, KPIs, historical analysis, and data relationships. The AI grounds its reasoning in this context before generating and executing a query, ensuring answers are relevant and accurate. The interface visually "lights up" the data sources used, providing transparency.
3. Centralized KPI & Metric Governance: Teams can define Key Performance Indicators (KPIs) once within The Library using natural language descriptions. Once defined, these KPIs are calculated consistently across every output channel—whether in an AI chat response, an embedded dashboard, a PowerPoint deck, or via the API/CLI. This eliminates metric sprawl and ensures alignment across finance, marketing, and operations teams.

Problems Solved

1. Pain Point: The "multiple versions of the truth" problem in data analytics, where different departments use different definitions and data sources, leading to conflicting reports and mistrust in data. Additionally, it solves the problem of AI chatbots "hallucinating" or providing generic, context-free answers when asked business-specific questions.
2. Target Audience: Non-technical business users (CFOs, Marketing Managers, Revenue Operations, Customer Success), IT & Data Governance teams who need to maintain control and security, and SaaS companies & consultants looking to embed or white-label analytics for their clients.
3. Use Cases: A CFO needs a last-minute variance analysis for a board meeting and asks, "Why did Q3 revenue deviate from forecast?" The Library ensures the AI uses the correct revenue recognition rules and data sources. A marketing manager compares ROAS across platforms; The Library ensures "ad spend" and "revenue" are defined consistently for Meta Ads and Google Ads. A SaaS company embeds The Library's governed analytics into their product for client-facing dashboards.

Unique Advantages

1. Strengths & Limitations (Pros & Cons):

   • Pros: Dramatically reduces time-to-insight for non-technical users (from months to minutes). Enforces a single source of truth, improving data trust. Provides enterprise-grade governance (SOC 2, ISO 27001, HIPAA) and row-level security that propagates to all AI interactions. The "teach it once" model reduces repetitive data definition work.
   • Cons: As a platform, it requires initial setup and context-building in The Library to reach its full potential; value is not instantaneous from a raw data connection. It is a managed SaaS solution, so it may not fit organizations with strict requirements for on-premise deployment of core analytics logic. The AI's accuracy is inherently tied to the quality and completeness of the business logic defined within it.

2. Key Alternatives & Differentiation:

   • Traditional BI Tools (Tableau, Power BI): These require significant SQL/modeling expertise to build reports and dashboards. The Library differentiates by allowing business logic definition in plain English and exposing it through an AI chat interface, making analytics accessible without pre-built dashboards.
   • AI Chatbots for Data (ChatGPT with plugins, Microsoft Copilot for Power BI): These often operate as thin interfaces over data without a persistent, governed business context layer. They can hallucinate or provide inconsistent answers. The Library's core differentiation is its persistent, teachable "brain" (The Library) that grounds all queries, ensuring consistency.
   • Custom Semantic Layer Solutions: Building a similar context-aware layer in-house requires significant data engineering and ML expertise. Querri provides this as an integrated, off-the-shelf platform with a pre-built AI agent, reducing development time from quarters to days.

Frequently Asked Questions (FAQ)

1. How does Querri's Library prevent AI hallucinations in data analysis? The Library acts as a grounding layer. Before answering a question, Querri's AI agent first references the business definitions, rules, and data relationships stored in The Library. It uses this specific context to formulate accurate SQL queries against the connected data, ensuring answers are derived from trusted sources and defined logic, not from the AI's general training data.
2. What data sources does The Library by Querri connect to? Querri offers read-only connectors to over 130+ common business applications and databases, including Salesforce, HubSpot, Google Ads, Meta Ads, Shopify, QuickBooks, Snowflake, PostgreSQL, and MySQL. This allows The Library to create a unified semantic model across marketing, sales, finance, and operational data.
3. Can I use Querri's Library inside other AI tools like ChatGPT or Claude? Yes, via Querri's Model Context Protocol (MCP) server and CLI. This allows you to connect your governed data from The Library to AI assistants like Claude Desktop, enabling you to ask business questions in your preferred chat interface while still receiving answers grounded in your company's specific data context and security rules.
4. How is data security and access control handled in The Library? Querri is SOC 2 Type II, ISO 27001, and HIPAA compliant. The Library enforces row-level and workspace-based security filters set by IT administrators. These filters are applied at the data layer, meaning every query—whether from chat, a dashboard, or an API call—inherently respects user permissions, ensuring individuals only see data they are authorized to view.
5. What is the difference between a KPI in The Library and a standard dashboard metric? A KPI defined in The Library is a centrally governed business metric. Once defined with its calculation logic (e.g., "Monthly Recurring Revenue is the sum of active subscription fees as of the last day of the month"), it is calculated consistently everywhere it appears. A standard dashboard metric is often defined within a single report, leading to potential inconsistencies if the same metric is recreated elsewhere with slight variations in logic.