Skip to main content

Waii Overview

Waii Overview

Introduction: Bridging the Gap Between Humans and Data​

In today's data-driven world, organizations collect vast amounts of information, but extracting meaningful insights remains challenging. Technical barriers separate business users from the data they need, while developers struggle with the complexity of building intuitive data interfaces. This is where Waii comes in.

Waii is an advanced AI platform that bridges the gap between databases and humans/applications. Rather than requiring users to learn complex query languages or developers to build custom interfaces for each use case, Waii provides a natural language layer that transforms how we interact with data.

At its core, Waii offers APIs that enable three fundamental capabilities:

  • Transforming natural language into precise SQL queries
  • Generating insightful visualizations directly from text instructions
  • Extracting meaningful insights from complex datasets through conversation

What makes Waii different? Unlike basic natural language to SQL tools, Waii truly understands enterprise data contexts. We measure accuracy based on capturing user intent rather than just generating syntactically correct SQL. This is possible through our unique knowledge graph approach, which automatically builds understanding of both database structure and domain knowledge, maintained by AI agents that continuously enhance this understanding.

Text to SQL: Beyond Simple Query Generation​

The ability to turn natural language into database queries isn't new, but Waii takes this capability to a new level of sophistication designed for real enterprise environments.

Understanding Intent, Not Just Words​

When a user asks a question, Waii doesn't just pattern-match keywords to columns. It builds a comprehensive understanding of the intent behind the question by leveraging its knowledge graph, which continuously learns from diverse sources including documentation, enterprise systems, user interactions, and external knowledge sources. This deep contextual understanding enables Waii to truly grasp specialized business domains and their unique terminology.

For example, in a healthcare analytics environment, when asked "Show me the 30-day readmission rates for cardiac patients by attending physician, excluding scheduled follow-ups," Waii understands complex domain concepts like "readmission," "cardiac patients," and "scheduled follow-ups." It automatically navigates a complex database schema, joining tables across patient records, admission events, procedure codes, physician assignments, and appointment types.

The resulting query might incorporate multiple joins across a dozen tables, complex filtering logic to distinguish between unplanned readmissions and scheduled care, windowing functions to calculate the 30-day periods, and appropriate aggregations to present the data by physician—all without requiring the user to understand the underlying data structure or SQL complexity.

Complete Transparency and Control​

Every generated query comes with:

  • A list of sources that justify why particular tables and columns were selected, and why particular operations were performed
  • Explicit assumptions made to disambiguate the query intent
  • Step-by-step descriptions of what the query does in either domain-specific or technical language
  • The ability to ask clarifying questions when assumptions aren't sufficient
  • A confidence score measuring how well the query is understood and related to the knowledge graph.
Intelligent Data Discovery

This transparency builds trust in the system and helps users understand the logic behind each result.

Guaranteed Execution​

One of the most frustrating aspects of AI-generated queries is when they fail to run. Waii's patent-pending compiler ensures every query is either executable or provides a clear explanation of limitations. Common DBMS issues are automatically identified and resolved through our optimizer, saving time and frustration.

Enterprise-Ready Database Support​

Waii provides deep integration with the databases that power modern enterprises:

  • PostgreSQL, MySQL, Trino, Snowflake, Databricks, Redshift, SQL Server, and more
  • Automatic selection of appropriate tables, columns, UDFs, keys, and indices
  • Context optimization to ensure each interaction fits within model limitations

Flexible LLM Integration​

Organizations have different preferences and requirements for AI models. Waii supports all major LLMs with the ability to:

  • Configure specific endpoints for different operations
  • Apply model-specific optimizations for prompts, settings, and capabilities
  • Control cost/performance tradeoffs at a granular level

Beyond Basic Queries: Advanced Data Management​

Waii goes beyond simply generating queries to provide a comprehensive data management solution.

Virtual Views for Abstraction​

Complex data models can be simplified through virtual views that exist within Waii rather than in the database itself. These views:

  • Abstract away complexity for end users
  • Maintain performance through intelligent compiler expansion
  • Provide a logical layer that can remain consistent even as the physical model evolves

Enterprise-Grade Security​

Data access must respect organizational security boundaries. Waii provides:

  • The ability to inject security filters across multiple tables
  • Column-level masks and hiding by user, tenant, or organization
  • Seamless integration with existing access control mechanisms
Access Policy

Intelligent Data Discovery​

Finding the right data is often the first challenge. Waii offers:

  • A dedicated table search API for discovering relevant tables by concept
  • Column-level search for specific data elements
  • Knowledge graph exploration to understand data relationships
  • Semantic search capabilities with confidence scoring
Knowledge Graph Visualization

Query Migration and Documentation​

As data environments evolve, Waii helps maintain continuity:

  • Migrate queries between different database systems (e.g., Snowflake to Databricks)
  • Support schema variations and different data modeling approaches
  • Generate comprehensive documentation for existing queries
  • Maintain complete interaction history for learning and improvement

Visualization: From Data to Insight​

Raw data rarely tells the complete story. Waii's visualization capabilities transform data into meaningful insights that drive decision-making.

Waii generates visualization specifications from datasets or queries, supporting multiple output formats including Vega(Lite), Pyplot, Metabase, Tableau, and Looker. This flexibility allows integration with your existing visualization tools.

Our system intelligently selects the appropriate chart type from a comprehensive library:

  • Line, bar, pie, and donut charts for comparing values
  • Heatmaps and faceted charts for multidimensional analysis
  • Multi-axis, bubble, area, and scatter plots for relationship exploration
  • Histograms and box-and-whisker plots for distribution analysis
Side by Side Chart

What sets Waii's visualizations apart is the attention to detail. Charts are automatically optimized with appropriate titles, tick marks, grid lines, legends, scales, colors, and interactive elements. This eliminates the time-consuming fine-tuning typically required when creating visualizations manually.

The Conversational Interface: Chat with Your Data​

While APIs provide powerful capabilities for developers, many users prefer a conversational interface. Waii's chat functionality leverages all the capabilities described above within an intuitive dialog format.

Chat with Data

The chat interface enables users to:

  • Ask questions about their data in natural language
  • Receive answers enhanced with appropriate visualizations
  • Explore assumptions and refine their questions
  • Understand the reasoning behind each answer

Each response provides:

  • Clear explanations of operations performed
  • Transparent assumptions used to generate the answer
  • Knowledge graph information leveraged for the response
  • Interactive clarification when needed

This conversational approach makes data accessible to a broader range of users without sacrificing depth or accuracy.

Integration: Fitting Into Your Ecosystem​

Waii is designed to integrate seamlessly with your existing tools and workflows through multiple pathways:

For Developers​

  • RESTful APIs for programmatic access to all capabilities
  • SDKs for Python, JavaScript, TypeScript, and Java
  • Command-line interfaces for scripting and automation
  • Direct integration with frameworks like Langchain and LlamaIndex
  • Support for Model Context Protocol (MCP) by Anthropic for enhanced AI model integration
Claude MCP Integration

For Applications​

  • React widgets that can be embedded in web applications
  • iFrame embedding options for simpler integration
  • Standalone components for quick deployment
  • Configurable visual elements including SQL editors, chatbots, and history viewers

For Collaboration Platforms​

  • Slack integration for data queries and insights directly in your team conversations
  • Microsoft Teams integration for seamless data access within your collaboration environment
Slack Integration

This flexibility ensures Waii can be implemented where it delivers the most value, whether that's enhancing existing applications, enabling new workflows, or bringing data insights directly to the platforms where your teams already collaborate.

Deployment: Your Data, Your Way​

Organizations have different requirements for data security, governance, and management. Waii offers deployment options to meet these diverse needs:

Self-Hosted Solutions​

  • Kubernetes deployment for scale and resilience
  • Single-node installation for simpler environments
  • Docker container support for standardized deployment

Software as a Service​

  • Fully managed hosting by Waii
  • Simplified setup and maintenance
  • Continuous updates and improvements

Both options provide the same powerful capabilities while giving you control over your data environment.

The Knowledge Graph: Waii's Intelligent Foundation​

At the heart of Waii's capabilities is its knowledge graph—a comprehensive representation of your data environment that grows more intelligent over time.

Automatic Understanding​

Waii automatically extracts semantic information from:

  • Database metadata and structures
  • Data patterns and relationships
  • System tables and constraints
  • User-defined functions and business logic

This creates a baseline understanding that grows with each interaction.

Continuous Learning​

The knowledge graph evolves through multiple pathways:

  • Document integration: Automatically parse documentation (HTML, PDFs, docs) to enhance understanding
  • External system integration: Import from semantic layers, data catalogs, and BI tools
  • Training from examples: Learn from question/query pairs to improve accuracy
  • User feedback: Incorporate corrections and improve over time

This continuous learning approach ensures Waii gets more accurate and valuable with each interaction.

Knowledge Sharing​

The knowledge graph isn't a black box—it's accessible through:

  • APIs for programmatic access
  • YAML metadata file export and import
  • Data catalog enhancement and integration

This transparency gives you control over how Waii understands your data environment.

Enterprise-Ready Performance​

Waii is built for enterprise-scale deployments with features that ensure reliability, performance, and governance.

Benchmarking and Accuracy​

The Archerfish benchmarking framework allows you to:

  • Run text-to-SQL workloads with accuracy reporting
  • Generate automatic benchmarks and test data
  • Ensure performance meets your specific requirements

Intelligent Caching​

Waii's sophisticated caching architecture:

  • Stores LLM interactions to reduce latency and costs
  • Caches previous answers for similar questions
  • Detects semantic equivalence between questions
  • Makes minimal modifications to adapt previous answers to new contexts

Multi-Tenant Support​

For organizations with complex user ecosystems, Waii provides:

  • Support for multiple organizations, teams, and users
  • External/federated identity integration
  • Single Sign-On (SSO) compatibility
  • Permission-based logical knowledge graphs
  • Optimized sharing where security controls permit

Conclusion: Transforming Data Interaction​

Waii represents a fundamental shift in how humans and applications interact with databases. By bridging the gap between natural language and data, we enable:

  • Business users to directly access the insights they need without technical barriers
  • Developers to build more intuitive data experiences with less effort
  • Organizations to extract more value from their existing data investments

The result is faster decision-making, reduced development costs, and more accessible data insights throughout your organization.

Resources​

For more detailed information about Waii's capabilities and implementation, please visit:

Ready to transform how your organization interacts with data? Contact us today to learn how Waii can fit into your environment.