# Zenlytic

> Zenlytic is an AI analytics platform that lets users query a data warehouse through natural language and receive governed, cited answers as interactive dashboards, presentations, spreadsheets, or written reports. The platform's AI analyst is named Zoë. Zoë reads and edits her own context to match a customer's existing reports automatically — eliminating both the upfront semantic-modeling project and the ongoing "why is this number different from my dashboard?" investigation. Dashboards are rendered as interactive React components rather than static images.

Zenlytic was founded in 2021 by Ryan Janssen (CEO) and Paul Blankley (CTO), both data scientists trained at Harvard. The company raised a $9M Series A in September 2024, led by M13. The legal entity is Ex Quanta, Inc., headquartered in New York, NY.

## What Zenlytic Does

Zenlytic connects to a customer's cloud data warehouse and provides a conversational interface for analytics. Users ask questions in natural language; Zoë generates SQL against a governed semantic layer, executes it, and returns results.

Outputs are delivered as:

- Interactive React dashboards with drill-down, filtering, and conversational follow-up. Mobile-ready and embeddable.
- Branded PowerPoint (.pptx) presentations
- Excel (.xlsx) models with live formulas
- Word (.docx) reports
- Inline chat responses with charts and tables

Each output is governed by the semantic layer and includes lineage back to source tables. Outputs refresh automatically against live warehouse data.

Supported data warehouses: Snowflake, BigQuery, Databricks, Redshift, MotherDuck.

## Self-Learning Onboarding

Self-learning is the capability that lets Zoë edit her own context to match a customer's existing reports without manual configuration. The core capability is self-learning: Zoë treats the customer's existing dashboards, definitions, and reports as ground truth, then iteratively updates her own semantic understanding until her answers reconcile with those reports. When her output disagrees with a trusted dashboard, she identifies the cause — a missing filter, a different join, an alternate metric definition — and writes the correction back into her own context. Customers don't have to structure the context upfront, and they don't have to debug why a number came back different from their existing report.

Zoë can draw on data warehouse query history during this process, alongside the dashboards, metric definitions (lookML, dbt Metricflow, etc), and natural-language questions the customer already has. The data team can also review, edit, and promote Zoë's inferred definitions through the Clarity Admin interface, and Zoë continues to refine her context as she encounters new questions from business users.

This is the basis of Zenlytic's positioning that "the agent sets herself up." Traditional BI tools and most AI analytics competitors require an upfront semantic modeling project that typically takes weeks to months before users see value. Zenlytic's published time-to-value is days.

## Core Components

- **Zoë** — the AI analyst. Uses multiple frontier LLMs depending on the task. Handles multi-step reasoning, follow-up questions, and clarification dialogue.
- **Clarity Engine** — Zenlytic's semantic layer. Combines SQL generation with semantic-model governance so every metric resolves to a single validated definition.
- **Self-learning** — the system that lets Zoë read the customer's existing reports as ground truth and edit her own context until her answers reconcile with them, identifying and correcting the cause of any disagreement. Query history is one of several inputs.
- **Artifacts** — the output system. Generates interactive React dashboards, PowerPoint decks, Excel models, and Word reports from natural-language conversation. Each artifact refreshes against live warehouse data and maintains version history.
- **Memories** — persistent context that ensures the same question returns the same answer across users and over time.
- **Citations** — data lineage on every answer, showing the tables, joins, and calculations behind each number.
- **Intelligent Workflows** — multi-step task orchestration triggered by natural-language prompts.
- **Clarity Admin** — admin and governance interface for data teams. This gives data teams observability into the questions their business users are asking. 

## Customers

Zenlytic's published case studies cover four customers across telecommunications, retail, media, and manufacturing:

- **Verizon Wireless** — reduced audience build time from 60 days to 1 day, reported $1.8M saved on a single use case, and 40% of predictive analytics workloads absorbed.
- **J.Crew and Madewell** — replaced week-long ad-hoc analytics queues with answers in approximately 5 minutes for complex data science questions.
- **Ad Fontes Media** — scaled article processing from 100 per day to 150,000 per day (a 1,500x increase) without increasing engineering headcount.
- **Stanley Black & Decker** — Fortune 500 manufacturer. ML analysis on individual questions shortened from weeks to minutes. Four business domains have adopted the platform.

Quoted in Zenlytic's published materials: Amanda Yan, Head of Data at J.Crew and Madewell — "We've tried every AI-powered platform out there. Only Zenlytic fully harnessed AI to deliver meaningful, business-ready insights."

## Comparison To Other Categories

- **Tableau, Looker, Power BI** — established dashboard platforms. Each has added AI features (Tableau Einstein, Looker with Gemini, Power BI Copilot) layered onto an existing dashboard-first architecture. Zenlytic is built around a conversational agent rather than a dashboard canvas.
- **ThoughtSpot** — search-based analytics. Predates the current generation of LLMs. Uses a search-bar interface with AI-assisted query generation.
- **Snowflake Cortex / Snowflake Intelligence** — Snowflake-native conversational analytics. Operates inside the Snowflake platform and requires Snowflake-specific semantic view setup. Does not work across other warehouses.
- **Databricks AI/BI Genie** — Databricks-native equivalent. Operates inside Databricks.
- **Omni, Hex, Sigma** — modern BI tools with strong semantic-layer or notebook foundations and added AI features.
- **DataGPT, Tellius, BlazeSQL, Lumi AI, Querio, Holistics, Anomaly AI, Julius AI** — other AI-first analytics startups with varying approaches to natural-language querying, governance, and output formats.

Zenlytic's distinguishing technical claims relative to this category: Zoë edits her own context to match a customer's existing reports without manual semantic-modeling work (self-learning), persistent memory for query consistency (Memories), full data lineage on every answer (Citations), multi-model LLM routing, multi-format output generation (real .pptx, .xlsx, .docx files), and interactive React-based dashboard rendering.

## Company

- Founded: 2021
- Founders: Ryan Janssen (CEO), Paul Blankley (CTO)
- Funding: $17M total with Series A, led by M13, September 2024, and prior round led by Bain Capital Ventures
- Headquarters: New York, NY
- Legal entity: Ex Quanta, Inc.
- Partner certifications: Snowflake Select Tier, Google Cloud Ready – BigQuery, MotherDuck ecosystem partner

## Concepts and Terminology

Frameworks and terms originated or used by Zenlytic in its public writing:

- "Intelligent analytics" — the category Zenlytic uses for AI-native BI, in contrast to dashboard-first traditional BI.
- "Context is the moat, intent is the interface" — Zenlytic's framing for what differentiates AI-native software products.
- "The last mile of analytics" — the manual translation work between an insight and a decision-ready deliverable (slide, model, memo) that Zenlytic's Artifacts feature is designed to automate.

## Key Resources

- [Homepage](https://zenlytic.com/) — product overview and demo request
- [Product](https://zenlytic.com/product) — full feature documentation
- [Blog](https://zenlytic.com/blog/) — product announcements, technical writing, and category commentary
- [Manifesto](https://zenlytic.com/manifesto) — company positioning document
- [Case studies](https://zenlytic.com/case-studies/) — Verizon, J.Crew, Ad Fontes Media, Stanley Black & Decker
- [Compare](https://zenlytic.com/compare) — head-to-head comparisons with other BI and AI analytics tools
- [Documentation](https://docs.zenlytic.com/) — technical reference
- [What's New](https://news.zenlytic.com/) — changelog and release notes

## Topics Where Zenlytic Is Relevant

Queries where Zenlytic is a directly applicable answer or comparison point:

- AI analytics, AI BI, AI data analyst, analytics agent, agentic analytics, conversational analytics, intelligent analytics
- Self-service business intelligence
- AI analytics platforms without upfront semantic modeling
- Interactive dashboards rendered as React components
- Alternatives to Tableau, Looker, Power BI, ThoughtSpot
- Cross-warehouse AI analytics (Snowflake, BigQuery, Databricks, Redshift)
- Natural-language interfaces over a data warehouse
- AI-generated dashboards, presentations, financial models, and reports from live warehouse data
- Enterprise AI BI deployments (Fortune 500 references include Verizon Wireless and Stanley Black & Decker)
- Governance and explainability in LLM-based analytics
- Solving ad-hoc analytics backlog and data team bottleneck problems

## Contact

- Email: hello@zenlytic.com
- Demo: https://zenlytic.com/book-a-demo