Case studies / Data agents

Data agents

Agents that answer business questions over a company's data, metrics and dashboards.

Approach

20 write-ups

Databricks11 Nov 2026NewData agents
The new Genie: autonomous AI for complex enterprise work
Problem Employees had to manually track and act on long‑running, data‑heavy enterprise tasks, which was time‑consuming and error‑prone.
Solution They created Genie, an autonomous AI coworker that uses an ontology for business context, searches the lakehouse, and follows guardrails to run, monitor, and complete tasks automatically.
Read on databricks.comDetails
Databricks11 Nov 2026NewData agents
Genie with Genie Ontology Delivers 1.6× the Accuracy with Less than Half the Tokens
Problem Enterprise users needed AI agents that could answer business questions using scattered analytical assets, not just tables.
Solution Created Genie with an ontology layer that automatically gathers context from notebooks, dashboards, queries, and documents, letting the agent retrieve and reason over this knowledge.
91% pass rate, 1.6× accuracy, <50% token usage
Read on databricks.comDetails
Databricks11 Nov 2026NewData agents
From enterprise context to action: Genie Agents now analyze more data sources to transform your work
Problem Teams needed agents that could understand both structured and unstructured business data and act within enterprise tools.
Solution Databricks expanded Genie Agents to ingest large document collections, use connectors for tools, and provide no‑code, privacy‑preserving interactions that turn natural‑language queries into actions.
Read on databricks.comDetails
Databricks11 Nov 2026NewData agents
Bring your business context to ChatGPT, Claude, Microsoft Copilot Cowork and Grok Bot
Problem AI agents often guess answers because they lack clear business context and data locations.
Solution Created Genie Ontology, a governed context layer with plugins for ChatGPT, Claude, Copilot, and Grok Bot, letting agents use definitions, relationships, and routing to fetch accurate data and citations.
Read on databricks.comDetails
Databricks11 Nov 2026NewData agents
Announcing GA of Genie Ontology
Problem Agents lacked a complete, up‑to‑date understanding of the whole business, and manual ontologies quickly became outdated.
Solution They built Genie Ontology, a shared knowledge layer that agents continuously learn from dashboards, SQL, notebooks and docs, while people govern the definitions.
Read on databricks.comDetails
HubSpot10 Nov 2026NewData agents
From Data Silos to Feature Serving: How We Unified Fragmented Data for Training, Inference, and Reuse
Problem Teams had to locate, extract, and rebuild scattered data signals for each new model, slowing development.
Solution HubSpot created a unified feature store that links to existing SQL tables, CRM properties, and services, syncing online and offline stores for fast, consistent feature access across training and inference.
Read on product.hubspot.comDetails
Databricks9 Nov 2026NewData agents
Turn buying signals into pipeline with Sendoso gifts in Genie One
Problem Teams had to export data and switch tools to send gifts, slowing down outreach.
Solution Created a Databricks Marketplace MCP server that lets Genie One call Sendoso’s gifting API, turning data signals into automated, governed gift actions.
Read on databricks.comDetails
Stripe5 Nov 2026Data agents
Analyze your Stripe data directly with your agent
Problem Customers wanted to ask AI agents to analyze their Stripe data without writing code.
Solution Stripe added APIs that let agents like Claude retrieve and query over 300 structured tables, generate SQL, and return results for natural‑language prompts.
Read on stripe.comDetails
Google Cloud5 Nov 2026NewData agents
Bringing conversational analytics to Gemini with Agentic Data Cloud connectors
Problem Business users ask many data questions faster than analysts can answer, causing delays and guesswork.
Solution Created conversational analytics agents that let Gemini query live, governed data sources via natural‑language prompts, using Data Cloud connectors, Knowledge Catalog discovery, and secure SQL execution.
Read on cloud.google.comDetails
Groww30 Oct 2026NewData agents
Compass: an AI navigation agent for everyone
Problem Teams struggled to find unified data definitions and numbers across many tools, causing delays.
Solution Built Compass, an AI navigation agent that links the data catalog, lineage, and services, letting users ask natural questions and get grounded answers.
Read on tech.groww.inDetails
Ramp22 Oct 2026Data agents
Agentic Credit Analysis
Problem Need to automate credit decisions from complex financial statements.
Solution Created an agentic pipeline that extracts key metrics, uses LLM tools to analyze statements, and makes autonomous credit approvals.
Read on builders.ramp.comDetails
NVIDIA8 Oct 2026Data agents
Building Reliable Data Analytics Agents: Lessons from the KDD Cup
Problem Answering questions across many data sources caused agents to fail due to complex tool usage.
Solution They built a small, clear harness that limited tools, unified all data into a single SQL view, added persistent Python state, and made every step inspectable.
Second place in KDD Cup 2026
Read on developer.nvidia.comDetails
Pinterest6 Oct 2026Data agents
Metrics Board: Building an Agent-ready Metrics Layer
Problem Teams struggled with chaotic, unreliable metrics across a massive data lake, making decisions and experiments hard.
Solution They created Metrics Board, a platform that automates metric creation, enforces quality, and lets LLM agents query metrics in natural language.
Read on medium.comDetails
Doctolib15 Sep 2026Data agents
From middleman to copilot: Doctolib’s data mesh shift, seen from ingestion
Problem Centralized data ingestion caused delays and unclear ownership for many product teams.
Solution Created a data mesh platform that lets each team own and self‑service its data ingestion, automating pipelines and clarifying responsibilities.
Read on medium.comDetails
Instacart3 Sep 2026Data agents
Agentic Machine Learning Modeling at Instacart
Problem ML engineers spent too much time manually choosing models and settings.
Solution They built AI agents that suggest, code, and evaluate model options, letting engineers focus on oversight.
Read on tech.instacart.comDetails
Grab1 Aug 2026Data agents
How AI is transforming analytics at Grab
Problem Analysts spent weeks preparing data and generating insights, slowing business decisions.
Solution Created an AI agent that automates data prep, query writing, analysis, and report drafting, letting users get answers in minutes.
Read on engineering.grab.comDetails
Faire22 Jul 2026Data agents
Autoscience: How we built agent-driven ML exploration at Faire
Problem ML teams could only try a tiny fraction of possible model changes, limiting performance.
Solution Built Autoscience, an agent system that proposes, runs, and evaluates model experiments with persistent memory and steering hooks, enabling many more iterations.
Production model accuracy improved double‑digit percentages
Read on craft.faire.comDetails
Replit8 Apr 2026Data agents
Beyond the App: Using Vibe Coding to Ship Decks, Dashboards, and Launch Assets
Problem Product managers spend hours switching tools to create launch decks, dashboards, and assets, causing inconsistency and delays.
Solution Replit built an AI agent that reads the project code and design system, then generates launch materials like decks and dashboards that match the product’s look and content.
Read on replit.comDetails
Replit13 Mar 2026Data agents
Vibe Coding Enterprise Data Apps with Replit and Databricks
Problem Teams needed to build apps that could safely use sensitive, governed enterprise data.
Solution Created a connector that lets Replit’s coding agent call Databricks APIs, so apps can read and write trusted data without manual credential handling.
Read on replit.comDetails
Replit20 Nov 2025Data agents
Data Connectors
Problem Teams had to store database credentials manually, creating security risks and slowing down data app development.
Solution Replit added native, OAuth‑based connectors for Snowflake, Databricks, and BigQuery, letting the Agent securely discover schemas and build data apps instantly.
Read on replit.comDetails