Datost Introduces the First AI Data Analyst With Its Own Computer – Inside Slack

Datost, a Y Combinator Spring 2026 company, is out to democratize data work inside the enterprise with a deceptively simple premise: give every team an AI data analyst that has its own computer and lives where the conversation already happens. The startup’s agent, accessible directly in Slack, sees and understands a company’s documents, databases, data lakes, Slack history, and codebase, then queries, debugs, and analyzes data in the same thread where the question was asked.

The problem Datost is solving will be familiar to anyone who has watched a product manager wait a week for a “quick” number from the data team. Business stakeholders have questions. Data is scattered across warehouses, BI tools, internal wikis, and ad-hoc spreadsheets. A small number of analysts hold the institutional knowledge about how tables join and which fields to trust, and their queue is always full. Datost’s bet is that large language models are finally capable enough – when paired with the right execution environment and the right organizational context – to take over the majority of those one-off analysis requests and hand time back to both the people asking the questions and the analysts who would otherwise be answering them.

What sets Datost apart from the growing roster of “chat with your data” tools is that it comes with its own computer. The agent does not merely translate English into SQL and pray that the query compiles; it operates a sandboxed environment in which it can explore schemas, write and run queries, iterate when results look off, open relevant documentation, read the codebase of the data pipelines that produced the data in the first place, and then explain its findings in plain language. The result is a workflow that looks far more like handing a question to a junior analyst than pressing a button on a dashboard.

“We built Datost to be the first data analyst that actually does the work, not just the talking,” said Maceo Cardinale Kwik, Co-Founder and CEO of Datost. “If you can ask your teammate a question in Slack, you can ask Datost. It will read the docs, check the database, write the code, and come back with an answer – and a trail you can audit.”

Kwik, 21 and based in New York City, co-founded the company with Jason Wang, whose work with the YC Spring 2026 batch has focused on building the agent’s tool-use and context-engineering stack.

The product lands in a fast-moving market. Enterprises have spent the last decade investing in data warehouses and BI platforms only to discover that the bottleneck has moved to the humans who write queries and interpret results. By embedding directly into Slack and natively understanding a company’s existing systems – docs, databases, data lakes, codebase – Datost aims to convert that unanswered demand for analysis into a conversation that teams already know how to have.