“Seek AI turns plain-English questions into database answers, helping business users explore enterprise data without writing SQL.”
Ever wish you could just ask your spreadsheet a question instead of digging through pivot tables? That’s the exact problem Seek AI was built to solve — and then IBM decided the idea was good enough to buy the whole company. Here’s what Seek AI actually does, why it got folded into IBM’s Watsonx AI Labs, and what that means if you’re considering it today.
What is Seek AI?
Seek AI is an agentic AI platform built for enterprise data analytics. The core idea is simple: instead of writing SQL or waiting on a data analyst, anyone in a company can type a plain-English question and get back an answer, a chart, or a written summary pulled directly from the company’s own data warehouse. It was founded in 2021 by data scientist Sarah Nagy and built its reputation as a “text-to-SQL” tool aimed at closing the gap between raw enterprise data and the business people who actually need answers from it.
Before its acquisition, Seek AI connected to major data warehouses including Snowflake, Google BigQuery, Amazon Redshift, Databricks, and Microsoft SQL Server, and had backing from investors like Battery Ventures, Conviction Partners, and NJP Ventures.
How does Seek AI turn questions into data answers?
Seek AI runs on a multi-agent system rather than a single model doing everything. Separate agents handle different parts of the process: one interprets the question (Dialogue), one maps that question to the company’s actual database structure and writes the query (Semantic Parsing), one explains the results in plain language (Explanation), and one suggests follow-up questions worth asking (Exploration). The platform’s proprietary model, called SEEKER-1, is reported to translate natural language into SQL with over 90% accuracy, and it uses reinforcement learning to keep improving as people use it, since every company’s data schema and terminology are a little different.
Seek’s marketing leans heavily on the idea of a “semantic model” — essentially teaching the AI what a company’s specific data actually means (what counts as “active users,” which table is the real source of truth, etc.) rather than just guessing from column names. That’s the part meant to separate it from a generic chatbot bolted onto a database.
Why did IBM acquire Seek AI?
On June 2, 2025, IBM announced it had acquired Seek AI for an undisclosed amount, alongside the launch of Watsonx AI Labs, a new AI accelerator based at IBM’s One Madison offices in Manhattan. IBM framed the deal as a way to make its Watsonx platform easier for non-technical business users to query, particularly in regulated industries like finance, healthcare, and government where explainable, natural-language access to data is valuable. Seek AI’s team, including founder Sarah Nagy, moved over to IBM as part of the deal, and the startup relocated its operations to the Watsonx AI Labs location.
Watsonx AI Labs itself functions partly as a startup accelerator, connecting outside AI developers and companies with IBM engineers, mentorship, and potential funding through IBM Ventures.
Is Seek AI still available as a standalone product?
Not in the way it was before the acquisition. Seek AI’s technology is now being folded into IBM’s Watsonx ecosystem rather than sold as an independent startup product. If you’re evaluating it today, treat “Seek AI” as IBM technology going forward rather than a separate company you’d sign a contract with directly — anyone interested should check IBM’s Watsonx pages for the current state of the integration, since post-acquisition roadmaps like this tend to change.
Because Seek AI was absorbed into IBM rather than continuing as a standalone vendor, pricing, sign-up flow, and even the product name could have shifted since the acquisition closed. Confirm current availability directly on IBM’s official Watsonx pages before making a purchasing decision.
How does Seek AI compare to other natural-language BI tools?
Seek AI sits in the same general category as tools like ThoughtSpot, Looker, and Veezoo — platforms trying to let business users query data without SQL. Reviewers who used Seek AI before the acquisition pointed to its ability to remember context across follow-up questions and generate complex SQL on the fly as a strength compared to traditional dashboard-first BI tools like Tableau or Power BI, which are built more around pre-made visualizations than open-ended conversation with your data.
| Detail | Seek AI |
|---|---|
| Founded | 2021, New York City |
| Founder | Sarah Nagy |
| Core function | Natural language to SQL / data querying |
| Underlying model | SEEKER-1 (proprietary, multi-agent) |
| Data warehouse support | Snowflake, BigQuery, Redshift, Databricks, SQL Server |
| Current status | Acquired by IBM, June 2025 — now part of Watsonx AI Labs |
Would a tool like Seek AI be useful for a game studio?
Natural-language data tools like Seek AI are aimed at enterprise teams generally, not games specifically, but the use case translates well. Live-service and mobile game studios sit on huge amounts of structured data — daily active users, retention cohorts, in-app purchase behavior, live-ops event performance — and a lot of that analysis still runs through a small data team that everyone has to wait on. A natural-language layer over that kind of warehouse is exactly the sort of tool a UA or live-ops team could use to self-serve basic questions without pulling an analyst off other work, similar in spirit to how workflow-automation platforms have been adopted by marketing and growth teams elsewhere.
What stood out
- Multi-agent design split across interpretation, querying, explanation, and follow-up suggestions
- Reported 90%+ SQL accuracy from its SEEKER-1 model
- Recognized by analyst firms including Gartner and Forrester before acquisition
- Broad data warehouse compatibility (Snowflake, BigQuery, Redshift, Databricks)
Worth knowing
- No longer operates as an independent standalone company
- Enterprise-focused, not built with small indie teams or hobby budgets in mind
- Future roadmap and pricing now depend entirely on IBM’s Watsonx strategy
- Gumloop AI automation: pricing and features — another AI platform aimed at automating business workflows
- What is Vapi? — a look at another enterprise-facing AI platform
- What is Leonardo.Ai? — a different kind of AI platform, built for creative generation
- Best password managers for teams in 2026 — more tooling picks for growing teams
- Roblox open-sources its AI safety tools — how another major platform is putting AI to work
Frequently Asked Questions
Can I still sign up for Seek AI directly?
Since IBM’s June 2025 acquisition, Seek AI is being integrated into IBM’s Watsonx ecosystem rather than sold as a standalone product. Check IBM’s official Watsonx pages for the current way to access this technology.
Who founded Seek AI?
Seek AI was founded in 2021 by data scientist Sarah Nagy, who joined IBM as part of the acquisition.
What is Watsonx AI Labs?
Watsonx AI Labs is an AI accelerator IBM launched in June 2025 at its One Madison offices in Manhattan, built to bring together IBM engineers, outside developers, and startups working on agentic AI, with Seek AI’s technology as a foundational piece.
What databases did Seek AI connect to?
Before the acquisition, Seek AI supported major data warehouses including Snowflake, Google BigQuery, Amazon Redshift, Databricks, and Microsoft SQL Server.
How much did IBM pay for Seek AI?
IBM has not disclosed the purchase price. Seek AI had previously raised roughly $10–11.7 million from investors including Battery Ventures, Conviction Partners, and NJP Ventures.