Sourcetable Raises $4.3M to Launch AI-Powered Autonomous Spreadsheet Platform

Sourcetable, a startup developing AI-native productivity tools, has announced the launch of its autonomous spreadsheet platform, marketed as the world’s first “self-driving spreadsheet.” The product is designed to automate spreadsheet analysis and operations using artificial intelligence. It offers natural language and voice-powered interfaces, eliminating the need for advanced formula knowledge and manual data manipulation. To support the platform’s growth, Sourcetable has secured $4.3 million in seed funding.

The funding round was led by Bee Partners, with participation from Julien Chaumond (Co-founder of Hugging Face), Preston-Werner Ventures (founded by GitHub Co-founder Tom Preston-Werner), Roger Bamford (MongoDB Distinguished Architect), and James Beshara (Co-founder of Magic Mind).

Sourcetable targets a common productivity gap: while over 750 million people use spreadsheets daily, only a minority are proficient in advanced functions like VLOOKUP, XLOOKUP, or pivot tables. The platform allows users to issue commands via text or voice, with Sourcetable’s AI completing tasks such as data cleaning, formula generation, chart building, template creation, and multi-tab analysis, no coding or spreadsheet expertise required.

The system supports over 100 live data integrations, connecting to tools like Shopify, HubSpot, Google Ads, Stripe, MySQL, and PostgreSQL without writing code. It also accepts file formats such as .csv, .xls, .xlsx, TSV, and plain text. For larger operations, Sourcetable processes datasets up to 10GB and supports advanced users with SQL and Python environments.

Key features of the product include:

• AI Formula Assist: Automatically generates formulas including SUM, INDEX, XNPV, and more.
• AI Charts & Graphs: Creates interactive visualizations from simple prompts.
• AI Data Cleaning: Cleans and structures messy data with one click.
• AI Research: Enriches spreadsheets by retrieving relevant data from the web.
• AI Summaries: Condenses large datasets and survey results into executive summaries.
• Voice-Driven Commands: Enables hands-free spreadsheet operations.
• Template Generator: Automates the creation of commonly used spreadsheet structures.

Sourcetable is currently trusted by users at top academic institutions, including Stanford, Oxford, Yale, Harvard, UC Berkeley, and the Indian Institute of Technology. The platform is free for students and educators, supported by credits from OpenAI and Anthropic, with a 50% discount applied once the credits are exhausted.

According to CEO and Co-founder Eoin McMillan, “Sourcetable is building the AI spreadsheet for the next billion users, be they human or AI. As analysis becomes easier, everyone can become an analyst.”

The company's underlying system includes a real-time code-based evaluation loop that verifies AI output for accuracy and reliability, critical for supporting the product’s autonomous functionality. The platform’s model architecture is also flexible, dynamically choosing between LLMs from OpenAI, Anthropic, Groq, Meta (Llama), NVIDIA, Hugging Face, and others depending on the task.

Originally designed for data professionals, Sourcetable shifted toward a more inclusive interface after observing increased engagement from non-technical users once friction was removed. Early users have reported dramatic time savings and improved insight generation. For example, a research assistant at Waterloo University described the tool as especially useful for visualizations and forecasting, while a training specialist at Allan Meyers significantly reduced spreadsheet time using Sourcetable’s AI assistant.

"For decades, we’ve lived in a world divided between those who know Excel and those who don’t," noted Michael Berolzheimer, Managing Partner at Bee Partners. “Sourcetable changes that. It’s not just making spreadsheets easier, it’s making data analysis accessible to all.”

Looking ahead, the company plans to evolve Sourcetable into a broader operating system for data interaction and AI agent collaboration. Future capabilities will include read/write access to third-party systems and deeper agent-to-agent interaction support.