Fundraise

Magnetic raises $5.2M to automate the grind of tax season

What's the deal? Magnetic, a San Francisco startup automating tax preparation for CPA firms, has raised $5.2 million in seed funding. Goodwater Capital and Griffin Gaming PartnersDealroom has a profile for this one. Try Dealroom → backed the round.

What Magnetic builds: Its software ingests unstructured tax documents — from handwritten receipts to multi-state K-1 forms — and feeds the parsed data into existing professional tax engines. The company says its proprietary vision model nearly doubles the accuracy of traditional OCR scanning tools.

Magnetic pairs a multi-agent AI architecture with human review. An internal team of US-based enrolled agents and CPAs checks return data before delivery, and firms do not pay for any return that needs manual correction due to a system error. To date, the platform automates more than 25% of the full tax return process.

Why now? Accounting firms face a persistent talent shortage as older practitioners retire and fewer graduates enter the field. During tax season, staff are pushed into 80-hour weeks to hit the April 15 deadline — and Magnetic pitches a way to handle more returns without new hires.

What's the endgame? Co-founder Thomas Shelley framed the goal bluntly: "our thesis is that we have to get the returns all the way from just a jumble of documents to review ready." The aim is to deliver completed returns ready for partner sign-off, rather than the partial scans that leave accountants validating output by hand.

Magnetic launched publicly in 2025 after Y Combinator's S25 batch and has grown to hundreds of thousands of dollars in annual recurring revenue across dozens of client firms. Founded in 2024, it is led by Shelley, former head of product at consumer tax platform Keeper, and CTO Patrick Fay. The two previously worked together at Zynga.

The signal: Backing for Magnetic shows investors are betting that AI can absorb routine professional-services work facing chronic labour gaps. By keeping humans in the loop and tying payment to accuracy, the company is testing whether firms will trust automation for a task where errors carry real cost.

Read more: zerogtalent.com

Image credit: Generated with Gemini

More top stories