Magical.

Here is a reliable way to build a large software company: find a task so boring that nobody has bothered to fix it, fix it for free, and wait. Magical is a fairly pure example of this. The task, in its case, was moving small pieces of data around a web browser - a name here, an email there, the same three sentences you type forty times a day to forty different people. Magical made a Chrome extension that did the typing for you. You give it a shortcut, like @intro, and it expands into the paragraph. You point it at a form, and it fills the form. It costs nothing. Roughly a million people installed it.

That is the setup for the interesting part, which is what you do once a million people have quietly handed you a map of the most tedious minutes of their workday. Magical's answer, refined over five years, is: build AI agents that do the whole workflow, sell those to the companies where all these people work, and keep the extension free as the thing that got you in the door. This is not a subtle strategy - the founders will tell it to you plainly - but it is a good one, because the free product is also the world's largest running study of which office tasks are worth automating.

Magical was started in 2020, during the COVID lockdown, by Harpaul Sambhi and Rosie Chopra - who are married - along with Zach Piepmeyer and Prashant Viswanathan. Sambhi is not a first-timer. His previous company, Careerify, was an HR-tech startup he bootstrapped to profitability and then sold to LinkedIn in 2015; he spent the following years in product roles at LinkedIn and Microsoft. Chopra came from the finance-and-strategy side, with stops at KPMG, GE, Deloitte and Atlassian. The founding story is the kind that sounds too tidy to be true but apparently is: stuck at home, they started cataloguing the repetitive digital chores that ate their days, and decided to automate them.

Automation without the integration tax

The technically clever thing about Magical is what it refuses to do. Traditional robotic process automation - the UiPath and Automation Anywhere world - typically wants integrations, APIs, and a rollout measured in quarters. Zapier, the friendlier version, still needs two apps that agree to talk to each other. Magical operates on the screen. It works with whatever is in front of you in the browser, which means it does not care whether the two systems you are shuttling data between have ever heard of each other. Magical's pitch is that it can automate a complex workflow in about a week rather than six months, and the reason is mostly that it skips the part where IT has to get involved.

This is also the source of the company's slightly awkward identity. Sambhi has described Magical as an RPA platform. Skeptics, including some early reviewers, pointed out that a lot of what it did was, functionally, a very good text expander. Both things are true, and the gap between them is exactly the distance the company has spent the last few years trying to close - from expanding snippets, to drafting AI replies, to running autonomous multi-step agents.

The money, and the operators behind it

Magical has raised about $41 million. The seed round, roughly $6 million, came from Greylock and Bain Capital Ventures with Lightspeed and SV Angel. The 2022 Series A - $35 million, led by Coatue - is the headline number. What is more telling than the sum is the angel list: founders and executives from Figma, Intercom, Lattice, Front, Atlassian and Morning Brew. When the people who built the productivity tools you already use put their own money into a productivity startup, it is at least a signal that the problem is real to people who would know.

The pivot nobody has to apologize for

The most recent chapter is a move toward healthcare. In 2025 Magical launched an enterprise agentic-AI suite - autonomous agents, process mining, human-assisted automations - and leaned hard into healthcare operations: prior authorizations, claims, credentialing, revenue-cycle management. It is not a glamorous market. It is a market absolutely stuffed with exactly the repetitive, high-stakes, screen-bound busywork Magical was built to eat. Pointing a copy-paste killer at an industry that runs on copy-paste is less a pivot than a homecoming.

Whether Magical becomes the next big automation platform or a well-liked feature that a larger company eventually absorbs is genuinely unsettled - the total funding is modest by 2026 standards, and the competitive field, from Glean to Zapier to the RPA incumbents, is crowded. But the underlying bet has been consistent since a locked-down apartment in 2020: most work software adds things to your screen, and the more valuable move is to quietly take things off it.