Sol Foundry Raises $4 Mn To Automate Email Promises

Suraj Prajapati
By
Suraj Prajapati
Suraj Prajapati, Contributing Editor at StartupFeed
Contributing Editor
Suraj Prajapati is a Contributing Editor at StartupFeed covering artificial intelligence, cybersecurity and India's fintech markets. He is a technology founder who built and runs a...
- Contributing Editor
Sol Foundry says its assistant uses more than 100 skills while keeping final action behind user approval.

Sol Foundry, a proactive AI startup built by three former CRED executives, has raised $4 Mn (about Rs 38 Cr). The round takes it out of stealth.

General Catalyst and Nexus Venture Partners led the round. DeVC, Peercheque and CRED founder Kunal Shah also put in money. Inc42 first reported the raise on September 22, 2026.

The startup was founded in 2025 by Anish Karan, Prateek Srivastava and Ranjith Nair. All three worked at CRED. Karan was in product, Nair in design and product, and Srivastava was the engineering lead.

Sol builds one thing: a proactive AI assistant. Most inbox AI waits for a prompt. Sol does not.

The product scans a user’s Gmail for commitments. If someone has written “I’ll send the deck” or “I’ll set up time”, Sol starts the work on its own. It can run research, write documents, build slide decks and draft replies pulled from many threads.

The user keeps final approval. Karan says Sol cannot send a mail, book a meeting or share work without a clear yes. Permissions can be pulled back at any time.

Sol runs in its own computer environment and can browse the web. It draws on a library of more than 100 skills. The stack is model-agnostic and uses OpenAI models today, running on Amazon Web Services in the United States.

The company is US-first and pre-revenue. It has no fixed pricing yet. Karan says pricing will be usage or outcome-based, in US dollars, with no India plan for now. Sol has more than 100 users on a waitlist product.

Before building, Karan spoke to about 55 enterprise teams in the US, South Korea and Australia. He checked whether the problem was only an Indian one. It was not.

What this means for you: If you build inbox or workflow AI, watch the approval step, because an approved draft carries the user’s name and that trust is the real moat.

Have a tip? Write to us at editorial@startupfeed.in.

Liked this story? Follow StartupFeed on Google so our reporting reaches you first.

Follow StartupFeed on Google News
Suraj Prajapati, Contributing Editor at StartupFeed
Contributing Editor
Follow:
Suraj Prajapati is a Contributing Editor at StartupFeed covering artificial intelligence, cybersecurity and India's fintech markets. He is a technology founder who built and runs a platform for unlisted shares, pre-IPO investing and SME IPO advisory, and advises companies on digital innovation and cybersecurity. He works with two organisations focused on cybercrime research and reporting, and has led more than 150 digital transformation projects, earning recognition for security practice in web development.
Newsletter signup illustration: an open envelope with a letter and a paper plane

Don’t Miss Startup News That Matters

Join thousands of readers getting daily startup stories, funding alerts, and industry insights.

Newsletter Form

Free forever. No spam.