Sol, a proactive AI assistant built by three former CRED executives, has emerged from stealth with $4 Mn in funding.

The bet is that the next wave of AI agents will not wait for prompts but act on commitments already made in email, a shift the biggest productivity incumbents are also chasing.

Backed by General Catalyst, Nexus Venture Partners, DeVC, Peercheque and CRED founder Kunal Shah, the assistant is US-first, runs on Google Workspace and is open through a waitlist.

So far, AI in the inbox has so far meant assistance in drafting.

Microsoft has pushed Copilot deeper into Outlook, Google has embedded Gemini across Workspace, to provide AI assistants designed to organise inbox, draft replies in your specific tone, and capture meeting notes directly. Almost all of them still wait for the final go-ahead.

Sol, a proactive AI assistant, wants to flip that order. Founded in 2025 by three former Cred executives, Anish Karan, Prateek Srivastava and Ranjith Nair, the startup emerged from stealth today with $4 Mn (over ₹38 Cr) from General Catalyst, Nexus Venture Partners, DeVC, Peercheque and Kunal Shah.

The offering: if a user has already written “I’ll send the deck” or “I’ll set up time”, an AI automatically begins working without waiting for a prompt.

The product scans a user’s Gmail (Google Mail) for commitments and then produces the work those promises imply, from research and documents to slide decks, calendar coordination and reply drafts assembled from several threads. The user retains final approval.

Sol runs in its own computer environment, can browse the web and draws on a library of more than 100 specialist skills, running on OpenAI’s model under-the-hood. Some of the use cases include:

• Creating or reviewing a document in Word or Notion, or building and reviewing slides


• Finding and shortlisting time slots across internal and external calendars


• Running research ahead of a reply or a decision


• Drafting replies that pull context from scattered email threads

The idea surfaced during the cofounders stint at Cred. While Karan worked in the product of the fintech major, Nair was in the design and product team, and Srivastava was the engineering lead.

After identifying the problem, Karan spoke to around 55 enterprise teams from the US, South Korea and Australia to check whether the problem was local for companies in India. It was not.

Commercially, Sol is pre-revenue with no pricing tiers fixed yet. Karan says the pricing model will be usage or outcome-based with a fixed component to start, since the work varies in volume and complexity per user. Pricing will be in US dollars with no India currency pricing plan yet.

He declined to commit to a gross margin target, arguing that model costs are falling and adoption matters more, for now.

The go-to-market motion is “prosumer” first: anyone can sign up (waitlist at the moment), and an enterprise sales motion follows once enough people in one organisation use it.

Integrations begin with Google Workspace, covering Gmail, Docs and Slides, while Outlook is open to a small group of testers. The stack is model-agnostic, using OpenAI models today with routers that switch between them, and runs on AWS in the US.

Karan argues Sol has no direct competitor, only the workflow it replaces, and that reliability plus a simple interface are the moats a challenger must cross. That claim faces incumbents bundling agents into the inbox for free, and the trust problem every email agent inherits: an unapproved draft is harmless, but an approved one carries the user’s name.

Whether proactive assistants become the default layer of work email, or get absorbed into Gmail and Outlook as features, only time will tell.