Imagine AI launches a LinkedIn agent used by Rippling and Fondo

Imagine AI launched the V2 of its LinkedIn agent on Y Combinator Fall 2025. The agent runs the LinkedIn work of founders, teams and company pages, Imagine says on its launch page. It writes posts, makes on-brand graphics, posts comments, sends connection requests, and messages people, all from one chat. The team talks to it in Slack or in the Imagine app.

Imagine says 100 accounts are already using the product. Those accounts have published 4,100 posts this year with "almost 18 million impressions", the launch page says. Named customers include Rippling, Fondo, MarqVision, Frame Data and AI, Eragon, and Paramify. MarqVision went from 1 to 7 seats. Rippling and Frame Data and AI each run 4 to 7 profiles on Imagine, according to the launch. The founders are Sky and Neo.

What the agent does on a day

The agent has its own computer and its own browser, the launch page says. On that computer it writes posts, draws graphics, replies to comments on behalf of a user, and sends connection requests. The agent reads a company's website, deck, past posts and brand kit. It then stores every teammate's writing voice. The company calls that memory "Fondo". It also plugs into a user's stack through MCP or an API key. One example in the launch: the agent reads the CRM, maps LinkedIn engagement to deals, and acts on new matches once approved.

The data behind the pitch

Imagine's claim is that it already knows what works on LinkedIn. The company runs a public engagement benchmark at benchmark.useimagine.ai, built from 30,000 people across more than 1,178 B2B companies. Imagine also studied 20,000 YC founder posts. In that study, the company says engagement follows a log-normal distribution. Companies with four or more people posting consistently get about seven times the reach per follower, Imagine says. Every one of these numbers is Imagine's own.

The one named customer figure in the launch comes from Frame Data and AI. Imagine says 70 of 85 matchable CRM records at that company had a direct connection to content from Imagine. That is a measurable claim and a specific one. It is also still Imagine's number, and the writer of this item has not seen the CRM.

Why it matters more than another writing tool

A LinkedIn post generator is a crowded market. The gain in V2, as Imagine presents it, is attribution. An engaged person is tagged as a buyer, a talent target, or a partner. The agent then reaches out once the account owner approves. If the agent is right about who is a buyer, the account team saves time. The LinkedIn work then feeds the sales pipeline. If the agent is wrong, the account team wastes connections.

The reader of this item is building something that needs sign-up or sales. A LinkedIn agent is one more place where an AI makes a decision about a person. The agent should not act without approval, and the launch is clear that it does not. Watch two numbers: match rate to CRM, and the fraction of replies that come from the people the agent flagged. Those are the ones a growth lead will ask about.

Source

• Primary source: Imagine AI on Y Combinator.
• Product site: Imagine AI.
• Benchmark: benchmark.useimagine.ai.