We reverse engineer B2B growth, starting with LinkedIn.
Imagine AI is a research and software company reverse engineering modern B2B growth, starting with LinkedIn.
As companies get more automated, more of B2B growth will happen between agents, and the companies that connect will compound each other's output. We're building the connective layer between them. We start with content because it's still the fastest way to build B2B relationships at scale.
Today our content agent decides what your team should post based on data from 1M+ people (who posts what, who engages, where they work and how they're connected). It drafts the posts and graphics, schedules them for your execs and company page, and maps engagement back to your pipeline.
Rippling, MarqVision, Slash, Corgi and Fondo use Imagine AI. Customers have generated $40M in revenue, 100M impressions and 14,341 MQLs.
Free public benchmark of how 1,000+ B2B companies and their execs perform on LinkedIn: https://benchmark.useimagine.ai
Sky Yang is the Co-founder & CEO of Imagine AI.
He was the youngest-ever elected Student Body President at UC San Diego, where he led 32,000 students and successfully raised $150M for university infrastructure improvement. He also founded the nonprofit Break the Outbreak, scaling it to a 400-person organization operating in 53 cities/18 States.
Sky Yang is the Co-founder & CEO of Imagine AI.
He was the youngest-ever elected Student Body President at UC San Diego, where he led 32,000 students and successfully raised $150M for university infrastructure improvement. He also founded the nonprofit Break the Outbreak, scaling it to a 400-person organization operating in 53 cities/18 States.
Neo Lee co-founded Imagine AI in 2025, which now serves 50+ B2B companies including MongoDB, Rippling, Slash, and Corgi Insurance.
He was previously a founding engineer at startups backed by MiraclePlus, Mysten Labs, and the SUI Foundation, and researched non-cooperative stochastic games at UC Berkeley under Prof. Alan Hammond, where he proved a previously open equilibrium-threshold result in stake-governed tug-of-war games.
Neo Lee co-founded Imagine AI in 2025, which now serves 50+ B2B companies including MongoDB, Rippling, Slash, and Corgi Insurance.
He was previously a founding engineer at startups backed by MiraclePlus, Mysten Labs, and the SUI Foundation, and researched non-cooperative stochastic games at UC Berkeley under Prof. Alan Hammond, where he proved a previously open equilibrium-threshold result in stake-governed tug-of-war games.
Company Launches
Imagine AI - You never need to go on LinkedIn again
Hey YC! We're Sky and Neo from Imagine AI (F25). Today we're launching the Imagine AI agent V2. Imagine is your teammate with its own computer and browser that runs LinkedIn for your founders, your team and your company page. It can take care of your entire company’s LinkedIn from one agent, it can post, graphics, comments, connection requests and messages, all from one chat.
It writes your posts and makes on-brand graphics based on what's working in your space
It tells you who engaged, which of them matter (buyers, talent, partners) and reaches out once you approve
Ask: YC founders get 50% off the first month and the setup fee waived. Email sky@imagineai.me with "YC Deal - [company]", or book a demo at https://useimagine.ai and we'll make sure your company has Imagine taking care of your LinkedIn.
The problem
LinkedIn is the bridge between companies. It's high value, but doing it well takes a lot of time and effort.
Nobody knows what works: Founders and team should be posting, but most Series A and B companies have a few growth/marketers at best, so people post whenever they remember
Underused: Most B2B companies communicate on LinkedIn, but it’s always something on the backlog. There is no clear and easy way to do LinkedIn for the entire company.
We spent almost two years on this problem of how B2B growth actually works and how companies actually communicate (especially in the future as we move to a closed loop production web), and we are starting with LinkedIn.
How it works
It knows what works. It's built on our research on how content distributes, like our benchmark at benchmark.useimagine.ai with millions of B2B engagement data points. Our study of 20,000+ YC founder posts found engagement follows a log normal distribution, meaning consistent posting of optimal content will help accounts get more value out of LinkedIn: Companies with 4+ people posting consistently get about 7x the reach per follower
It learns your company. Website, deck, past posts, brand kit and every teammate's voice, in one shared memory
It does the work from its own computer. Posts, graphics, tags, comments, connection requests and messages on demand
It plugs into your stack. Give it an MCP or API key, like your CRM, and it maps LinkedIn engagement to your deals and acts on it
Why us
We research how B2B companies actually communicate on LinkedIn. Our YC founder study and our benchmark of 30,000+ people across 1,178+ B2B companies (benchmark.useimagine.ai) go straight into the agent.
100+ LinkedIn accounts run on Imagine, including teams at Rippling, Fondo, MarqVision, Frame Data & AI, Eragon and Paramify, and overall accounts connected on Imagine have published 4,100+ posts this year with almost 18M impressions. At Frame Data & AI, 70 of 85 matchable CRM records had a direct connection to the content.
We grow with you. We’re a small team and we care a lot about every customer. MarqVision went from 1 to 7 seats, and companies like Rippling and Frame Data & AI each run 4 to 7 profiles with us.