Imagine AI teaches a machine to sound exactly like you - then hands it your LinkedIn

There is a fork in the road that every founder who has ever tried to build in public knows by heart. It arrives on a Tuesday afternoon. You can ship the feature the customer asked for, or you can write the LinkedIn post that might, eventually, bring in the next customer. You ship the feature. You always ship the feature. And the post that could have found your next ten buyers never gets written.

Imagine AI, a San Francisco company in Y Combinator's Fall 2025 batch, was built inside that exact moment of hesitation. Its founders, Sky Yang and Neo Lee, are not selling the idea that founders are bad writers. They are selling the observation that founders are experts who have run out of hours. The company's job is to take the expertise that already lives in a person's head - the talks they gave, the calls they took, the way they explain their own product when no marketer is listening - and turn it into a steady stream of content that sounds like them and, more to the point, brings in business.

01 / What it actually doesA voice engine, not a template

Most AI writing tools start with a blank box and a prompt. Imagine AI starts with a dossier. Before it publishes a single post, the platform builds what the company calls a persona report - a document that can run past 100 pages, assembled from a person's transcripts, recorded interviews, past writing, and conversations. The point is to capture not just how someone writes but how they think: their worldview, their recurring arguments, the way they position their company against everyone else.

That report feeds a voice engine that drafts the actual output - roughly four LinkedIn posts a week, strategic comments left on prospects' posts, and longer authority-building blogs tuned for search. A human stays in the loop. Every account gets a dedicated content engineer, a real strategist who manages the plan and keeps the AI from drifting into the flat, over-eager register that gives machine writing away. Nothing publishes without approval.

02 / The metricContent market fit

The phrase the company keeps returning to is "content market fit," a deliberate echo of the product-market-fit gospel that YC founders are raised on. The idea is a small act of rebellion against the way most content gets judged. Impressions are easy to inflate and easy to ignore. Imagine AI's analytics are built to watch a narrower, more uncomfortable number: how many of the right people, the actual buyers, started a conversation because of a post.

The traction claims the company puts forward are framed in that language. It reports helping clients generate somewhere between $10M and $40M in attributed revenue, more than 100 million impressions across accounts, and one striking anecdote it likes to tell: a client who traced a $5M deal back to a single optimized post. Those are self-reported figures, the kind that should be read as a company describing its best days rather than an audited ledger. What is easier to verify is the shape of the growth - after a mid-2025 launch, the company reported week-over-week gains of 30 to 70 percent.

*Client-reported figures. Read as the company's account, not audited results.

03 / Who buys itThe whole leadership team, posting together

The customer list skews toward sales-led B2B: SaaS founders, GTM leaders, VPs of sales and marketing, and content teams that are tired of a single overworked founder carrying the whole company's public voice. Named clients include MongoDB, Rippling, Slash, and Corgi Insurance, alongside sixteen fellow Y Combinator companies. There is a quiet irony in the mix, too - some of the ghostwriting agencies that would normally be competitors have turned into customers.

That points to what may be the company's sharpest structural idea: content as a team sport. Instead of one founder shouting into the feed, Imagine AI coordinates a whole leadership bench - VP Sales, VP Marketing, Head of GTM - posting aligned narratives on a shared cadence through one calendar. A company with a coordinated point of view reads very differently from five executives independently guessing at what to say.

It also changes the internal politics of content, which is usually where these efforts quietly die. When a single founder owns the feed, every post is a personal errand that competes with the actual job. Spread across a team with a shared calendar and a strategist keeping score, posting stops being a favor someone reluctantly does at 11pm and starts looking like a channel the company runs on purpose. For a sales-led business, that reframing may matter more than any single clever post.

04 / The market it's crashingBetween the agency and the chatbot

To understand where Imagine AI fits, picture the two doors a founder currently has. Behind the first is a content agency: real strategists, real quality, and a bill that can start north of $150,000 a year with months of onboarding before the first post lands. Behind the second is a generic AI tool: instant, nearly free, and prone to producing the kind of hollow, everyone-sounds-the-same copy that can quietly damage a brand.

Imagine AI is trying to build a third door. The deep persona report and dedicated content engineer borrow the substance of the agency model; the AI drafting borrows the speed and price of the software model. Whether that blend holds up under scale is the open question, but the positioning is clean: agency-grade voice at software-grade cost.

05 / The foundersTwo dropouts, zero safety net

Both founders left something to be here. Sky Yang walked away from a Master's program at UC San Diego, where he had earlier been the youngest person ever elected student body president, leading a student body of 32,000 and helping raise $150M for campus infrastructure. Along the way he built a nonprofit, Break the Outbreak, into a 400-person organization spanning 53 cities and 18 states. Neo Lee left a founding-engineer seat; he had worked at startups backed by MiraclePlus and Mysten Labs and studied stochastic games at UC Berkeley.

The company's own account of its origins leans on family history - grandparents in rural China without running water, parents who arrived as asylees - and lands on a line the founders repeat like a thesis statement.

Eight weeks in, the company says it had earned its first real revenue and gathered a waitlist of more than 250 companies. It raised a $130,000 seed alongside its YC acceptance, with partner Pete Koomen. The business model underneath all of it is subscription software sold by the seat, wrapped in a service layer - the kind of hybrid that is easy to start and hard to keep profitable, which is precisely the tension a young team gets to spend the next few years resolving.

06 / What you can take from itThe idea is portable

Even if you never buy the product, the core move is worth stealing. Your best raw material already exists. It is sitting in the transcript of a webinar you gave, in a Slack thread where you explained your product better than any brochure, in the voice memo you recorded walking to lunch. The work is not creation. It is extraction. Imagine AI's whole company is a wager that mining what experts already know beats asking them to produce something new on a deadline they do not have.

For a startup this young, the honest framing is that of a promising bet rather than a settled outcome. The claims are large and self-reported, the model is unproven at scale, and the space is crowded with tools promising some version of the same thing. But the underlying observation - that founders are drowning in expertise and starving for time - is real, and Imagine AI has aimed itself squarely at the gap.