Minna Song

EliseAI, the New York software company Minna Song runs, sells a boring product to a boring industry: it answers rental inquiries. Someone emails an apartment building, wants to know if there is a two-bedroom under $4,200 with a dishwasher, and Elise replies within seconds, in a voice indistinguishable from a leasing agent, and if the prospect is qualified, books the tour. This is not glamorous. Andreessen Horowitz recently valued it at $2.2 billion.

Song is the cofounder and chief executive. She is in her late twenties, holds a computer-science degree from MIT, and started the company in 2017, which is a slightly odd year to have started a conversational-AI company because ChatGPT would not exist for another five. The polite way to describe her timing is early. The impolite way is that she spent several years running a chatbot business at a moment when chatbots were something a normal person might reasonably describe as "annoying," "broken," or "please transfer me to a human."

Before she wrote any code for EliseAI, Song did something you would not necessarily expect from an MIT computer science graduate with internships at Microsoft and MIT Lincoln Laboratory on the resume. She got a job at the front desk of a real estate firm. Not to make rent. To watch. She wanted to see, without an intermediary, which inquiries went unanswered, which calls went to voicemail, which prospects filled out an online form on a Saturday night and then heard nothing until Tuesday afternoon and then heard nothing again. This is the sort of thing you can guess at in a slide deck. You do not really know it until you are the one deciding which email to reply to next.

What she found is what everyone finds, which is that leasing offices are staffed by human beings, and human beings work business hours, and inquiries do not. The Sapphire Ventures announcement note when EliseAI hit unicorn status included a data point that is not really a data point but a diagnosis: nearly half of all rental inquiries go unanswered, and the ones that get answered often take more than 24 hours. If you sell apartments for a living, half of your top-of-funnel is a rounding error to your own inbox. This is what Song built EliseAI to fix.

The company was originally called MeetElise, which is charming and slightly cutesy and suggests a company that answers your emails. In 2022 it became EliseAI, which is less charming and suggests a company that intends to answer everything else too. This turned out to be a fair rebrand. EliseAI now handles the leasing conversation, the renewal conversation, the maintenance request conversation, the payment reminder conversation, and, since 2023, the healthcare appointment conversation, through a sibling product called HealthAI. Song runs all of it.

The cofounder is Tony Stoyanov, sometimes called Stoyan Stoyanov, who is EliseAI's chief technology officer and who Song met studying computer science in college. In an interview with Thesis Driven, Song described the founding decision in terms that sound almost unfashionably deliberate: they decided they wanted to start a company, and then they thought, from first principles, about which company would maximize their impact. Most founders arrive at their idea by accident or grievance. Song and Stoyanov arrived at housing communication by reasoning.

What EliseAI actually sells, when you strip out the labels, is a way for a property manager to stop losing money on prospects. A landlord who owns 20,000 apartments cannot manually track which inbound inquiry got a same-day response and which got ignored for four days. EliseAI does the tracking, replies in the interim, and hands back a scheduled tour. In the Bessemer writeup of the company, the framing is "vertical automation," which is venture-capital speak for a piece of software that lives inside a specific industry's workflow rather than trying to be everything to everyone. EliseAI knows what a "concession" is. It knows what "PMS integration" means. It knows that the person emailing about the studio on East 82nd is either going to sign a lease in three weeks or vanish, and the difference is how fast the reply comes.

The Forbes cover story in June 2026 put the customer count at more than 350, including 70% of the fifty largest rental housing operators in the United States. The Series E announcement raised that share to roughly 75%. This is the sort of penetration that, in most industries, takes decades. EliseAI has been selling for about eight years.

There is a version of the AI moment we are in where every founder is racing to demonstrate the newest model, the flashiest agent, the most impressive multi-step reasoning trick. Song is not really in that race. She is in the adjacent, less-decorated race, which is the one where you build a durable business by picking an ugly workflow, embedding yourself in it, and never leaving. When she talks about EliseAI in interviews, the words are not "foundation model" and "reasoning" but "renewals," "maintenance," "compliance," "dynamic pricing." The verbs are "collect," "schedule," "follow up." A close read of what she is describing is not artificial intelligence in the science-fiction sense. It is an operations team that never sleeps.

The healthcare expansion is the tell. In 2023 EliseAI launched HealthAI, aimed at practices in dermatology, ophthalmology, ob-gyn, orthopedics, and the various outpatient clinics that share housing's particular problem: an operations layer buried under phone calls, scheduling requests, follow-ups, and reminders. A patient wants to reschedule; a resident wants a maintenance visit; both are, at bottom, an inbound message that needs a fast, correct, polite reply. Song's insight is that the underlying workflow, once you name it precisely, is portable. Different industry, same category of pain.

The a16z-led $250 million Series E closed in August 2025 and pushed total funding to about $421 million. Song used part of the round to double headcount past 450. She did not, by all reporting, use it to buy a superyacht or a Superbowl commercial. The interviews she has given since (Thesis Driven, the Funded podcast with Jason Yeh, an appearance on the a16z podcast about AI, housing, and healthcare affordability) all sound roughly the same: measured, technical, patient. She is not selling anything. She is describing the workflow.

The CNBC Changemakers list named her in 2026. Forbes ran a cover story in June of the same year, headlined around the $2.2 billion figure and the fact that the founders are both 29. That last data point is the one Song seems least interested in. In the Thesis Driven interview she was asked, gently, what it was like to be a young founder selling to landlords, an industry not famous for treating young founders as equals. Her answer was essentially: we already got them to sign contracts, so.