AI is fueling renewed interest in consumer tech, including online dating.
Sitch, a matchmaking app that uses AI to connect singles, raised a $2 million pre-seed investment, the startup revealed exclusively to Business Insider. The round was led by Andreessen Horowitz’s startup accelerator, A16z Speedrun, and includes the angel round Sitch raised in 2024 from investors like Jeremy Liew, who wrote Snapchat’s first check.
Fresh out of Speedrun’s most recent cohort, Sitch cofounders Nandini Mullaji and Chad DePue are hiring full-time engineering and growth staff, planning to expand to new cities, and introducing voice-based AI.
While many dating app users feel burned out by constant swiping and rampant ghosting, and dating industry giants like Tinder and Bumble face headwinds, new startups like Sitch are trying to shake up the dating app experience.
“We understand people have been burned in the past,” Mullaji told BI. “We are coming in and saying, ‘Hey, we have a business model shift and we have a total platform shift.'”
The platform shift? AI.
Sitch’s AI matchmaker chatbot — built using OpenAI — is trained on the hundreds of real-life introductions Mullaji has made as a part-time matchmaker.
Mullaji sees AI as a way to “democratize” the matchmaking experience (which can cost individuals thousands of dollars) and bring it “to the masses.”
When signing up for Sitch, users answer a slew of questions about their dating priorities, values, and backgrounds, which Sitch uses to create a profile and curate potential matches. It presents users with a maximum of five potential “setups” each week. Users have to pay up front to access the setup features. Sitch offers three tiers of packages: $90 (for three setups), $125 (for five), and $160 (for eight).
Once Sitch’s AI matchmaker presents users with someone it deems may be a good fit, users can ask the chatbot questions about the other person, and the AI responds using information from their respective profiles. If both parties are interested in meeting, the AI introduces the two in a group chat. However, if an introduction occurs and you get ghosted, Mullaji said you’ll be refunded.
Sitch manually reviews new user applications, which Mullaji said is the “one part of our process that’s still completely human-driven,” as a quality control and safety measure while the platform grows.
AI in dating is still nascent. Other early-stage startups, like Gigi or Amori, use AI to coach singles and help curate matches. Meanwhile, larger dating apps like Tinder and Grindr have introduced AI wingman features.
Can AI make dating feel more … human?
“This is not about building AI girlfriends or trying to replace human contact and connection with AI,” Mullaji said. Instead, she thinks AI can be used to help better connect people.
To give Sitch users a more “human-like experience,” it is introducing voice-based AI features, Mullaji said.
Starting this week, Sitch is rolling out a voice-based AI onboarding experience for new users as the app plans to expand into more US cities. Mullaji said San Francisco and Los Angeles will be added in May, and Chicago and Washington, D.C. will quickly follow. Users will otherwise be added to Sitch’s waitlist, and if a particular US city reaches “critical mass,” which Mullaji defined as between 2,000 and 4,000, Sitch will begin to admit users.
It will also soon expand its voice AI tools to current app users. Instead of texting the AI matchmaking agent on the Sitch app, users will be able to talk to it on the phone with feedback about setups.
Voice AI tech has become a hot category among venture capitalists. In 2024, voice AI startups raised over $398 million from VCs, according to PitchBook data.
Sitch is using ElevenLabs, a voice cloning AI startup that announced a $180 million Series C round with a $3.3 billion valuation in January, to clone Mullaji’s voice.
“We spent a lot of time recording and re-recording my voice to see how we could actually have it sound human,” Mullaji said. “The one thing you do not want this to feel like is a customer support bot.”
Read the full article here
