O n a Tuesday evening in early July , Anushka Ahuja, a dentist in Gurgaon, India, opens the door for the house help she booked ten minutes ago on Pronto, an app that provides instant professional cleaning services like sweeping, mopping and laundry across 11 cities in India. A woman clad in a crisp uniform greets her with a polite smile, then asks: Can she strap a small camera onto her head to record her work?

Ahuja isn’t surprised to see the camera. In fact, she paid an extra nine rupees (about nine cents) so her cleaner would record everything from her elbows to her fingertips. Now she’ll receive a video recording a few hours later, giving her peace of mind that her house was cleaned properly and nothing was stolen or broken in her absence.

Despite disclosures in the consent form she signed, Ahuja initially didn’t realize that there’s another purpose to the camera footage: AI training data. After a cleaning session, the video is stripped of personally identifiable information like personal documents and people’s faces and then sold to American AI companies.

Bangalore-based Pronto launched the video recording service, called Pronto Verified, in May, stirring widespread backlash in India and government scrutiny over privacy concerns. Customers consent to the recording and workers are paid double for Pronto Verified bookings, CEO Anjali Sardana says.

“Trillions of dollars are going into AI capex every year. All the money in the world is getting invested in AI, but your blue-collar worker in India is not capturing any of that,” Sardana, 24, tells Forbes . “And there is an opportunity to actually redistribute that capital.”

Ahuja says she doesn’t mind that footage of the inside of her home will end up in an AI model’s training dataset. “Everyone is just collecting data… I think it's just everywhere, so it's fine,” she says. “It’s beyond our control now.”

Demand for Pronto’s video footage has exploded in the past 12 months as robotics becomes the new hot category in AI, says Ajay Agarwal, a partner at Bain Capital Ventures, which invested $2 million into the startup when it was founded in February 2025. Unlike large language models, which were trained using data scraped from all corners of the internet, robotics startups require data that’s not as readily available, including videos of people interacting with physical spaces. “The thing we're able to offer the AI labs is incredible diversity of environments with 50,000 cleanings a day, a range of humans performing different cleaning tasks and that's a scale that no one else has, frankly,” Agarwal says. Pronto declined to share which AI startups are using its data.

“The thesis behind Pronto is just that broken labor markets should not exist.” Anjali Sardana, CEO of Pronto

The most immediate application for Pronto’s data: building a general-purpose home robot that can do much of what the human house help does today. Pronto’s workers don’t have much to fear about being replaced though, at least in the short term: customers pay from 49 rupees (50 cents) to 249 rupees ($2.60) an hour. It’ll take a very cost-efficient robot to compete with that.

Pronto isn’t the only company that’s stumbled onto a potential data treasure trove. In May, Shift , a startup founded by German company MicroAGI, started offering free housekeeping services in New York City with a catch: its cleaners wear head-mounted cameras to film people’s apartments and collect data for training humanoid robots. “You get a spotless apartment. We get training data. Everyone wins,” reads the company’s website.

Selling training data is a new line of business for Pronto, which primarily sends cleaners to people’s homes across cities like Bangalore, Mumbai and Pune within 10 minutes. Pronto does 46,000 regular bookings and thousands of Pronto Verified orders each day. Valued at $200 million, the nascent startup has raised about $60 million in funding from American investors including Bain Capital Ventures, General Catalyst, Glade Brook Capital and Lachy Groom, CEO of $11 billion-valued robotics startup Physical Intelligence.

The company has onboarded 7,000 workers through referrals, influencer campaigns in villages and local street vendors and tea sellers who are well connected. Workers are assigned to what’s called a “polygon,” a geographic area with a 500 meter-radius where a team lead helps resolve issues, answer questions and take care of day-to-day operations. Pronto trains staff on household chores like cleaning, ironing and food prep as well as how to use phones, plan finances, speak to customers and maintain hygiene and grooming. In areas that are especially unsafe for female workers, which make up 99% of the startup’s workforce, the company also teaches self defense. It also provides electric scooters so that when matched, workers can get to a customer’s home quickly.

“I feel like I got out of jail.” Puja Kumari, a professional at Pronto.

Today, more than 50 million people across India work as domestic help. But in this unorganized, chaotic industry, everyone loses: customers struggle to find reliable, high-quality help while workers grapple with unsafe working conditions, income instability and lack of dignity and respect.

“The thesis behind Pronto is just that broken labor markets should not exist,” Sardana says. “There's an opportunity to build a platform that structures informal, unskilled labor and provides incentives for workers to stay in those structures, essentially bridging the gap between what these service providers can do on their own and what the customers expect.”

Pronto is part of India’s booming quick commerce industry, which has birthed a slate of successful startups promising near-instant services. Blinkit, a service that delivers groceries in 10 to 15 minutes, brought in about $3.9 billion in revenue in FY2026, accounting for about 75% of parent company Eternal Limited’s revenue. Zepto , another rapid grocery delivery app, recorded $2.35 billion in revenue in FY2026 and plans to go public at a valuation of at least $3 billion. In March, home services giant Urban Company launched “Insta Maids,” a service similar to Pronto, sparking uproar over derogatory language it used in its promotional campaign. It later changed its name to “Insta Help.”

T he first time Varsha Singh , 33, wore a camera to a customer’s home, it felt strange and heavy, she says. But then the Go Pro was replaced with a lighter camera. Now she can barely tell it’s there.

“At first, we would be scared of going to a stranger’s house. Someone might falsely accuse us of stealing something or might speak rudely,” Singh says. “But with the camera on we don’t need to worry. We feel safer. Everything is being captured, how they’re speaking and how they’re treating us.” She had no idea what AI is, let alone that the video recordings were being used to train it.

Singh used to earn 18,000 rupees (about $180) a month working about 45 hours a week at a cafe in Uttar Pradesh. Then she joined Pronto last January, and now she earns almost double: About 30,000 rupees ($310), for a similar amount of work. Singh says she’s been paid about two times her regular earnings for the verified bookings.

Other workers Forbes spoke to say Pronto has given them a new identity, recognition and respect at a time when domestic low-skilled laborers are looked down upon and treated poorly. They’ve gotten access to better working conditions, more agency to decide their work hours, location and time off. “We have a zero tolerance policy for anyone misbehaving and we will assist you. There's an SOS button in the app, you press it, someone will be there in minutes if necessary,” Sardana says.

Some have even escaped toxic employers. Puja Kumari, a domestic worker originally from a village from Bihar, says she had been working as a nanny for more than three years for a family that didn’t let her cross the gate of their apartment complex and tracked everywhere she went. In January, a friend told her about Pronto.

“I feel like I got out of jail,” she says. “I used to feel suffocated. I had no freedom to talk, to laugh, to go anywhere.”

Pronto customers are relieved too. Many Indian households have dedicated full-time or part-time house help. But it’s not always reliable: workers will quit without advance notice, leaving employers scrambling to find a replacement and train them from scratch. “The biggest thing for me is convenience,” says Anushka Kotwanni, a Pronto customer who lives in Bangalore. “The help comes when I want them to come for a fixed time during the day.”

S ardana, who grew up in Northern Virginia, says the idea for Pronto came during a trip to India in her senior year of college. She witnessed how quick-commerce grocery apps like Blinkit had taken off in India, even as American instant delivery companies like Gopuff struggled . Speaking to friends and family, she realized that the unpredictability around domestic help was a common pain point. She decided to build a business structuring and formalizing the country’s inefficient labor markets, she says.

She started building Pronto during her first job as a private equity investor at Bain Capital in Boston, working on the startup at night while it was daytime in India. Six months into the job, she quit to focus on Pronto full time.

Sardana’s family knows a great deal about running an operations-heavy business. Her father and uncle cofounded Innova Solutions, a company that made billions in revenue outsourcing IT jobs to engineers in India. Her father sits on the board and retired last year, while her billionaire uncle Raj Sardana runs the company as CEO ( Forbes profiled him in 2023). But when Sardana first told her father about her idea for Pronto a month before she graduated from Georgetown in May 2024, he quickly dismissed it, saying no one would want such a service and that she was incapable of building it because she had never lived in India. “And I was like, ‘Okay, screw you. I'm going to do it anyways, but I don't want any of your support," she says.

Sardana vibecoded the first version of the app in October once she realized that running it on WhatsApp didn’t give it any legitimacy. Gaining trust among workers was also an uphill battle in the early days, when Pronto had 50 bookings and just three workers. Sardana says one of Pronto’s first employees was almost beat up when he tried to hire workers to the platform because they didn’t trust him and had never heard of the company. But when he bought a button-down shirt and a cheap gold chain, the pitch was well received. “He showed up the next day with 30 people” who wanted to work for Pronto, she recalls.

After the startup’s first pilot in Delhi failed, Sardana launched a second one in Gurgaon that quickly gained traction. She spent days going door to door and stopping domestic workers at the gate, asking about their wages and experiences. “We were literally sleeping on park benches in a market... And we learned a lot of hard lessons,” she says.

Sardana sees a much bigger opportunity than organizing domestic work, or even selling AI training data. Pronto could extend to other low-skilled labor markets that struggle with supply shortages, like factories, small retail stores and construction. “I am not interested in building a home services company. That's very, very uninteresting,” she says. “The vision doesn't just stop at India. It's pretty global in terms of going across a lot of developing nations where you have a lot of great people who have been suffering from broken labor markets.”