Pronto Founder Anjali Sardana Launches Provision Labs to Build Physical AI Data Infrastructure

Anjali Sardana Expands Into Physical AI With New Venture

Anjali Sardana, founder and CEO of Indian household services startup Pronto, has launched Provision Labs, a San Francisco-based AI venture focused on building data infrastructure for AI systems that perform physical tasks1

1

2

2

. Sardana will continue leading Pronto while simultaneously serving as CEO of Provision Labs, positioning herself at the intersection of household services and physical AI development2

2

.

Backing and Strategic Focus

Provision Labs has secured backing from prominent investors including General Catalyst, Bain Capital Ventures, Lachy Groom, and Glade Brook Capital, according to the company's website1

1

. The startup has not disclosed its funding amount, valuation, or initial customer base1

1

. The venture will collect data to train and test AI systems designed to perform real-world tasks using robots and other machines, working directly with developers building AI models for robotics and other applications2

2

3

3

.

The Pronto Connection and Unanswered Questions

The launch of Provision Labs comes as Pronto develops a parallel initiative involving data collection through household service sessions. Pronto has introduced Pronto Verified, an opt-in programme that records household service sessions using wearable cameras, with the stated ability to anonymize recordings for training AI systems for robotics1

1

. However, the relationship between Provision Labs and Pronto's data collection efforts remains unclear. Neither company has publicly confirmed whether Provision Labs will access Pronto's recordings or if a data-sharing agreement exists1

1

. Provision Labs' website describes its focus as "the data, evaluation and deployment layer for physical AI," but the startup has not announced commercial partnerships or explained how it will source data across industries1

1

.

Scrutiny Over Pronto's Recording Pilot

The launch follows months of scrutiny over Pronto's pilot programme involving recordings of household chores to generate training data for physical AI and robotics systems3

3

. During the pilot, service professionals wore cameras while performing tasks such as washing dishes, folding laundry, and meal preparation3

3

. Pronto maintained the pilot was strictly opt-in and covered less than 0.01% of its users, with participating customers actively choosing the feature at booking and paying for it3

3

. The startup stated that anonymized footage would be accessible to customers on the app for 48 hours before being deleted from servers, and would not be shared with third-party AI labs3

3

. Despite these assurances, the pilot drew criticism over whether customers fully understood how recordings made inside their homes could be used to develop physical AI systems, raising questions around data storage, processing, and commercial dataset development3

3

.

Growing Demand for Real-World Training Data

Provision Labs enters a market with increasing demand for real-world training data needed to train robots to manipulate objects, navigate environments, and perform everyday tasks3

3

. Several startups are already tapping India's workforce to build these datasets. Y Combinator-backed Human Archive collects recordings of workers performing physical tasks, processes the footage, and sells resulting datasets to frontier AI labs and robotics companies3

3

. Companies like Awign and Objectways are building pipelines to supply first-person footage of human activity to AI developers3

3

. This emerging opportunity raises important questions around worker consent and compensation, privacy of individuals captured on camera, and how much value Indian data suppliers retain when resulting models and intellectual property belong to overseas companies3

3

. Watch for how Provision Labs addresses these ethical and commercial considerations as it scales its data collection operations and whether clearer boundaries emerge between Sardana's two ventures.