Pronto is now not positioning itself as only a family providers platform. Behind the fast progress story and investor pleasure, the startup can also be exploring one thing far greater and doubtlessly controversial: changing into a real-world information layer for Bodily AI and robotics.
And this isn’t merely trade hypothesis. The clearest indication comes from its investor paperwork reviewed by Entrackr.
In an inside memo, Glade Brook Capital states: “Pronto is looking for to formalize India’s huge casual labor markets and within the course of generate information to assist prepare bodily AI and robotics.”
The memo goes additional, saying the corporate is already “piloting actual world coaching information with main bodily AI labs.”
That line essentially modifications how the corporate must be seen.
Till now, Pronto was largely seen as one other fast-growing on the spot dwelling providers startup connecting households with employees for cleansing, laundry, utensil washing, cooking help, gardening, and automotive washing. Based in 2025 by Anjali Sardana, the corporate has expanded rapidly throughout high ten cities and attracted sturdy investor curiosity.
However the Glade Brook memo suggests Pronto’s ambitions lengthen past family providers.
In response to the investor be aware, the corporate is “growing an information enterprise leveraging its workforce to seize real-world family information for robotics labs.” The memo additionally says early partnership curiosity has been “encouraging” and that the corporate is “transferring rapidly to commercialize the technique.”
Responding to Entrackr’s queries, Pronto acknowledged that it has been working a restricted pilot round AI-related information initiatives.
“We have now been eager about this since our inception,” the corporate mentioned, including that clients can voluntarily select to have jobs recorded. In response to Pronto, the skilled carries “a small digital camera that faces outward on the work,” and clients obtain the footage afterward.
The corporate additionally added that Bodily AI programs require real-world behavioural information.
“One sort of information they want is first-person video of individuals doing actual duties, resembling washing dishes and folding laundry, in actual environments,” Pronto mentioned, including that work carried out by its professionals “can type a foundational information layer for bodily AI.”
The implications are important.
Not like web platforms that gather shopping exercise or clicks, Pronto doubtlessly operates inside probably the most non-public environments doable: folks’s houses.
If family workflows grow to be coaching information for robotics programs, then on a regular basis home actions might theoretically evolve into helpful AI datasets. How folks clear kitchens, fold garments, organise dwelling areas, wash utensils, transfer objects, navigate litter, or work together with home equipment could all grow to be related inputs for Bodily AI programs.
And that’s exactly why buyers globally have gotten on this class.
Bodily AI, in contrast to generative AI, requires real-world behavioural coaching. Robots want publicity to bodily environments and repetitive human actions to discover ways to function successfully inside houses and workplaces. Artificial simulations assist, however real-world information stays way more helpful.
The Glade Brook memo seems to immediately acknowledge this chance. It says Pronto is already executing greater than 25,000 orders per day whereas piloting “actual world coaching information” initiatives with Bodily AI labs.
Sources conversant in the matter indicated that the corporate has explored using mounted cameras or workflow commentary programs in sure operational settings.
Pronto mentioned utilizing cameras as a part of a restricted pilot linked to AI-related initiatives, however denied any large-scale deployment. The corporate mentioned the pilot is “strictly opt-in and chosen by the shopper at reserving, job by job,” including that “by default, nobody is in it” and that it has “no plans to increase it to nearly all of clients.”
When requested in regards to the pilot’s phrases and circumstances, Pronto didn’t share particular particulars and mentioned the initiative would stay restricted to pick out clients.
Underneath India’s Digital Private Knowledge Safety Act, 2023, consent have to be particular and purpose-bound. Agreeing to recordings for monitoring functions is legally completely different from consenting to their use as AI coaching information. Whereas Pronto frames the recordings as a customer support function, the investor memo describes them as potential AI coaching materials, elevating questions round compliance with the DPDP Act’s goal limitation necessities.
The corporate additional mentioned faces and figuring out particulars are blurred robotically, no personally identifiable info is uploaded or shared, and photographs is deleted inside 48 hours.
Nonetheless, Pronto’s claims round deletion additionally increase questions. The corporate says footage is deleted inside 48 hours and can’t be accessed by anybody aside from the shopper. But each Pronto and Glade Brook Capital individually describe the initiative as producing real-world coaching information for Bodily AI labs.
The 2 claims seem troublesome to reconcile, since AI coaching information sometimes requires curation, annotation, processing, and structured storage earlier than it may be commercialised or used to coach robotics programs.
If family workflow information is being captured or analysed for robotics purposes, the apparent query turns into whether or not clients and employees totally perceive what they’re consenting to.
May family layouts and behavioural routines finally grow to be business AI belongings?
Who owns such information?
Can it’s shared with third-party robotics firms or labs?
Would customers meaningfully perceive how Bodily AI programs are educated?
These issues grow to be sharper as a result of houses reveal excess of typical web behaviour. Home environments can expose deeply private way of life, cultural, monetary, and behavioural alerts. Even anonymised family datasets can carry extremely delicate contextual info.
The timing additionally issues.
Globally, capital is more and more shifting towards robotics, humanoids, warehouse automation, and embodied intelligence after the generative AI growth. Traders are inclined to imagine the subsequent AI giants shall be constructed not simply on compute and fashions, however on proprietary real-world datasets.
That makes Pronto strategically fascinating.
The startup has raised practically $60 million to date, together with funding from Glade Brook Capital and a latest $20 million funding from Lachy Groom that reportedly doubled its valuation to round $200 million.
There is no such thing as a proof suggesting Pronto is amassing information particularly for its investor Lachy Groom’s robotics pursuits or affiliated firms resembling Bodily Intelligence, which is constructing general-purpose AI fashions and studying algorithms designed to assist robots carry out real-world duties throughout environments. Nonetheless, the overlap between family workflow information, investor references to robotics coaching datasets, and rising investor curiosity in Bodily AI is certain to draw scrutiny.
Queries despatched to Groom and Glade Brook Capital on Thursday didn’t obtain a response.
Pronto mentioned its bigger goal is to assist casual employees take part within the AI financial system relatively than substitute them.
“Execs on Pronto get to take part within the AI financial system immediately and earn from the information their work generates,” the corporate mentioned.
The bigger query, nonetheless, extends past one startup.
If firms working inside houses start evolving into AI information infrastructure layers, India might finally grow to be one of many world’s largest suppliers of real-world family behavioural information for overseas robotics programs.
And in that future, the road between comfort platforms and AI infrastructure could grow to be more and more troublesome to separate.