The problem we're solving
Indian businesses are drowning in calls they can't answer. A telecaller costs about ₹2.8 lakhs a year — call-centre salaries jumped 18% in 2026 alone. Your customer speaks Hinglish, Tamil, Marathi, Punjabi; most agents only do English and Hindi. Three million people work in Indian customer service today, and the volume keeps growing. The supply doesn't.
What we built
An AI employee that talks like India — 14 Indian languages live today, including Hindi, Hinglish, Tamil, Telugu, Bengali, Marathi, Punjabi, Gujarati. Indian English by default, never American or British. Mid-call, it sends meeting links and brochures over WhatsApp, emails attachments, and books real calendar events. It remembers the conversations and picks up where the previous one left off — like an employee who actually paid attention.
Indian by design. Hosted on Indian servers, customer data stays in the country, caller consent captured at every touchpoint. Sub-200ms response latency so the conversation feels human. Every call has a recording, a transcript, a structured summary, and a tool-by-tool audit log — triage any call in 60 seconds.
Pre-launch. Already paying clients.
We're early, and the validation is across verticals that matter. Two paying clients on launch: ICAR (Indian Council of Agricultural Research, Ministry of Agriculture, GoI) and DocHome India (at-home healthcare platform). In active conversations with Shiv Nadar University and several Indian startups across edtech and healthcare. Backed by Masters' Union.
The founder
Karan Khanna is a builder at heart. Six years across enterprise automation, with stints as a Senior RPA Developer at Tech Mahindra (delivering automation for Graphic Packaging International, USA) and an RPA Developer at IBM, embedded with Unilever in Bengaluru where he shipped 10 end-to-end production bots. Currently a PGP candidate in Technology and Business Management at Masters' Union, with a Computer Science degree from KIIT Bhubaneswar.
Now building halloo.ai.