In most offices, the newest colleague arrives with a laptop, a learning curve and a long list of questions. In a growing number of enterprises across India, the Middle East and the United States, the newest colleague arrives with none of those things — it simply joins the video call, listens to the discussion, and then gets on with the work. That colleague is HumAIn, and it is built in Hyderabad.
TensorGo, founded in August 2023 by Srini Chilukuri and Lakshmi Chilukuri, calls HumAIn the world’s first Pre-AGI teammate — an autonomous AI colleague that shows up and performs, rather than a tool that waits to be asked. The company says HumAIns join Zoom, Meet and Teams meetings as live participants, see and hear what happens in the room, reason over the organisation’s own knowledge, and then execute: interviewing candidates, closing deals, training workforces, screening patients, serving citizens and resolving customer queries, in more than 130 languages, around the clock.
Nia, one of TensorGo’s HumAIn AI teammates. (Image: TensorGo)
A teammate, not a tool
The difference, as TensorGo frames it, is between assistance and performance. A copilot drafts; a HumAIn delivers. The teammates read micro-expressions and physiological signals from live video, speak with natural lip-synced expression, and reason over institutional knowledge in under 200 milliseconds, according to the company. They remember what happened last time, carry decisions forward, and keep working across 38 time zones while the humans sleep. One HumAIn trainer, the company says, replaced the need to deploy trainers across 200 locations — the same depth, the same coaching quality, in every employee’s own language.
That is the pitch, and the deployments are where it gets tested. The company reports more than 15 live enterprise deployments, over 250,000 autonomous interaction minutes processed, and more than 16,000 job applications screened autonomously. In hiring, the company says time-to-hire fell by 94% and candidate conversion improved 3.5 times at one Fortune 500 enterprise, where HumAIn interviewed every single candidate with the same rigour.
More Read
One stack, built entirely in-house
What makes the story unusual is how much of the machinery TensorGo owns. Most AI companies assemble their products from third-party models, voice APIs and cloud infrastructure they do not control. TensorGo built every layer itself: hyperspace, its deep-learning platform for designing, training and deploying models; HumAIn Brain, the enterprise knowledge platform, which the company says retrieves answers in under 50 milliseconds; Voice Studio, the voice-intelligence layer that works across phone and SIP in 130-plus languages; and Go-X, the cognitive intelligence engine that ties it all together. The company says there are zero third-party AI dependencies anywhere in the stack.
That independence is the point. Because nothing in the system is rented, the whole platform can run on-premise, on a private cloud, or air-gapped inside a customer’s own walls — the sovereign deployment that Big 4 consulting firms use for HumAIn, according to the company. Global semiconductor leaders have gone further, co-publishing whitepapers with TensorGo on the hyperspace platform and optimising inference at the chip level.
HumAIn’s five core capabilities — to see, listen, reason, decide and execute. (Image: TensorGo)
In production, not on a slide
The customer list the company names spans Fortune 500 technology companies, global conglomerates, enterprise software providers and government institutions. In public services, the company says citizens across multiple Indian states now access government schemes in their own language, 24 hours a day, through a single HumAIn deployment. In manufacturing, HumAIns carry shift readings, observations and decisions across rotations, so nothing is lost at handover. In customer experience, the company reports satisfaction scores consistently above 90% while HumAIns handle the volume.
The recognition has followed the deployments. Gartner named TensorGo a Cool Vendor in AI; Oracle awarded the company its GenAI Innovation Award 2025 for the Japan and Asia-Pacific region; and Nasscom selected TensorGo for its Emerge 50, recognising it as one of India’s most promising deep-technology companies. The company reached production-grade enterprise deployment in eleven months from founding — a sprint its founders credit to two decades spent building enterprise software at scale, at companies including Microsoft, before they started TensorGo.
By 2030, TensorGo says HumAIn will power one billion enterprise interactions a year. Whether or not the number lands exactly there, the direction is clear: the work of building the intelligence that does the work is increasingly happening in Hyderabad, by a team of around 28 people that, according to Tracxn, has grown without raising venture funding — and the teammates it builds are already sitting in meetings on three continents, listening, reasoning, and getting on with it.
Keep reading: The AI Tutor That Speaks 12 Indian Languages: Inside YoLearn.ai’s Pragya, India’s First LLM for K-12 · No Signal, No Problem: Inside 169Pi’s Alpie Seed, the 9-Billion-Parameter AI That Runs Entirely on Your Phone · More Industrial Automation & Technology stories
Like stories like this? Add Exceptional Technology as a preferred source on Google to see more of our reporting in your search results and Discover feed.
