Why Tensile exists
An agent can sound right, pass a test and still leave the customer’s work unfinished. We are building Tensile to close that gap—with evidence that follows the work from production failure to a verified outcome.
Our conviction
A document obligation can disappear before it reaches a review queue. A correct query can bind the wrong financial period. A retry can create a second effect after the first one already committed. These are different failures with the same consequence: the customer cannot depend on the result.
— Investigate the first upstream failure, alongside the business context.
— Test the correction with the identities, permissions and state the agent will actually encounter.
— Read back the result outside the agent’s own account of what happened.
Explore the product ↗
Experience into product
Work across document review, financial analysis and multi-tool agents taught us to preserve what generic scores leave out. Source coverage, calculation lineage, content identity and lost responses after a write now shape the investigations, checks and environments in the product concept.
— A missed source becomes a coverage check.
— A wrong interpretation becomes a versioned expectation.
— A customer correction becomes a condition the next run must respect.
Read the engineering field notes ↗
Co-founders
Co-founder
Unmukt sets company direction and stays close to product and engineering work. His public background includes Chief Data Architect at Goldman Sachs, VP Applied AI/ML at JPMorgan, AVP at Airtel / Wynk, and co-founder of Thena.
LinkedIn ↗
Co-founder
Ankit leads engineering, architecture and technical delivery. He co-founded Thena, backed by Lightspeed and First Round; was Branch’s first international hire and general manager for Asia-Pacific, the Middle East and Africa; and was CTO of LetsIntern through its acquisition by Aspiring Minds.
LinkedIn ↗
Co-founder
Deepika is a Chartered Accountant turned product leader and AI-native builder. Her background spans risk advisory at Mahajan & Aibara, a CA practice supporting startups, and product roles at Scaler and Thena. She has worked on 0→1 products, internal tools and complex B2B workflows, bringing business context, product judgment and execution together to shape practical AI and automation solutions.
Meet the founders ↗
Bring the customer job, the expected result and the evidence your team would accept.
Discuss a concrete workflow ↗
AI reliability, from production to customer outcome.