Ishavi is a knowledge-verification platform for hiring teams that want decisions they can defend -- to candidates, to auditors, and to themselves. Built on the conviction that an AI voice interview can be both rigorous for the recruiter and fair for the candidate, with neither serving as the excuse for the other to suffer.
01
Founders
Ishank is the founder and sole maintainer of Ishavi. The thesis is direct: an AI voice interview can produce the same evidence-anchored, appealable recommendation that a senior human interviewer would, at a fraction of the cost and without the inconsistency that comes from human schedules. The product is built end-to-end -- prompt composition, voice loop, rubric, scorecard, recording, appeals, and platform operations -- so the standards are consistent across every surface a candidate or recruiter touches. There are no investors, no board pressure, no growth-at-all-costs incentive. The constraint is taste, and the goal is a product that survives an audit.
Founded 2026 · Based in India · Reaches globally
02
Origin story
The AI interview category in 2025 had a clean split. On one side, surveillance-heavy platforms that treated every candidate as a suspect, ran gaze-tracking and keystroke-fingerprinting, and produced opaque numerical scores recruiters were asked to trust. On the other, lightweight chat tools that asked one or two follow-up questions, generated upbeat summaries, and failed to distinguish the candidate who really knew the work from the candidate who had memorised a template. Both sides optimised for the recruiter's convenience. Neither asked what the candidate's experience would be like if the system got it wrong.
Ishavi started from the other end of the loop. If a candidate disputes the outcome, what does the platform owe them? The answer became the product: evidence the candidate can read, an appeal path with reviewer rotation enforced at the database, jurisdiction-aware AI disclosure, accessibility profile that adjusts the interview itself rather than just the score, and a public Bill of Rights that the founder is personally on the line for.
01
Principle 01
The product is judged against both axes simultaneously. A rigorous interview that is not fair fails the candidate. A fair interview that is not rigorous fails the recruiter. We do not optimise one at the expense of the other.
02
Principle 02
Every score Ishavi produces is anchored to a transcript quote a reviewer can replay. Speed comes from the surface (one-click deep-link to the evidence) -- it never comes from skipping the citation.
03
Principle 03
Practice Mode, Re-Do, Human Escalation, Accessibility Profile, Scorecard Transparency. Five named rights, wired into the database with CHECK constraints and audit-logged transitions. If a feature would compromise any of them, the feature loses.
04
Where we are based
Ishavi is based in India and the data plane runs in Mumbai (ap-south-1) for the first year. The platform is region-aware from the first commit: every tenant carries a region tag, every domain row carries it end-to-end, and the application refuses to read across the boundary without an explicit, audited override. EU customers land on Frankfurt as soon as demand calls for it. US customers land on us-east-1.
The DPDP Act (India), GDPR (EU), CCPA (California), and UK GDPR are mapped one-to-one to the same internal controls. The compliance posture is published on the Trust Center and matures along a stated path -- pre-GA tenants get the same controls as GA tenants, marked by their certification stage.