A multi-agent AI architecture designed for objectivity, confidentiality, and methodological rigour
Every assessment is produced by five independent AI credit analysts, synthesised by a presiding team lead — no single model, no human discretion
PreRatings does not rely on a single AI model, and it does not rely on human analyst discretion. Every rating request flows through a committee of five specialised AI credit analysts, each with a distinct analytical personality and weighting. They score the same seven factors independently, then a presiding team lead reviews their submissions, resolves disagreements, and synthesises a single consensus rating.
Because the pipeline is deterministic — the same data always produces the same output — no individual can be influenced to alter a rating. Objectivity is enforced by architecture, not by policy.
A conservative AI analyst focused on debt levels, liquidity buffers, downside protection, and worst-case scenarios. Applies the most demanding thresholds to stress-test the rating.
A growth-oriented AI analyst focused on revenue trajectory, market expansion, competitive moat, and future earnings potential.
A balanced, methodical AI analyst calibrated to the published frameworks of CRISIL and ICRA. Weights every factor equally and follows traditional rating methodology.
An AI analyst that weights external risk heavily: industry headwinds, regulatory risk, competitive intensity, and macroeconomic exposure.
A quantitative AI analyst that decomposes capital structure, cash-flow, and debt-service capacity through deep ratio analysis.
Does not score independently. Reviews all five submissions, identifies where the analysts disagree, and synthesises a single consensus rating.
Principles that govern every interaction — with clients, partners, regulators, and each other
We apply our 7-factor model consistently and objectively. Every rating request flows through the same multi-agent pipeline. No manual overrides, no special treatment, no favours. The same data always produces the same rating.
Client data is never disclosed, shared, or published. Unlike SEBI-registered CRAs, we have zero public disclosure obligations. Confidentiality is built into the platform — not dependent on individual discretion.
We do not invest in, advise, or have any financial interest in the companies we rate. We do not accept success fees, equity, or contingent compensation. Our only revenue source is the fee agreed in advance for each rating.
Our methodology is fully disclosed. Every client receives a detailed breakdown of how their score was computed, which benchmarks were applied, and what drove their rating. No black boxes.
Encouraging and protecting those who report genuine concerns about misconduct
PreRatings maintains a whistleblower mechanism that allows clients, partners, and other stakeholders to report concerns about unethical behaviour, fraud, data misuse, or any other serious wrongdoing — anonymously and without fear of retaliation.
Reports can be submitted to ratings@preratings.com. All reports are investigated promptly, thoroughly, and confidentially.
PreRatings expressly prohibits retaliation against any individual who, in good faith, reports a concern.
Ensuring objectivity and independence in every rating assessment
PreRatings has zero tolerance for conflicts of interest. Our architecture is deliberately structured to eliminate the conflicts present in the traditional credit rating industry. Specific safeguards include:
No human discretion in rating outcomes: The rating is produced by a deterministic multi-agent AI pipeline. No individual can be influenced to alter a rating, because no individual decides the rating — the same data always produces the same output.
No ancillary services conflict: We do not provide consulting, advisory, or restructuring services to companies we rate. We only provide preparatory ratings, so there is no incentive to inflate a rating to win downstream business.
No success-based compensation: Our fee is a flat, pre-agreed amount tied to the assessment itself. It is never contingent on the rating outcome, so there is no incentive to favour a higher or lower result.
Zero tolerance for bribery, corruption, or improper payments — in any form
PreRatings complies with the Prevention of Corruption Act, 1988 (India), and all applicable anti-bribery and anti-corruption laws in jurisdictions where we operate. We prohibit all forms of bribery, kickbacks, facilitation payments, and improper gifts.
Because every rating is generated by our AI system with no human discretion, there is no path by which an improper payment could influence a rating outcome. Key provisions: (1) No person acting on PreRatings' behalf may offer, promise, give, or authorise any payment, gift, or advantage to any person to obtain or retain business or secure an improper advantage. (2) All third-party intermediaries and channel partners must undergo anti-bribery due diligence and contractually commit to our standards. (3) Any suspected violation must be immediately reported to ratings@preratings.com. Violations result in termination of the relationship and may be reported to relevant authorities.