Trust Score Methodology

Transparency is fundamental to what we do. This page explains exactly how we calculate Trust Scores for every AI agent in our directory.

Overview

The Trust Score is a number from 0 to 100 that represents our editorial assessment of an AI agent's overall trustworthiness for Indian businesses. It is not a quality rating or a performance benchmark. It measures how transparent, credible, and compliance-aware an agent maker appears based on publicly available information.

The five scoring categories

1. Transparency and documentation (25 points)

We check whether the agent maker publicly discloses: pricing (clear, published pricing vs "contact us"), product documentation (public docs, API references, guides), privacy policy (does one exist and is it substantive), data handling practices (where data is stored, how it is processed), and how the agent works (any explanation of the underlying technology or methodology). Each item earns up to 5 points. Purely factual: either the information exists on their website or it does not.

2. Compliance claims (25 points)

We report what the agent maker claims about regulatory compliance on their website, documentation, and published materials. We check for mentions of: DPDP Act compliance, IT Act and IT Rules 2021, sector-specific regulations (RBI, SEBI, IRDAI, FSSAI), international standards (GDPR, SOC 2, ISO 27001, EU AI Act), and data localisation. We use careful language: "claims DPDP compliance" not "is DPDP compliant." We do not independently audit or verify these claims.

3. User reviews and reputation (20 points)

We aggregate ratings from G2, Capterra, Product Hunt, Google Play, App Store, and published testimonials. We calculate a weighted average. For agents with no external reviews, this category scores lower. As our platform grows, we will add our own verified review system.

4. Company credibility (15 points)

We assess publicly available information about the company: how long it has been operating, whether founders are identifiable, whether it has disclosed funding (from Crunchbase or press releases), whether it has a verifiable physical address, and whether it offers a real support or contact channel.

5. Feature completeness and integration (15 points)

We evaluate the agent's published feature list: whether it integrates with common Indian business tools (Tally, Zoho, Razorpay, etc.), whether it supports Indian languages, whether it offers an API, whether a free trial is available, and the breadth of integrations listed.

Data sources

All scoring data comes from publicly accessible sources: the agent maker's official website, published documentation, pricing pages, privacy policies, press releases, Crunchbase, G2, Capterra, Product Hunt, LinkedIn company pages, and Google Play or App Store listings. We do not use any private, confidential, or proprietary information.

Updates and corrections

Agent makers can request corrections or updates to their listing by contacting hello@agentvault.in. We review requests within 5 business days and update scores when publicly verifiable information warrants a change.