Conversion.Business
Accessibility White Paper → View Live Demo →
Technical White Paper

Frictionless Security at Scale

Architecting a privacy-first, WCAG-compliant human verification primitive for the modern enterprise web.

1. The Problem with Legacy Verification

For over a decade, the enterprise standard for bot mitigation has been Google reCAPTCHA. However, as the web has matured, the costs associated with legacy image-grid verification have compounded into three critical liabilities for risk-averse organizations:

2. The Invisible Telemetry Engine

Conversion.Business solves this by replacing historical data analysis with real-time physics and behavioral heuristics. We verify humanity not by what the user has done previously on the web, but by how they interact with the page in the present moment.

Non-Linear Entropy Analysis

When our lightweight frontend widget loads, it begins tracking high-frequency coordinate vectors (mouse movements, touch trajectories). Automated headless browsers and basic scripts operate with distinct, linear mathematical efficiency. Humans, conversely, generate immense "entropy"—smooth, imperfect, non-linear curves.

Once a specific entropy threshold is met within acceptable timing bounds, the engine automatically verifies the session. The user is never required to interact with a puzzle; the security operates entirely invisibly in the background.

Mobile Safari & Throttling Mitigations

To preserve battery life, some mobile operating systems aggressively throttle high-frequency touchmove DOM events. To ensure mobile users aren't incorrectly flagged as bots due to sparse data, our telemetry engine applies heavy entropy weights to absolute touchstart coordinates, guaranteeing instant auto-verification even on highly restricted environments.

3. Uncompromising Accessibility (WCAG 2.1 AA)

To ensure 100% compliance with global accessibility standards, the telemetry engine is equipped with our proprietary Accessible Behavioral Bypass Gateway. Legacy visual CAPTCHAs attempt to solve this via audio challenges, which are frequently impossible for cognitively or hearing-impaired users, and often break on mobile devices.

We bypass the puzzle entirely. When the engine detects a screen reader or keyboard-only navigation, it exposes a visually hidden focus-trap that accepts explicit desktop keystrokes (Space/Enter) and native mobile screen reader gestures (VoiceOver/TalkBack double-taps). As the user interacts with this gateway, the engine measures the microsecond variance between their physical activations. Natural human timing variances (cadence cryptography) instantly generate massive entropy, allowing visually impaired users to bypass the security wall seamlessly while remaining impenetrable to scripted bots.

4. Cryptographic Handshakes (Roadmap to Server-Side Validation)

Client-side telemetry must be protected against malicious actors injecting fake JavaScript events. To secure the pipeline, Conversion.Business is transitioning to an industry-standard cryptographic handshake.

The V1 Architecture (Current)

Currently, during the V1 Beta period, the widget acts purely as a client-side telemetry engine. When the engine determines a session is human, it generates a raw JSON payload containing the clientId, a unique timestamp, and the verified: true boolean.

This payload is passed via postMessage to your frontend application. For V1 integration testing, your frontend can simulate validation, but this remains a client-side trust model.

The V2 Architecture (Upcoming)

In our upcoming V2 release, the raw telemetry payload will be transmitted to the Conversion.Business Node.js backend. Using the native Web Crypto API, our backend will sign the payload using a pre-shared secret key, generating a true HMAC SHA-256 signature.

When your user submits their form, your backend server will hash the payload with your private key. If the resulting signature matches, and the timestamp has not expired, you will have mathematical certainty that the request originated from an authentic human session—with zero risk of client-side key extraction.