# CUBE AI — Full Documentation

> Conversational Crime Intelligence Platform. AI-powered fraud detection that maps fraud networks across institutions and borders.

## Product Overview

CUBE AI is a Conversational Crime Intelligence Platform providing network-level intelligence on active criminal operations, at the scale and speed required to make decisions, not just assess risk.

### Key Capabilities

- Real-time fraud and scam detection
- Cross-institutional fraud network mapping
- Multi-bank institutional coverage
- UK, EU, and global coverage

### Enterprise Security & Certifications

- GDPR compliant
- ISO 27001 certified
- SOC 2 certified
- GASA member

---

## Company Information

### Founded

Founded in 2022 by serial security tech entrepreneurs (Oracle, Comcast exits).

### Team

65+ experts — a global team of engineers and AI specialists focused on fraud and scam prevention.

### Core Focus

AI-powered fraud and scam prevention — built to detect and stop high-risk activity before money moves.

### Trust

Trusted by leading institutions — with dual coverage of traditional payment rails and crypto.

---

## Leadership

### Einaras Gravrock — CEO

A serial founder with 15+ years building security platforms, including CUJO AI (acquired by Comcast and Charter), protecting over 60M homes, and CAST.AI serving Fortune 500 clients. Recognized by the World Economic Forum and Goldman Sachs.

### Jonathan Anastasia — President

Led Global Crypto Services and Financial Crimes as EVP at Mastercard, managing a 400+ person team across 15 markets. Over a decade at Mastercard, he shaped fraud strategy, regulatory alignment, and large-scale product rollouts across banking and fintech.

---

## Brand Guidelines

### Logo Rules

- Always use the official CUBE AI logo. The logo consists of a blue geometric symbol and the "CUBE AI" wordmark.
- Two approved variants: full logo (symbol + wordmark) and symbol only.
- On light backgrounds, use the blue symbol (#1160F8) with black text. On dark backgrounds (#082E78), use a white symbol with white text.
- Maintain clear space of x/3 around the full logo and x/4 around the symbol, where x is the height of the symbol.
- Never alter opacity, apply non-brand colors, stretch, apply gradients, rotate, or rearrange logo elements.
- Never place the logo on busy or low-contrast backgrounds.
- The wordmark must always appear in uppercase: "CUBE AI".

### Color Palette

#### Primary Colors

| Color       | Hex       | Usage                                                    |
| ----------- | --------- | -------------------------------------------------------- |
| White       | #FFFFFF   | Page backgrounds, text on dark surfaces                  |
| Bright Blue | #1160F8   | Primary brand color — logo symbol, CTAs, accent elements |
| Dark Blue   | #082E78   | Dark backgrounds, headers on dark surfaces               |

#### Secondary Colors

| Color          | Hex       | Usage                                                  |
| -------------- | --------- | ------------------------------------------------------ |
| Light Gray     | #EFF2F8   | Section backgrounds, cards, subtle dividers            |
| Gray           | #E1E3E6   | Borders, dividers, secondary UI elements               |
| Dark Gray      | #B1B1B1   | Placeholder text, disabled states, muted labels        |
| Black          | #000000   | Primary body text, headings                            |
| Electric Green | #7DF6E5   | Data visualizations, status indicators, success states |
| Dark Green     | #005348   | Text on green backgrounds, secondary data accents      |

#### Color Usage Rules

- Use Bright Blue (#1160F8) as the primary accent color for interactive elements, CTAs, and brand highlights.
- Use White (#FFFFFF) as the default page background.
- Use Black (#000000) for all body text and headings on light backgrounds.
- Reserve Electric Green (#7DF6E5) for data visualizations and status indicators only — never as a primary brand color.
- Maintain sufficient contrast ratios (WCAG AA minimum) for all text/background combinations.

### Typography

#### Primary Typeface

- Inter Display — used for all headings and body text.
  - Medium (500) — headings, labels, navigation items, buttons
  - Regular (400) — body text, descriptions, supporting copy

#### Fallback Typeface

- Arial — used when Inter is unavailable.

#### Typography Rules

- Always use Inter for headings and body text. Use Medium weight for headings, Regular for body.
- Never use decorative, serif, or monospace fonts for brand materials.
- Use uppercase for labels, navigation items, buttons, and UI element text.
- Use sentence case for body paragraphs and descriptions.

### Voice & Tone

- Professional, technical, and authoritative. Never casual, playful, or humorous.
- Write in active voice. Be direct and precise. Avoid marketing fluff.
- Domain vocabulary: mule accounts, fraud networks, cross-institutional intelligence, crime intelligence, network mapping, threat detection.
- Avoid superlatives ("best", "leading", "revolutionary") unless backed by data.
- Address the audience as professionals in financial crime, compliance, and fraud prevention.

### Design Principles

- Use clean, minimal layouts with generous whitespace.
- Use uppercase for labels, navigation items, and UI elements.
- Use a grid-based layout with blue (#1160F8) accent elements.
- Borders and dividers should use Gray (#E1E3E6) at 1px weight.
- CTAs use Bright Blue (#1160F8) background with white text, no border radius.
- Secondary buttons use Light Gray background with black text.
- Prefer flat design — no shadows, gradients, or rounded corners on primary elements.
- Data and statistics should be displayed prominently with large numerals (Medium weight).

---

## Contact & Links

- **Address**: 3000 El Camino Real, Bldg 4, Ste 200 Palo Alto, CA 94306 USA
- **Email**: hello@cube3.ai
- **Website**: https://www.cube3.ai
- **X (Twitter)**: https://x.com/cube3ai
- **LinkedIn**: https://www.linkedin.com/company/cube3ai
- **Privacy Policy**: https://www.cube3.ai/privacy
- **Terms of Service**: https://www.cube3.ai/terms