Cascade Technical Documentation

Version Date: January 24th, 2025

1. Product Overview

Cascade is a generative AI platform designed to assist with employee support processes and inquiries. It automates routine employee support tasks and provides responses to employee inquiries at an individual level.

Two-Sided Product

Cascade operates as a two-sided product:

1. Employee-Facing Platform:
   1. Accessed via the web application at https://app.gocascade.ai.
   2. Employees can submit questions and receive contextual responses tailored to their specific needs.
   3. For additional convenience, Cascade offers a Slack app that integrates directly into existing communication channels, enabling employees to interact with the system within their familiar workflows.
2. Admin-Facing Platform:
   1. Accessed via a separate web application at https://portal.gocascade.ai.
   2. Employee support staff can manage access, configure integrations, review usage metrics, and oversee the data driving the AI model.

2.  System Overview

Architecture

Cascade is a cloud-based system hosted on Microsoft Azure. The platform leverages generative AI models from Azure OpenAI and Google Cloud’s Vertex AI to provide responses from inputs from users. As a web application, Cascade is accessible from anywhere with an internet connection. Cloudflare sits in front of all components to ensure robust security and performance.

Dependencies

Since Cascade uses third party systems, Cascade’s functionality and operability depends on availability of the systems listed below:

Name	Location	Website	Purpose
Microsoft Azure	USA	https://azure.microsoft.com/en-us/	Hosting of the Services
Google Cloud	USA	https://cloud.google.com/	Hosting of the Services
Cloudflare	USA	https://www.cloudflare.com/	Web security for the Services
Azure Open AI	USA	https://azure.microsoft.com/en-us/products/ai-services/openai-service/	Large language model processing
Vertex AI	USA	https://cloud.google.com/vertex-ai?hl=en	Large language model processing
Finch	USA	https://www.tryfinch.com/	API provider for HR data integration
Fullstory	USA	https://www.fullstory.com	Frontend app analytics and feedback
LogRocket	USA	https://logrocket.com	Frontend app analytics and feedback
Langsmith	USA	https://www.langsmith.com	Tracing and evaluation for model outputs
Sentry	USA	https://www.sentry.io	Error monitoring and performance tracking for applications
Elastic	USA	https://www.elastic.co	Search and analytics database and vector database

Integrations

Cascade AI is built to integrate  with existing HR ecosystems. The current available integrations include :

1. Single Sign-On (SSO)
   1. Works with enterprise identity providers (e.g., Okta, Azure AD) to simplify user authentication and authorization.
2. HRIS Systems via Finch
   1. Connects to popular HRIS platforms through Finch for real-time updates of employee and benefits data.
3. Ticketing & Workflow Tools (e.g., Jira)
   1. Allows complex HR cases to be routed and tracked in Jira, streamlining issue resolution and collaboration.

Authentication

Single Sign-On (SSO): Enables easy and secure access through integrations with leading identity providers like Okta, Microsoft, and Google.

UI Overview

• Employee-Facing UI: An interface accessible via https://app.gocascade.ai, optimized for natural language queries and personalized responses.
• HR-Facing UI: A management portal at https://portal.gocascade.ai for configuring settings, managing integrations, and reviewing analytics.

API Availability

Cascade does not currently offer APIs for direct integrations. However, we provide users with the capability to export data files from the platform.

Testing

Cascade runs automated tests on all features and updates before deploying changes to production.

3. Data Usage

Employee & HRIS Data

Includes basic profile information (e.g., name, department, role) and key records regarding benefits, payroll, and other HR processes. Cascade AI uses this data to tailor responses, automate HR tasks, and deliver precise, individualized support.

Interaction Data

Consists of user queries, feedback, and usage logs that help optimize Cascade AI’s accuracy and functionality over time.

4. System Requirements & Compatibility

Browser Requirements

• Internet access with a minimum of  2Mbps speed.
• Google Chrome (latest stable version), Mozilla Firefox, and Microsoft Edge are recommended.
• Safari (macOS/iOS) is supported, though certain features may require the latest version.

Operating Systems

• Cascade AI functions on Windows, macOS, and Linux devices with any modern web browser.

Network Requirements

• A stable internet connection (2 Mbps or higher).
• If applicable, VPN or firewall rules should allow traffic to/from Cascade AI’s authorized domains/IP ranges.

5. Security & Privacy

Encryption

Cascade ensures data protection by encrypting all data in transit using TLS/SSL and data at rest with industry-standard encryption algorithms.

Access Control

We implement role-based access control (RBAC) to ensure that users only have access to the data necessary for their roles.

Compliance Frameworks

Cascade follows security and privacy best practices aligned with frameworks such as SOC 2 Type 2 and NIST.

Monitoring & Incident Response

Our platform undergoes continuous monitoring to detect anomalies or unauthorized activity. Regular risk assessments and platform testing enable swift identification and mitigation of potential issues.

6. Implementation & Onboarding

1. Setup
   1. Administrators configure organization-wide settings, user roles, and SSO options within Cascade AI.
2. Integration Configuration
   1. IT teams connect Cascade AI to HRIS systems via Finch and set up Jira (or similar tools) as needed, following our API documentation.
3. User Training
   1. Cascade AI offers quick-start guides, video tutorials, and ongoing support to ensure HR teams and employees can fully leverage the platform’s features.

7. Implementation & Onboarding

While Cascade AI is designed to provide a secure, reliable, and efficient platform, there are inherent limitations to be aware of:

• No API Integration: Currently, Cascade does not support direct API integrations. Users can export data files as an alternative for data transfer and integration.
• Custom Infrastructure Dependencies: Cascade relies on third-party cloud providers such as Azure and GCP. While we leverage their robust security and compliance measures, our services are dependent on the availability and reliability of these providers.
• Role-Based Access Control (RBAC): While RBAC limits access to data based on user roles, it is the responsibility of customers to ensure role assignments and permissions reflect their organizational needs.
• Data Purging: Cascade depends on cloud provider protocols for data purging, which may vary based on the provider’s processes.
• Customer Responsibility: Security practices such as strong password policies for Single Sign On (SSO) must be maintained by customers to fully secure their usage of the platform.

8. Contact & Support

For further information, technical support, or to report any issues, please contact us via:

• Email: support@gocascade.ai
• Website: https://gocascade.ai

We encourage you to review our Privacy Policy and Terms of Service for additional details on how we handle data and maintain compliance.