Optura AIEOS User Manual

TOC (GP to update once all sections are finalized)

We will update all the better design images once the flow is finalized

1. Introduction

EOS equips enterprises with the tools to overcome common barriers to scaling AI initiatives. The platform addresses critical pain points such as fragmented AI initiatives and the inability to measure ROI effectively. Key features like Value Driver Management, ROI tracking, and a robust pipeline development framework empower organizations to identify high-impact projects and scale them with confidence.

To ensure immediate success, EOS integrates governance, security, and compliance workflows, providing organizations with the control needed to manage sensitive data and streamline AI program approvals. These foundational capabilities enable enterprises to transition rapidly from experimentation to scalable solutions that deliver measurable business outcomes.

The platform’s modular design ensures scalability and adaptability, allowing seamless integration of emerging AI technologies. EOS’s long-term vision positions it as a strategic enabler, helping businesses create sustained value, gain competitive advantages, and establish leadership in an AI-driven future.

Purpose of This Manual:

This manual serves as a comprehensive guide for administrators, analysts, and business users to effectively onboard, configure, and use AIEOS features. It includes step-by-step workflows, UI navigation guidance, conceptual overviews, and best practices.

Intended Users:

• AI COE Leaders

• Digital Transformation Executives

• Business Analysts and Ops Managers

• IT and Security Admins

• AI Developers and Data Scientists

System Requirements:

• Latest versions of Chrome, Firefox, or Edge browsers

• High-speed internet connection

• Access to email for MFA

• Admin-provided credentials or SSO integration

Onboarding Steps:

1. Receive invitation from admin

2. Configure user profile and department

2. Platform Overview

Vision and Architecture

Bridging the Short and Long Term

AI EOS bridges the gap between immediate operational needs and long-term strategic goals. In the short term, it focuses on essential capabilities:

• Value Driver Management: Capturing and prioritizing key business objectives.
• Pipeline Development: Building a robust pipeline of scalable AI opportunities.
• Security Management: Ensuring compliance and protecting critical data.

Over time, these capabilities evolve into the long-term vision supporting real-time simulations, autonomous orchestration, and seamless integration.

• EOS ensures enterprises achieve immediate results while laying the foundation for sustained innovation and growth in an increasingly AI-driven world.

Key Capabilities

• Intent Engine for aligning business objectives with AI projects

• Operator for orchestrating task execution and LLM integration

• Executive Dashboards for performance monitoring and decision support

• Agentic Workflows to scale execution across repetitive or complex tasks

Compliance and Governance

AIEOS embeds regulatory templates (e.g., HIPAA, SOC2) into workflows, tracks decision trails, and supports on-premise model routing to ensure security and compliance.

3. Getting Started

Login and Initial Setup

• Visit your company-specific AIEOS URL

• Login with credentials or

• Complete the MFA prompt

Company Configuration

Admins can:

• Define business units, departments, and initiative categories

• Map internal metrics to platform KPIs

• Assign primary use case owners and collaborators

Roles and Permissions

Role	Permissions Summary
Admin	Full access to settings and user management
COE	Review all use cases and approve
User	Only see existing use cases and submit new ones
Demo	Full access to give a demo ?

Connecting Data

Users can:

• Upload CSV batch datasets

4. Core Concepts

4.1 Intent Engine

• Role:

  • Serves as the foundational entry point for capturing business goals and aligning them with enterprise strategies.
  • Structures organizational objectives into a hierarchical framework to ensure coherence across departments and initiatives.
  • Refines broad objectives into actionable, detailed intents through a conversational interface that clarifies ambiguities and gathers missing details.

• Capabilities:

  • Hierarchical Structuring: Links intent to broader corporate goals, ensuring alignment from legal and regulatory compliance to department-specific initiatives.
  • Iterative Refinement: Engages stakeholders to refine intents using structured workflows, enabling consistent updates and re-prioritization.
  • AI-Driven Insight: Uses LLMs to interrogate intent, suggest enhancements, and flag conflicts.

  

  

  

4.2 Operator

• Role:
  • Manages the routing and execution of tasks through the appropriate Large Language Models (LLMs), tools, and systems.

4.3 Librarian

• Coming Soon in Next phase

5. Use Case Lifecycle

MVBC Statuses

IO = intent owner; CO = collaborator; CE = COE; AD = administrator

Status	Description	Permission (R,W)
		IO	CO	CE	AD
Draft	Initial status of intent, prior to initial pre-approval submission.	W	W
Identified	Use case has been submitted to COE for initial review	W	W	W
Refinement	COE is completing V&V and augmenting with any additional data required.
Approved	Approved for build. Will stay in this status until budget and build team is identified.
Build	Budget assigned and passed off to delivery team.
Delivered	Development complete and in production.
Realized	Use case has achieved it proposed value.
Archived	Archived use cases do not show up on COE list. They can be archived for any number of reasons, denied approval, duplicate,

MVBC Tags

Duplicate
Use case can be marked duplicate through the “Mark as duplication” function found in the item status dropdown.

Step 1: Create Use Case

• Initiate from Intent Engine or Use Case Library

• Define business owner, objective, expected ROI, risks

• Attach any available data inputs or mockups

Step 2: Scoring & Simulation

• AIEOS generates:

  • Value Score: Strategic impact, adoption potential

  • Viability Score: Technical feasibility, data readiness

• Optional: Run simulated agent to forecast potential KPIs

Step 3: Approval Workflow

• Stakeholders receive notification

• Review panel evaluates readiness and risk

• Approval is version-controlled with notes stored

Step 4: Pipeline Progress

• Once approved, use case appears in pipeline dashboard

• Track development, deployment, monitoring status

6. Executive Dashboards

Dashboards help business and AI leaders monitor:

• Use Case Impact: Adoption, value realization vs. plan

• Operational Metrics: Task completion rates, error rates

• Cost Insights: LLM usage, execution time, cloud costs

• Workforce Involvement: Departments, owners, decision logs

Features include:

• Interactive filters (time, department, stage)

• Export to Excel/CSV/PDF

• Role-specific widgets and alerts

View 1 : Use Case Pipeline

View 2 : Use Case Pipeline

Reports : Delivery

Reports : Pipeline

7. Model & Agent Setup

Registering Models

• Models can be imported via API or uploaded directly

• Tag with version, type (classification, generative, etc.), and owner

Access Controls

• Define usage rules by role, department, geography

• MFA required for modifying active production models

Agent Creation

• Combine model + rules + triggers = executable agent

• Example: "Eligibility Checker Agent" uses a preloaded ML model, public benefits criteria, and runs on claim submission event

Agent Workflows

• Drag-and-drop interface to build sequences:

  • Input validation → Model call → Result aggregation → Notification

• Support for branching logic, fallback steps, and retries

8. Monitoring & Audit

Traceability

• All agent executions are logged with input/output pairs

• Re-run historical traces to verify reproducibility

Logs

• Structured logs: Status codes, latency, model name, LLM cost tokens

• Errors flagged with context and suggestions

Audit Trail

• Tracks who changed what, when, and why

• Exportable for compliance audits

Version Control

• Each model, workflow, and use case maintains full history

• Compare diffs and rollback

9. Reports and ROI

Real-time Reports

• Visual and tabular reports updated with live data

• ROI dashboards show gains vs. forecast by initiative

Use Case Analytics

• Custom metrics: Time saved, revenue added, errors avoided

• Report builder for department heads

Export Options

• PDF, CSV, or via secure API

• Embedded in emails for exec distribution

10. Security & Compliance

• SSO and MFA: Azure AD, Google Workspace, Okta

• Encryption: TLS 1.2+ for data in transit; AES-256 for at rest

• Logging: All interactions logged for auditability

• Regulatory Modules: Pre-built templates for HIPAA, GDPR, SOC2

• Admin Alerts: Trigger alerts on anomalies or policy violations

11. FAQs & Troubleshooting

Login Troubles:

• Reset via email link

• Ensure browser cookies are enabled

Model Execution Fails:

• Check payload validation errors

• Verify model is active and agent is configured

Slow Dashboard Load:

• Ensure data sync jobs completed

• Contact support if latency > X mins/secs

Contact Support:

• Email: TBD
• Contact No.

12. Appendices

• Glossary: Detailed terms used in UI and documentation

• Acronym Index: AIEOS, ROI, HIPAA, MFA, COE, SOP

• Release Notes: Linked per quarterly release

• API Reference: Available at https://demo.optura.ai/