Faysal AzizAdvanced Cassidy AI Controls and Automation
Build reliable workflows with stronger logic, integrations, and oversight.
Why this course
Duration: 1 Day
Effective AI automation must do more than generate useful responses. It must follow business rules, use approved information, handle exceptions, and preserve human control over important actions. Cassidy AI provides these capabilities through configurable Agents, multi-step Workflows, permission-aware knowledge access, organizational instructions, and approval-controlled connectors.
This course concentrates on improving the reliability, control, and operational value of Cassidy AI solutions used for document processing, underwriting support, referrals, and business-system activities. The instructor has over 30 years of industry experience and will use real, industry-demanded content rather than an academic approach.
Learning outcomes
Participants will be able to:
- Design specialized Cassidy Agents with precise operating boundaries
- Build Workflows containing conditions, variables, approvals, and exception paths
- Configure controlled access to knowledge and connected applications
- Develop advanced underwriting and referral processes
- Test AI outputs and workflow behaviour systematically
- Apply governance and deployment controls
Prerequisites
- Practical familiarity with Cassidy Agents and Workflows
- Understanding of routine insurance and underwriting processes
- Access to an approved Cassidy AI workspace
- Approved non-confidential training documents
- No programming experience required
5 modules
01Advanced Agent and Knowledge Design3 topics
Specialized Agent architecture
- Defined responsibilities and boundaries
- Structured instructions and outputs
- Mandatory escalation conditions
Knowledge configuration
- Source and collection selection
- Conflicting information handling
- Permission-aware access
Organizational controls
- Global instructions
- Confidentiality requirements
- Decision-authority restrictions
02Advanced Workflow Construction3 topics
Workflow architecture
- Triggers, actions, and variables
- Sequential and conditional processing
- Reusable workflow components
Business-rule automation
- Eligibility conditions
- Authority and referral conditions
- Missing-information paths
Exception management
- Failed actions and integrations
- Manual intervention routes
- Workflow recovery controls
03Underwriting Process Automation3 topics
Submission review
- Document classification and extraction
- Completeness and consistency checks
- Risk-information summarization
Guideline and referral processing
- Risk appetite alignment
- Exception identification
- Referral-package preparation
Renewal and policy review
- Material change identification
- Claims and exposure comparison
- Outstanding requirement tracking
04Connectors and Human Oversight3 topics
Connected application controls
- Read and write permissions
- User and shared connections
- Approval-controlled actions
Human review design
- Risk-based review points
- Authority-based approvals
- Communication approval
Audit and traceability
- Source verification
- Action records
- Approval evidence
05Testing and Operational Governance3 topics
Agent and Workflow testing
- Normal and exceptional cases
- Restricted and incomplete requests
- Conditional-path validation
Output quality assurance
- Accuracy and completeness
- Source support
- Business-rule compliance
Controlled deployment
- Pilot scope and success criteria
- Ownership and monitoring
- Change management and ongoing review
This course outline is provided as a general instructional guideline and does not constitute a fixed or contractually binding programme. The trainer reserves the right to amend, reorganize, substitute, expand, reduce, or omit topics when reasonably necessary to address participant requirements, organizational priorities, platform developments, operational constraints, or professional considerations, without prior notice.
A programme built around your team.
Share your training goals and requirements.