Inspiration
Client onboarding in wealth management is a perfect example of an exception-heavy business process. A single onboarding case can involve email intake, multiple client documents, identity validation, tax review, address proof, duplicate client checks, risk scoring, compliance review, supervisor escalation, account provisioning, and audit closure.
In many organizations, these steps are handled through manual handoffs between Operations, Compliance, Supervisors, and provisioning teams. That creates delays, duplicate work, inconsistent decisions, and audit gaps.
We built WealthCaseIQ to show how UiPath Maestro, AI Agents, API Workflows, Coded Apps, Autopilot, and Codex with UiPath Skills can turn this process into an intelligent, auditable, end-to-end case journey.
What it does
WealthCaseIQ automates wealth management client onboarding from intake to closure.
The solution starts with an onboarding request and documents, then uses an AI Document Review Agent to classify and validate required documents such as the onboarding form, identity document, tax document, and address proof.
After document validation, the case moves through deterministic duplicate checking, deterministic risk scoring, human Operations Review, Compliance Review, Supervisor Escalation, account provisioning, audit closure, and final AI-generated audit summary.
The case can follow different paths depending on the facts:
- Missing or incorrect documents route to Operations or customer resubmission.
- Strong duplicate matches route to Operations Review.
- Medium, high, or screening-related risk routes to Compliance Review.
- Compliance exceptions can route to Supervisor Escalation.
- Approved cases move to provisioning.
- Failed provisioning routes back to Operations.
- Completed or rejected cases go through Audit Closure.
The result is a full case management journey with AI support, deterministic controls, human-in-the-loop review, and audit-ready closure.
How we built it
We built WealthCaseIQ using UiPath Maestro as the orchestration layer for the client onboarding case.
We used UiPath Autopilot and Codex with UiPath Skills to accelerate the development of the solution, including API Workflows, AI Agent prompts, Coded Apps for human review, structured JSON schemas, and deployment-related logic.
The case flow includes:
- Email intake and attachment handling
- Document Review Agent for document classification and extraction
- Document decision gateway
- DuplicateCheckApiWorkflow for deterministic duplicate validation
- Operations Review Coded App for human operational decisions
- RiskScoringApiWorkflow for deterministic risk scoring
- RiskSummarizationAgent for business-readable risk explanation
- Compliance Review Coded App for compliance decisions
- Supervisor Escalation Coded App for exception approval or rejection confirmation
- ProvisionAccountApiWorkflow for mock account provisioning
- AuditClosureApiWorkflow for final closure status and audit trail
- AuditSummaryAgent for final human-readable audit narrative
- Completed or Closed Rejected terminal case stages
A key design decision was to separate AI from deterministic business logic.
AI is used for document understanding, risk explanation, and final narrative summaries. Deterministic API Workflows are used for duplicate scoring, risk scoring, provisioning, routing, and final audit closure. Human review Coded Apps are used when Operations, Compliance, or Supervisor judgment is required.
This design keeps the process explainable, auditable, and repeatable while still using AI where it adds the most value.
Challenges we ran into
One of the biggest challenges was designing a realistic enterprise case flow without making the demo overly complicated.
We had to decide which steps should be AI-driven and which steps should remain deterministic. In financial services, it is risky to let AI make final business or compliance decisions, so we used AI for document extraction and summarization while keeping routing, scoring, provisioning, and closure rule-based.
Another challenge was building human review apps with clean input and output schemas. Operations Review, Compliance Review, and Supervisor Escalation each needed to support multiple decisions, null-safe inputs, next-stage routing, audit notes, and structured outputs for Maestro.
We also had to handle realistic case loops, including customer document resubmission, duplicate review, compliance send-back, supervisor escalation, provisioning failure, audit closure, and terminal case states.
A major learning point was making the solution modular. Each stage had to return a clean JSON result that the next stage could safely consume, even when some inputs were null or optional.
Accomplishments that we're proud of
We are proud that WealthCaseIQ demonstrates a complete enterprise-grade case management pattern, not just a single automation.
The project includes:
- AI-powered document review
- Deterministic duplicate matching
- Deterministic risk scoring
- AI-generated risk explanation
- Human Operations Review Coded App
- Human Compliance Review Coded App
- Supervisor Escalation Coded App
- Mock account provisioning
- Audit Closure API Workflow
- AI-generated audit summary
- Wait/resume and exception-routing patterns
- End-to-end case visibility
We are especially proud of the design pattern: AI assists the process, deterministic workflows control the decisions, and humans review exceptions. That makes the solution practical for regulated industries like wealth management.
We are also proud of how UiPath Autopilot and Codex with UiPath Skills helped accelerate development. They helped us move from design to working components faster while still keeping the architecture structured and enterprise-ready.
What we learned
We learned that agentic automation is most powerful when AI, workflow, and human review are used together with clear boundaries.
AI is excellent for reading and summarizing unstructured information, but deterministic workflows are still critical for compliance-sensitive decisions. Human reviewers remain important for exceptions, duplicate handling, compliance approvals, and supervisor decisions.
We also learned how important clean case variables, structured outputs, null-safe schemas, and audit-friendly JSON are when building Maestro case management solutions.
Using UiPath Autopilot and Codex with UiPath Skills improved our workflow by helping us design and iterate faster across AI Agents, API Workflows, Coded Apps, and deployment logic.
What's next for WealthCaseIQ
Next, we would expand WealthCaseIQ with deeper integrations and production-ready controls.
Planned enhancements include:
- Real CRM or core wealth platform integration
- Real KYC, AML, PEP, sanctions, and adverse media APIs
- Advisor portal for document upload and resubmission
- Automated customer/advisor notifications
- Case analytics dashboard
- SLA tracking and escalation rules
- Enhanced audit reporting
- Role-based reviewer experiences
- Additional onboarding scenarios such as trusts, joint accounts, retirement accounts, and entity accounts
- Deployment hardening using reusable Coded Agent patterns with Codex and UiPath Skills
The long-term goal is to make WealthCaseIQ a reusable UiPath blueprint for regulated, document-heavy, exception-heavy onboarding processes.
Built With
- codex
- gmailconnector
- javascript
- maestro
- uipath
- uipathactioncenter
- uipathapiworkflows
- uipathapps
- uipathautopilot
- uipathintegrationservices
- uipathorchestrator
- uipathskills
Log in or sign up for Devpost to join the conversation.