Vaibhav Dhawan
Marina McCulloch
FSI, Governance, Apps, Vibe coding, Developer API
FSI Dept Leads, Application Administrators
The challenge:
Financial services institutions operate under some of the most demanding AI governance requirements in any industry, from SR 26-2 model risk management guidelines to regional regulatory frameworks that vary by a combination of geography, job role and risk classification. Meeting these requirements typically means enforcing controls at every stage of the AI lifecycle: who can access what data, how models are validated before deployment, and what audit trails exist after the fact.
The challenge is not just compliance. It is compliance without friction. Heavy-handed governance configurations slow onboarding, frustrate end users, and ultimately push work outside the governed environment entirely, creating the exact shadow AI risk regulators are trying to prevent. Governance needs to be embedded in the platform itself, not bolted on top of it, so end users can be productive from day one.
The solution:
Domino’s LLM skills, covered in the Claude code blueprint, allow teams to build and deploy applications directly within the platform using coding agents. Combined with the extended identity propagation feature introduced in Domino version 6.2, administrators can create applications that act on behalf of the authenticated user, allowing new workflows to be embedded quickly and easily in the product.
In this blueprint, we use those capabilities to build a role-aware onboarding application for financial services teams. The app derives a user's business unit from their Domino organisation membership, maps their internal role to the appropriate Domino project permissions, and connects them to the governance documentation, compliance requirements, and use cases relevant to their team. Showcasing how governance controls can be embedded in the workflow from day one, not added as a separate step.
This approach works well when your organization is onboarding multiple teams with different roles, governance requirements, and access permissions. The application guides a user through a process allowing them to quickly discover learning paths, resources and best practices based on their role and organisation in the business.
Consider this approach when:
A few things to keep in mind before you build.
It is also worth noting that while this blueprint uses Claude, many other coding assistants, including Copilot and Cursor, recognize and incorporate Domino skills. You can also use internal models via the Claude CLI, keeping queries private and adding enterprise-specific context to your results.
The goal of this app is to both streamline the Domino experience for new users, while maintaining the context of their individual internal role and business impact at the org-level.
Since Claude Code on Domino already understands the Domino ecosystem, including how to develop and deploy apps within the platform, you can get straight into building. Rather than spending time on configuration, your focus is on providing the organizational context that makes the app useful: internal role structures, business unit requirements, and governance documentation specific to your FSI workflows.
In this blueprint, we us the VSCode IDE with the default Domino Standard Environment, and optionally the VSCode Domino Extension. This gives you immediate access to Domino skills and Domino’s coding assistant, as a result you can start shaping the onboarding app without any additional environment setup.
Once Claude Code is set up and confirmed working in your Domino workspace, you are ready to start building.
Before writing any code, start by giving Claude the overall goal and scope of the app. At this stage you are not asking for code. You are asking Claude to organize the information it needs to build something useful.
Initial prompt:
I want to build a Domino Onboarding app that incorporates knowledge, best practice and organisation/role specific onboarding scope for users joining the platform. This app should link business level roles’ skillsets with Domino best practice and recommend specific labs/code samples in accordance with the scope of the user’s org. Bare in mind this is an FSI onboarding app and our company is onboarding several different teams with their own specialities
Please hold off any code development as we need to gather:
Business Units / Organisations and the specific roles that mapWhat internal business roles map to which domino rolesInternal links to documentation on our internal model governance, AI policy (Both BU/Org level), Compliance and use cases for BU'sClaude response:
I'll help you plan this Domino FSI Onboarding app. This is a great initiative to streamline platform adoption across different business units with varying specialties. Let me organise the information gathering phase.With your next prompt you provide the specific details of your organization. Replace the example values below with your own business units, roles, and internal documentation links.
Second prompt:
Information needed:1. Organizations being onboarded include Retail Banking, Investment Banking, Risk ManagementWe have data scientists in all Orgs, we have Quant Analysts in Investment Banking, we have Model Validators and ML Engineers in Risk Management, we have Business Managers in all OrgsNo role hierarchies needed for now.
2. Domino Project Roles: All above internal roles except Business Managers have can have Project Owner access, they will be Domino PractitionersBusiness managers can be Practitioners or be Launcher UsersBusiness managers can also be lite users
3. Model Governance Doc: https://bank.sharepoint.comAI Policy Doc: https://bank.sharepoint.comCompliance Requirements: https://bank.sharepoint.comBU-Specific Use Cases: Separate sharepoint areas for BU/Org's include https://bank.retail.sharepoint.comhttps://bank.ib.sharepoint.comhttps://bank.risk.sharepoint.comInternal Domino Standards (naming conventions, environment standards, project templates): https://bank.confluence.com
App Capabilities:1. Users organisation will be derived from matching domino organisation i.e. Domino Organisation exists for all Orgs (Retail Banking, Investment Banking, Risk Management)User can select their Internal Role if it is relevant to the Org, this correlates with their domino role (i.e. Practitioner if regular data scientist or Launcher User/Lite User etc for a business manager)
Questions to consider:1. Hold off for now2. Hold off3. Hold off4. Streamlit app with Domino theme
Please create the structure and start the development of the App based on answers above.Claude response:
Perfect! I have all the information needed. Let me start building the FSI Domino Onboarding App. I'll create a structured data model and then develop the Streamlit application with Domino theming.Once the initial app is built you can add features that improve the user experience. The prompt below adds a personalized welcome screen, a support section, and a best practice section tailored to each role.
Here’s an example of such a prompt.
Refinement prompt:
Final feature changes.
When a user clicks on the onboarding app, they should be greeted by their name and welcomed to the onboarding platform before they select their role within the organisation.
Could you also add two extra sections: “Support” and “Best Practice”
Under “Support” please provide a link to our internal incident ticket system and best practice for incident creation, i.e. uploading evidence to your problem ticket, providing correct URL and context for support
In “Best Practice” please include details about key tools, programming languages for the user’s specific role.Claude response:
Perfect! Let me enhance the app with these important improvements. I'll:
Add a personalized welcome greeting with the user's nameCreate a dedicated "Support" section with incident ticket system infoEnhance "Best Practice" with role-specific tools and programming languagesWe’ll conclude with our theme prompt, since we’re developing in a Domino workspace we can use the domino-ui-design skill to help us publish a sleek Domino themed app.
Theme prompt:
/domino-ui-design
Without changing any existing backend code, please use the Domino UI theme to improve the Streamlit app frontend Claude response:
I'll help you improve the Streamlit app frontend to use the Domino UI theme. Let me re-examine the current codebase to understand what we're working with.In four prompts, you have built a working FSI onboarding app that surfaces role-specific governance documentation, learning paths, and Domino project permissions automatically based on each user's business unit. This is a foundation you can continue to build on.
When expanding the app with additional content and features, follow your organization's data protection and compliance policies when storing or linking internal documentation in your Domino app.
Once your Domino SysAdmin has set up Organisations that reflect your internal business structure in your Domino deployment, you can extend this app to call the Domino REST API endpoints directly, connecting users to their correct organisations without any manual selection. You can also link role-specific project templates to give users a relevant place to start on their first model development project.
From here, you can continue to build on top of this using Domino skills for Claude:
Contact your Domino Professional Services team who has expertise in applying enterprise governance for FSI across geographies and can help you get the most out of your platform.
Principal Solution Architect
I work to support large and complex customer deployments to meet their requirements for security, cost, tool integration, data and processes both in the cloud and on-prem. A number of these solutions and best practices are packaged into reusable Blueprints for our larger customer base, and some are later integrated into the Domino platform.
Solutions Engineer
Marina is a Solutions Engineer at Domino Data Lab, she works with Financial Services (FSI) customers to deploy and scale the Domino platform. She provides technical guidance on architecture, configuration, and best practices, helping organizations successfully adopt the platform and realize business value. She supports implementations end-to-end, including environment setup, upgrades, and platform enablement, while partnering closely with support teams to troubleshoot and resolve issues. Her role combines expertise in Kubernetes, Docker, Cloud Platforms, and Python to deliver tailored solutions aligned to customer needs.
