Intelligent CPD_Louis Mutton

Intelligence CPD uses AI to turn anonymized client needs, advisor skill gaps, and CPD data into targeted learning paths, next actions, and admin insights.

Comment

Inspiration

Intelligent CPD was inspired by the reality that advisors are expected to serve clients well while also keeping up with continuous professional development. In a fast-changing advisory environment, learning cannot be separated from client needs. Advisors need to know not only what courses are available, but what they should learn next based on the clients they are preparing to serve.

We wanted to build something more useful than a normal CPD tracker. The idea was to connect anonymized client demand, advisor skill gaps, CPD progress, and learning recommendations into one practical workflow. In simple terms, the platform helps answer:

  • What do clients need?
  • What is the advisor missing?
  • What should the advisor learn next?

What it does

Intelligent CPD is an AI-powered advisory learning platform. It helps advisors and administrators make better learning decisions using real operational signals.

For advisors, the platform provides:

  • A general CPD dashboard with progress, completed courses, and next study priorities.
  • An Insert Client tab where advisors can add anonymized client information.
  • AI-generated next actions for newly inserted clients.
  • Client Insights showing readiness, product/content demand, and client issues or learning risks.
  • AI-generated learning paths based on client insight statistics.
  • Saved Learning Path records for advisors to revisit recommended learning plans.
  • A searchable CPD course library.

For administrators, the platform provides:

  • Firm-level CPD health.
  • Compliance risk visibility by advisor.
  • Expandable advisor details.
  • Client interest and skill gap summaries.
  • Admin insight cards generated by AI to make analysis easier to understand.

The main value is that CPD learning becomes connected to business needs. Instead of advisors choosing random courses, the system recommends learning based on client demand, skill gaps, and case signals.

How we built it

We built Intelligent CPD as a full-stack web application using TypeScript.

The frontend is built with React, Vite, Tailwind CSS, Recharts, and Lucide icons. The interface uses a glassmorphism-inspired design system and is organized into separate advisor and admin workspaces.

The backend is built with Express and connects to Supabase for structured data storage. The main database tables include:

  • advisor_profiles
  • client_profiles_anonymized
  • advisor_client_cases
  • advisor_cpd_progress
  • cpd_catalog
  • saved_learning_paths
  • ai_recommendation_runs

The AI flow is organized around three advisor-facing agents:

  • ClientInsightsAgent identifies client urgency, product demand, and loss analysis.
  • SkillGapAgent identifies what skills the advisor is missing.
  • RecommendationAgent recommends validated CPD learning paths using real catalog course IDs.

We also use AI to generate short, practical next actions when a new anonymized client is inserted. The system avoids collecting personal client information and uses client_code values such as CLIENT-001 to keep client records anonymized.

Challenges we ran into

One challenge was making AI recommendations useful without allowing hallucinated courses. To solve this, the backend validates AI-generated learning paths against real cpd_catalog course IDs before showing them.

We also ran into database constraint issues while inserting clients. For example, some fields such as priority level and case status had database rules that needed to match the frontend options. We adjusted the form and backend logic so inserted records follow the actual Supabase schema.

Another design challenge was making analytics understandable. Long AI paragraphs were hard to scan, especially for administrators. We redesigned the admin AI output into concise insight cards with severity, key metric, short insight, and recommended action.

Accomplishments that we're proud of

We are proud that Intelligent CPD is more than a static dashboard. It connects multiple pieces of the advisory workflow:

  • Client needs
  • Advisor skills
  • CPD progress
  • Course catalog data
  • AI-generated recommendations
  • Admin monitoring

We are also proud of keeping client data anonymized while still making the insights meaningful. The advisor can understand what type of client demand exists without exposing personal client details.

Another accomplishment is the saved learning path flow. Advisors can generate AI learning paths from Client Insights, save them, and revisit them later. This makes the recommendation system feel like a real working tool rather than a one-time AI response.

What we learned

We learned that AI becomes more valuable when it is grounded in structured data. A generic AI recommendation is easy to produce, but a useful business recommendation needs context, validation, and clear constraints.

We also learned the importance of designing for different users. Advisors need direct actions and learning priorities, while administrators need organization-level visibility and compliance risk. The same data can support both roles, but the interface must present it differently.

From a technical perspective, we learned how to combine React, Supabase, Express, and an OpenAI-compatible AI API into a working workflow. We also learned how important it is to handle missing data, schema constraints, duplicate saved records, and invalid AI output gracefully.

What's next for Intelligent CPD

The next step is to make Intelligent CPD more proactive. Instead of only showing insights when advisors open the dashboard, the platform could send reminders when client deadlines are near or when an advisor has a high-priority skill gap.

Future improvements could include:

  • Calendar integration for upcoming client meetings.
  • More detailed partner or referral ecosystem visibility.
  • Admin approval flows for recommended learning paths.
  • More advanced CPD completion tracking and certificates.
  • Branch-level benchmarking and advisor performance trends.
  • Stronger audit trails for AI-generated recommendations.

Long term, Intelligent CPD can become an organization-wide learning intelligence platform. Its goal is to help advisory firms prepare advisors better, improve client conversations, and make professional development more targeted, measurable, and sustainable.

Built With

Share this project:

Updates