Available from 25 August 2026

Claude Code for Investments

Build the research, due diligence and reporting agents an investment desk actually runs.

  • 5+ hoursOn demand, with ongoing updates
  • 6 modulesFrom scratch, no assumed knowledge
  • Live supportWebinars and office hours
claude — research

> Review the annual reports in ./filings and draft a risk summary

Read filings/annual-reports (24 files)

Grep "risk factors"

Write reports/risk-summary.md

Drafted reports/risk-summary.md. Every point cites the filing and page it came from.

> Turn that into an agent that re-runs it every Monday

Write .claude/agents/risk-monitor.md

Write CLAUDE.md (project conventions)

Created the agent and recorded the project conventions, so the next run needs no set-up.

>

Illustration of a Claude Code session. Folder names are examples, not real holdings.

Register your interest

Register and we'll send you the research and analytics workflow from the demo, along with access to the apps built on it.

Then one more email on 25 August 2026, when the course opens. Nothing else.

Why this matters to investment teams now.

Giving a strong investment analyst access to Claude Code is one of the biggest productivity unlocks available to a research team. Not because it replaces analysis, but because it removes the manual work that sits between an investment question and a well-supported investment view.

It doesn't just answer questions

Most AI tools return text. The analyst still has to gather source documents, extract tables, reconcile figures across reports, update spreadsheets, compile market data, and turn notes into a research memo.

Claude Code does the work. It gathers the documents, runs the analysis, writes the outputs, checks for inconsistencies, and leaves the analyst with something to review rather than something to finish.

One analyst, repeatable research

Every investment team has workflows that repeat. Periodic reviews, document and transcript analysis, comparable screening, valuation and pricing updates, portfolio and exposure monitoring, and investment committee materials.

With reusable agents and automated workflows, those processes are built once and run again whenever they are needed. Instead of rebuilding the same process every cycle, analysts review the results, investigate what changed, and spend their time on higher-value research.

The work compounds

The first reporting cycle requires building the workflow. The second requires refining it. By the fourth, much of the data collection, reconciliation, and report generation happens automatically.

Research becomes more consistent because the process improves over time instead of being recreated from scratch.

Judgment stays with the analyst

The mechanical parts of research, including finding information, collecting data, formatting outputs, and assembling reports, are increasingly automatable.

Deciding what matters, questioning assumptions, interpreting the evidence, and communicating a clear investment thesis remain the analyst's responsibility.

Claude Code handles the assembly work so analysts can spend more of their time thinking, investigating, and making informed judgments.

From setup to a system your desk runs on.

Six modules that move from configuring Claude Code safely to putting it to work on real research tasks, then packaging what works so the rest of the desk can use it.

Format
5+ hours on demand, with updates
Support
Webinars and office hours
Skill level
Beginner to Intermediate
  1. 01Claude Code Foundations
  2. 02Working with Data, APIs, MCP servers and Documents
  3. 03AI Driven Research and Due Diligence
  4. 04Automating Analytics, Workflows, Presentations
  5. 05Building Custom Apps
  6. 06The AI Investment Harness

What the course covers.

The techniques behind the demo above, and the material that comes with the course. Each is built up from first principles, so nothing here assumes you have met it before.

Techniques covered

SKILL.md
Write a procedure down once so Claude Code follows it the same way on every run, instead of re-explaining it each session.
Subagents
Hand a scoped job to a dedicated agent, keeping research, checking and drafting separate rather than one long thread.
MCP servers
Connect Claude Code to the systems that hold the numbers, so it reads from the source rather than a copy.
Loops
Run the same job across a folder of files, or on a schedule, without sitting and watching it.

Included

Skills guides
A written guide for each skill built during the course, to adapt for other workflows.
Code repository
The full course repository, cloned and run locally.
The demo apps
The research and analytics apps from the session above.
Claude Code folders
The project layout from the demo above, CLAUDE.md and .claude/ included, set up for investment work.

What you'll be able to do.

Each outcome is built and run during the course, not demonstrated on a prepared example.

  1. 01

    Configure and optimise Claude Code

    Understand agentic AI and modern AI workflows, and connect multiple financial data sources, including documents and APIs.

  2. 02

    Build specialised agents for research

    Generate investment memos and portfolio insights, and conduct due diligence.

  3. 03

    Automate existing workflows

    Automate analytics, report generation and on-brand slide deck creation.

  4. 04

    Create custom AI applications

    To integrate into your current investment or analytic processes.

Who it is for.

Analysts, portfolio managers and operations teams who already know what analysis they need and spend most of their time assembling the inputs for it.

The course starts from scratch. No coding background and no prior AI experience is assumed. It begins with installing and configuring Claude Code, and you will read and edit code by the end.

Requirements. A Claude subscription, which is separate from the course fee and not included in it. A laptop you can install software on.

Who runs it.

Michael Wang

Michael Wang, CFA

Investment
Fund management roles across analytics, multi-asset, prop trading and equities.
Data science
Ex Head of Data Science and AI, CAIO, Fintech.
Current
Founder and principal, WhyPred. Chief Executive, Data Science and AI Association of Australia.
Teaching
Taught at the University of Sydney, UNSW and Kaplan Business School.

One subscription, everything included.

An ongoing subscription, billed every quarter and renewing until you cancel. It covers every module, plus anything added while it is active.

Quarterly subscription

$399 per quarter

Billed every three months, renewing until you cancel. A Claude subscription is required and not included.

Seven-day money back

  • All six modules
  • Ongoing updates and new modules
  • Webinars
  • Office hours

The course opens 25 August 2026.

Register now for the demo workflow and apps, and you will be notified the day it opens.

Register your interest