Danny Morris
← Back to work Case study · March 2026 · Client trial

Wealth management pipeline

A content system for a wealth management brand's YouTube channel, aimed at high-income professionals.

Tools · Claude Projects, ChatGPT, deep research across several models, Google Docs Role · Solo build
73/100trial score, with strong marks on systems thinking and architecture
9pillars, each covering one area of the brand
9stages in every outline

The problem

The channel targets high-income professionals: a sophisticated, sceptical audience. The owner needed research-backed, on-brand video outlines without researching every topic by hand, and needed to stay credible while publishing more, and more consistently. The brief arrived scattered across several documents, so the first job was pulling the brand's positioning into one place.

Before Manual research for every topicNo consistent structure across videosSources checked case by caseRepurposing done separately each time
After Structured outlines from a topic inputOne nine-stage structure, with optionsSource checking built in, with citation linksShort-form clips drawn from the same checked sources

What I built

A nine-pillar system in Claude Projects, with a knowledge layer drawn from the client's own documents. Each pillar covers one area and keeps the output from straying from it:

  • Foundation: core mission and non-negotiables
  • Brand identity: tone, positioning and output standards
  • Audience: who they are, how they think and what they respond to
  • YouTube performance: titles, hooks and retention
  • Voice and script: rhythm and natural framing, so it doesn't read as generic
  • Plus: short-form, written content, success metrics and channel status

Why my video background mattered

The hooks follow retention principles from years of editing. The nine-stage outline mirrors how I build a video narrative: make the strongest case against the idea first, then turn it. The credibility test came from knowing what makes finance content trustworthy and what makes it feel like clickbait. I planned the short-form clips around natural, quotable moments.

What went wrong, and the fix

The first outline took a stance that contradicted what the client had already said on camera. The cause: the client's reference scripts hadn't been loaded into the system before that first output. I diagnosed the gap, loaded the scripts into the knowledge layer, and rebuilt the outline around the client's own framework. The revised outline matched the channel's strategy and was approved. The system now requires reference material before it produces anything that depends on positioning.

A sample of the output

Redacted. The client's name and positioning are removed, and the structure and sources are condensed.

Video outline If AI replaces white-collar jobs, why might stocks still go up? Angle: who actually buys stocks when workers stop? It starts from the viewer's fear, then gives a structural reframe they haven't heard elsewhere. Hook A, time-coded [0:00-0:05] AI is going to take millions of white-collar jobs. [0:05-0:10] I've made an entire video about why that's probably true. [0:10-0:18] But here's the question nobody is asking: if the people who own 80% of the stock market lose their jobs, who's left to buy stocks? [0:18-0:20] [PAUSE - let the question land] [0:20-0:25] The answer surprised me. [0:25-0:30] Because it might mean markets go UP during the transition, not down. Structure Hook · Strongest case against · The reframe · Historical proof · The three buyers · When this thesis breaks · Positioning · What the smart money is doing · Close
ClaimSource
White-collar workers own 80% of stocksFederal Reserve, Survey of Consumer Finances
S&P 500 buybacks hit a $942.5B record in 2024S&P Dow Jones Indices
Norway's fund: $2.2T, 71% in equitiesNorges Bank Investment Management

Every claim in the delivered outline came with a clickable source. Looking back, the first one needed a second check: the Federal Reserve data shows the wealthiest households own most stocks, not white-collar workers as a group.

The outcome

I delivered the system and the client approved the revised outlines. The trial scored 73/100, with strong marks on systems thinking and architecture, and the main issue was the positioning error above, which I fixed. The project was at outline stage when the trial ended, with no video filmed yet. The engagement did not continue beyond the trial.

← The 600 Next case study: Find Catering Engineers →