Screening companies, building the pitch one-pager, updating the model, monitoring the book. The same workflows, rebuilt from scratch every time. The Leverage Program turns each one into a reusable Claude Code skill your team owns, and gives every investment professional 10-15 hours a week back.
Investment professionals are ambitious, high-agency, and expensive. Exactly the people whose judgment compounds returns. Yet the recurring grind of sourcing and analysis eats the week, and the know-how stays trapped in individual heads instead of becoming a shared asset.
Open a blank chat, re-explain the task, paste in the document, get an 80%-good answer, fix the output by hand, then start from zero again tomorrow. Useful, but it never compounds.
Each recurring workflow becomes a documented Claude Code skill the team runs on demand. When the output isn't right, you fix the instructions, so it's better for everyone, next time and every time after.
We start with the workflows that are documented, recurring, and data-heavy, the ones the models are unreasonably good at, where the judgment stays human and the grind doesn't.
A manager forms a hypothesis on an industry, researches the companies in it, and builds a templated one-pager per company to pitch internally. The single most recurring, time-sunk task on the desk.
Hours per company digging through annual reports and footnotes for private financials, then hand-assembling the same template in Excel and PowerPoint, and the data has to be right, or you lose the room you're pitching to.
Drop in the IM or prospect → the skill drafts the templated executive summary, pulling figures from reliable sources, and runs a built-in review loop that flags confidence before a human ever pitches it. The judgment stays yours; the assembly doesn't.
The same investment type, modelled over and over, the middle step between sourcing and pitching that quietly consumes the most time.
Rebuilding waterfalls and analysis from a blank sheet each deal, with structure and assumptions living in one person's workbook and nowhere else.
A skill that takes the source material and produces a first-pass model and analysis in your firm's structure, so the professional spends their time interrogating the numbers, not assembling them.
Standing on top of the book: counterparty and tenant risk, holdings you can query in plain language, and the recurring reports that go out to investors.
Hunting through scattered spreadsheets to answer basic questions about the book: what you hold, at what terms, against what plan, and closing recurring investor reporting by hand.
A queryable, always-current view of the portfolio, automatic counterparty risk reads from public filings, and reporting that assembles itself from a single skill.
Find the recurring, documented workflows hiding in your team's heads.
Capture each one as a reusable Claude Code skill the team owns.
When output falls short, fix the instructions, not the one-off result.
One improvement lifts every person who runs it. The asset grows.
The leverage you apply to capital, applied to your people.
A finance team's monthly investor reporting, once closed entirely by hand, collapsed into a single skill that does it all.
A support function's most repetitive, time-consuming tasks, automated end-to-end with reusable skills.
The target for a deal team: hours moved off assembly and back onto sourcing, analysis and judgment.
One two-hour session a week, over four weeks, with a cohort of five investment professionals. Across the program each participant builds one reusable skill for a task they all do, then shares it with the group at the end, so the whole team compounds. A firm runs as many cohorts as its headcount calls for.
Everything runs inside your own Claude Team or Enterprise environment, with no training on your data. A short infosec one-pager is available for compliance on first contact.
“I spent two years learning to build, went fully AI-native, and now I install that same capability in teams, one reusable skill at a time.”
I'm a former founder of a tech company, so I know first-hand what it costs to build something real and to make a team move faster. I learned to code inside an engineering team, then pushed AI into every workflow I touched. With 1,000+ hours operating Claude Code across both engineering and non-engineering work, I'm one of the few people with both the technical depth to build real workflows and the ability to teach them to people who never will.
I've already run Claude Code programs across multiple cohorts. Now I'm focused on the teams this fits best: private equity deal teams.
This is my AI clone. It knows the program, it knows how I think and work, and it can tell you whether this is a fit for your team, then collect the details I need to follow up personally. Talking to it is the fastest way to see what I'd help you build.