About DecisionPrint

Every Investment Decision Deserves a Record of Why.

Your brokerage statement tells you what you bought, what you sold, and how much you made or lost. But it rarely tells you why.

What did you believe? What evidence supported your view? What risks did you accept? And what would have made you change your mind?

Those answers often disappear into spreadsheets, research notes, conversations, and memory—if they were recorded at all.

Financial markets have built remarkable systems for analyzing prices, managing portfolios, and executing trades. Yet the reasoning behind investment decisions remains surprisingly difficult to preserve, evaluate, and improve.

We believe investing needs more than a record of transactions. It needs a way to continuously examine decisions, challenge assumptions, and learn from what happens next.

That's why we're building DecisionPrint.

Our Story

01 — The Observation

Everyone Knew Review Mattered. Few Had Time for It.

I began my career in institutional derivatives sales, spending much of my time speaking with hedge fund traders and portfolio managers.

Something kept coming up in those conversations: experienced investors already understood the value of documenting and reviewing their decisions. Many would tell me so themselves. But actually doing it properly was tedious, time-consuming, and difficult to maintain alongside the demands of managing money.

There was always another market move to understand, another position to manage, another trade to consider.

The problem wasn't that investors didn't value review. It was that meaningful review required time and effort they rarely had to spare.

02 — The Experience

The Outcome Was Never the Whole Story.

Later, as a family office CIO and investment allocator, I experienced these challenges from a different perspective.

I worked across asset classes and evaluated hedge fund managers pursuing different strategies, including quantitative and momentum approaches.

When markets sold off, my job wasn't simply to ask how much a manager had lost. I needed to understand whether the strategy had behaved as expected, whether its downside protection had worked, and whether the manager had respected the risk framework behind the allocation.

I also saw capable managers identify excellent investment opportunities yet struggle with position sizing and exits. Momentum investing makes this especially visible: a trade that looked like a successful exit at the time could later look like a significant missed opportunity as the trend continued.

But hindsight alone doesn't tell us whether that decision was wrong.

Evaluating an investment manager requires understanding not just the returns, but the decisions, assumptions, and constraints behind them.

03 — The Turning Point

What If the Record Could Take Care of Itself?

While working as a CIO, I started vibe coding my own internal investment research and portfolio management system.

I wasn't trying to build a software company. I was trying to solve problems I faced every day as an investor.

The system brought together work across asset allocation, portfolio volatility and drawdowns, hedge fund performance monitoring, derivatives exposures, attribution, and investment reporting workflows.

It wasn't an autonomous investment manager. It wasn't DecisionPrint. It was a practical attempt to make a lean investment operation more efficient.

But building it changed how I thought about what AI could make possible.

Work that once required hours of collecting information, organizing data, and preparing reports could increasingly be handled through automated workflows.

And it brought me back to the frustration I'd heard from professional investors years earlier.

If AI could dramatically reduce the cost of recording and organizing investment decisions, could it also make continuous evaluation and learning practical?

The opportunity wasn't merely to create a better investment journal.

It was to rethink how investment decisions themselves could be made, evaluated, and improved.

A Lesson From Poker

A Good Decision Can Lose. A Bad Decision Can Win.

Poker has been a constant influence on how I think about investing.

Both involve making decisions under uncertainty, with incomplete information and outcomes partly shaped by chance.

Give me pocket aces before the flop at a cash table, and if I have the opportunity to get all-in in a favorable situation, I'll take it—even knowing I might still lose roughly one time in five against certain hands.

Losing that hand doesn't make the decision wrong. Winning a hand doesn't automatically make a decision right.

Serious poker players understand this. They spend substantial time reviewing hands, studying alternative strategies, and examining where their reasoning could have been better.

And there's something particularly interesting about poker: many hands already have a record.

Hand histories and recorded games allow players to reconstruct decisions without relying entirely on memory. The analysis still requires thought, but much of the information needed to begin is already there.

When recording becomes effortless, systematic review becomes much easier.

Investing is more complicated. Markets don't reveal the true probabilities beforehand, and there is no universal solver that can tell us the correct investment decision.

That makes preserving what we believed, how confident we were, and what evidence we had even more important.

This idea became central to DecisionPrint: using AI not only to reduce the work of recording decisions, but to make evaluating, monitoring, and learning from them part of a continuous process.

The Idea Behind a Print

One Thesis. Multiple Paths. A Living Decision.

Every investment idea presents choices: act now, build a position gradually, wait for better evidence, or do nothing.

Choosing one path doesn't make the alternatives irrelevant.

A Print should preserve not only the decision that was made, but also the meaningful alternatives that were considered.

At the heart of DecisionPrint is the idea of preserving a decision's original evidence, assumptions, opposing arguments, risk constraints, and possible paths—its OUTS.

As new information emerges, Living Monitor is designed to detect meaningful changes and initiate fresh evaluations within the user's authorization boundaries.

The original decision remains preserved. New assessments, competing paths, and observed outcomes build upon its history rather than rewriting it.

Over time, comparing actual decisions with carefully evaluated alternatives can provide feedback for improving future judgment.

The goal isn't simply to remember a decision. It's to understand how it evolved, what other choices were possible, and what we can learn from the difference.

Thesis
Evidence
Alternatives
Checkpoints

What We Believe

01 — Intelligence Will Evolve. Accountability Must Endure.

We believe AI models may eventually become more capable investors than even the most accomplished human portfolio managers.

And we expect the capabilities of those models to keep improving.

Betting on having the smartest AI is a fragile business strategy.

The most capable model today may not be the most capable model tomorrow. We don't want DecisionPrint's long-term value to depend on owning the most intelligent model at any particular moment.

Instead, we want to build systems that become more useful as intelligence improves.

But greater intelligence doesn't eliminate the need for accountability.

In serious investment organizations, those who take risk are not necessarily the only people responsible for reviewing that risk. Investment committees, risk management, and independent oversight exist for a reason.

We believe the same principle should apply to autonomous AI agents.

The agent making a decision should not be the sole authority over the record used to evaluate it.

Regardless of how intelligent an agent becomes, investors should be able to examine what it believed, what it was authorized to do, what it actually did, and what happened afterward.

02 — Accountability Has Value. Better Outcomes Remain Our Ambition.

A well-researched investment can lose money. A poorly reasoned trade can produce an excellent return.

That's why transparent decision records, independent evaluation, and institutional accountability have value even when they don't directly generate alpha.

But accountability is also a starting point for improvement.

Without preserving what we believed before an outcome was known, it becomes much harder to distinguish weaknesses in a decision process from unfavorable results.

For active investors, generating alpha and improving risk-adjusted returns remain important objectives.

We believe that systematic evaluation, meaningful feedback, and disciplined experimentation can help identify weaknesses, refine strategies, and create opportunities to improve investment decisions over time.

Accountability has value. Better outcomes remain our ambition.

That ambition must be tested against real evidence, not assumed from the sophistication of the technology.

03 — Autonomy Should Be Governed, Not Micromanaged.

We believe AI should take on an increasing share of the intellectual work involved in investing.

Not everyone wants to spend their days reading financial statements, monitoring markets, or reviewing individual trades. Even professional investors should be able to delegate much of that work.

But greater autonomy requires clearly defined objectives, permissions, risk boundaries, reliable records, and independent evaluation.

The goal isn't to require human approval for every minor action.

Good governance should make responsible autonomy possible—not make autonomy impossible.

Our Vision

An AI-Native Investment Team. A Decision Infrastructure for the Future.

We don't believe the future of investing is simply a person asking an AI chatbot what to buy.

We believe AI should be capable of doing much of the work of investing itself.

Investors should be able to define their objectives, preferences, and risk boundaries—and delegate the work to AI agents: discovering opportunities, conducting research, evaluating competing ideas, exploring alternative strategies, and making decisions within clearly defined authority.

That is the investment system we're building toward.

But making decisions is only half the challenge.

Every decision creates alternatives that weren't chosen. Every investment thesis faces new evidence. Every outcome offers an opportunity to examine what worked, what failed, and what may simply have been luck.

DecisionPrint is designed to preserve those paths through its Prints, continuously revisit decisions through Living Monitor, and establish a foundation for systematic learning and improvement.

We don't just want AI to make investment decisions. We want a system capable of examining and improving how those decisions are made.

Our first step is a governed environment for investment research, evaluation, monitoring, and paper trading—where strategies and decisions can be examined without executing trades using live capital.

Over time, we envision making DecisionPrint's decision memory, independent evaluation, monitoring, and audit capabilities available to third-party agents and platforms.

Different models. Different strategies. Different agents.

We don't need to own the smartest AI to pursue that future. We want to build the systems that increasingly capable investment agents can operate within—and be held accountable to.

One accountable decision infrastructure for an increasingly intelligent financial world.

01Research
02Evaluate
03Decide
04Monitor
05Re-evaluate

Decision Memory

New Evidence / Next Review

The Founder

Chenhao Yu

Founder, DecisionPrint

Institutional derivatives professional, family office CIO and investment allocator, and builder of internal investment research and portfolio management systems through vibe coding.

A longtime poker player with a particular interest in probability, decision-making under uncertainty, and systematic learning.

Help Shape DecisionPrint

We're building DecisionPrint for investors who care about how decisions are made—and for a future where increasingly capable AI agents participate in financial markets.

Whether you invest independently, manage capital professionally, or build financial agents, we'd like to understand the decisions you face and where existing tools fall short.