# Build for the Memo, Not the Demo — Shawn Chan, China Resources Holdings AI Engineer · 2026-07-30 Shawn Chan, a veteran investor at China Resources Holdings who has sat on roughly 200 investment committees, argues that most AI finance products are built to impress for five minutes but cannot survive a room whose job is not to be impressed—the difference between a demo and a memo. He defines the memo as the real document that must survive an argument, with hundreds of pages where half the sources disagree. He recounts how one wrong sentence in a tech company's AI demo erased $100 billion in market value, how an AI invented six court cases for a lawyer's brief, and how an airline's chatbot made a fake policy that cost the airline in tribunal. He identifies six ways trust breaks: treating all sources equally, numbers that disagree, hiding contradictions, melting facts and guesses, unprovable claims, and no accountable human. His five fixes are: every claim linked to its source with trust level, facts and guesses visibly separate, automatic number reconciliation, surfaced contradictions, and a logged human approval gate. The product that wins, he says, is the one that lets a tired finance person trust its output without opening seven tabs at midnight. ## Questions this episode answers ### What does Shawn Chan mean by 'build for the memo, not the demo'? Shawn Chan distinguishes between a demo, which is a polished, fluent output designed to impress for a few minutes, and a memo, a comprehensive document that must survive rigorous scrutiny from an investment committee whose job is to find flaws. While demos prioritize smoothness, memos require trustworthy, verifiable, and reconcilable information, as real money decisions depend on trust, not just intelligence. [5:47](https://aiengineer.podhood.com/e9890546-bfde-48ef-9628-a2e44e4664f0?t=347000) ### What was the $100 billion AI demo mistake that Shawn Chan mentions? In February 2023, a major tech company's AI assistant gave a wrong answer about a space telescope in a marketing demo. That single unchecked sentence caused the company's stock to drop about 8% in a day, erasing roughly $100 billion in value. Chan uses this to illustrate that even a demo can fail the memo test when real money is watching, as every sentence becomes subject to verification. [7:31](https://aiengineer.podhood.com/e9890546-bfde-48ef-9628-a2e44e4664f0?t=451000) ### What five fixes does Shawn Chan recommend for building trustworthy AI finance products? Chan demands that every claim link to its source with a trust level, facts and guesses remain visually distinct, numbers automatically reconcile, contradictions are surfaced rather than smoothed over, and a human approval gate with an audit log is built in. These are not about smarter models but about 'plumbing and honesty,' ensuring a skeptical user can trust the output without fact-checking across multiple tabs. [20:56](https://aiengineer.podhood.com/e9890546-bfde-48ef-9628-a2e44e4664f0?t=1256000) ## Key moments - **[0:00] Confession** - [1:10] "Being sure of yourself and being right are two different skills," says Shawn Chan. - [1:32] Shawn Chan: Almost all AI finance products impress for five minutes but cannot survive a skeptical room. - [4:05] "Money doesn't follow intelligence. Money follows trust." — Shawn Chan - [5:46] Shawn Chan: A demo makes a room go 'oh' for five minutes; a memo must survive an argument. - **[5:47] Demo vs Memo** - **[7:31] Costly Typo** - [7:31] An AI demo's one wrong sentence about a space telescope erased $100 billion in market value in February 2023, says Shawn Chan. - [9:45] Shawn Chan: An AI tool quoted a group chat guess because it sounded more enthusiastic than the real audit number. - **[9:46] Source Trust** - **[11:22] Numbers Agree** - [11:22] Shawn Chan: A real estate algorithm's overconfidence in house prices caused a $500 million write-off and layoffs. - **[13:20] Contradictions** - [13:20] Shawn Chan: A contradiction between a CEO's earning call number and filing is a gift, not a bug — but AI silently picks the nicer version. - [14:59] Shawn Chan: 'Approval likely next quarter' can harden into 'approval received' across three drafts if facts and guesses aren't separated. - **[15:00] Facts vs Guesses** - **[16:48] Provenance** - [17:06] Shawn Chan: A New York lawyer filed a brief with six AI-invented court cases and the chatbot falsely confirmed they were real. - **[19:12] Accountability** - [19:21] Shawn Chan: An airline chatbot invented a bereavement discount policy, and the airline's defense that it was a separate legal entity failed in court. - **[20:56] The Fix** - [21:30] Shawn Chan's five fixes: link claims to source trust levels, separate facts from guesses, auto-reconcile numbers, surface contradictions, log human approvals. - [22:52] Shawn Chan predicts the winning AI product will be the one that lets a skeptical finance person approve a deal at 11 PM without opening seven tabs. - **[23:14] Pitch Memo** ## Speakers - **Shawn Chan** (guest) ## Topics AI Product Design ## Transcript ### Confession **Shawn Chan** [0:13] Good afternoon. Thank you for being here, day 4 of our conference, a room with no windows,right just after lunch. You're the strongest people in this building. Let me start with a confession: for 15 years, my job has been vancing. Sitting in a room, a very smart, confident person hands me a piece of paper, and before my company spends $100 million, I have to decide: do I believe this paper? This year, my job still exactly that. Except now, some of the time, the very confident person handing me the paper is a chatbot. And honestly, the chatbot is often better writing than humans. Better grammar, nicer formatting. Never get defensive when you ask a follow-up question. Very, very sure of itself. The problem is, being sure of yourself and beingright are two different skills. Some of the most confident people I have ever met in finance were also the most wrong. AI just learned that trick faster than the rest of us. So here's my whole talk in two sentences. First: almost every AI finance product is built to impress people for 5 minutes, and almost none are built to survive a room whose entire job is to not be impressed. Second, and this is the part I promised the organizers: at the edge, the exact same skills that fix your product are the skills that get investors like me to write you a check. Same muscle. I've proved both. And I'm telling this now, this week, because a lot of you are about to get pulled into exactly this: your CEO saw a demo somewhere, your biggest customer suddenly has a compliance department, or you're 3 months from releasing your next round. When any of those days arrive, I'd rather you hear the hard part from someone who's sitting on the other side of the table than discover them live in a room in front of the people who bill by hour. Quick bit about me. I promise this is the boring part, and I'll keep it fast. Fifteen years of cross-border deals: Hong Kong, Mainland China, the UK, the US. Mergers, IPOs, big strategic investment. On names you'd actually recognize: the kind of companies that go public and your LinkedIn feed won't shut up and about it for a week. Along the way, I've sat in about 200 investment committee meetings. I've agreed hires to prove it. I checked. It's not genetics. And here, the part that matters for the second half of this talk: I've also read hundreds and hundreds of pitch decks, founders decks, bankers decks, 47 slide seed decks. I have seen funds that should be illegal. 200 committee meetings taught me one thing: no textbook says out loud. The number on the page is not what gets a deal approved. Trust is a number. What gets a deal approved. Money doesn't follow intelligence. Money follows trust. And trust is fragile, especially when the thing that wrote the page has never once, in its entire life, said the words, "I'm not sure." Let me give you one small taste of what those rooms feel like. Early in my career, a very polished, very expensive banker presented a beautiful slide full of confident wrong numbers. One senior person in the room asked one quiet question: "Where does this number come from?" The banker posed what felt like an entire physical quarter. The pose taught me more about finance than 3 years of exams did. So I'm not sure as a—so I'm not here as a builder. I don't build these systems. I sit across the table from them and from the founders selling them, deciding whether to trust them. Today, I'm going to give you both halves of that: what breaks trust in your product, and what builds trust in your pitch. Let's define two machines that most everyone keeps confusing. Machine 1 is a demo. One clean document in, one fluent answer out, is the whole job: to make a room go "oh" for 5 minutes. Machine 2 is a memo. ### Demo vs Memo **Shawn Chan** [6:06] The real document is a real committee raise before real money moves. Hundreds of pages, filings, transcripts, broker notes, spreadsheets, someone's rush notes from a call last Tuesday. Half the sources disagree with each other. And the memo's job is not to make your go "oh." Its job is to survive an argument. Basically, a family dinner, except somebody's uncle brought a spreadsheet. Here's the everyday version of the gap: ask your phone to summarize a long email that a demo is not just sound plausible for 10 seconds. Now imagine that. Same phone has to stand in front of a bank and defend, out loud, why you deserve a mortgage. Suddenly, plausible isn't enough. Now it has to beright, and it has to prove it. That second situation is what a memo actually is. Now, you might think the demo world and the memo world never touch. Let me tell you about the most expensive typo in history. February 2023, one of the biggest tech companies on the earth launches its shiny new AI assistant with a promotional demo. ### Costly Typo **Shawn Chan** [7:40] In that demo, the assistant answers a simple question about a space telescope, and it gets the wrong answer. One sentence. One wrong fact about a telescope. In a marketing demo, the market noticed the company's stock dropped around 8% in a day. That's roughly $100 billion of value gone because of an unchecked sentence. $100 billion for one sentence. Nobody in that company asked the one question. Every junior analyzed on my team is trained to ask before anything leaves the building. Wait. Where does this claim come from? Did everyone check it? So here's the punchline: even the demo failed the memo test. The moment real money is watching, and real money is always watching. Every sentence becomes a memo sentence. There is no safe demo anymore. Keep that story in your head, because of the same trust-breaking moment happens in 6 smaller, quieter, very predictable ways inside of your product every day. Let's go through them fast. Six ways trust quietly breaks. Which source do you believe? Do the numbers agree? Do you hide contradictions or show them? Is that a fact or a guess? Can you prove it in 30 seconds? And whose name is actually on the decision? Six. Keep count with me. I'll keep each one short, and I'll bring receipts. Model 1. Not every source deserves the same trust, but most AI systems treat them like they do. Think of it like this: a number from an audit filing is your accountant speaking under oath. A number from an analyzed note is a friend at a party, confident, probably from someone's internal email. ### Source Trust **Shawn Chan** [10:14] It's a thing you overheard in an elevator. Most retrieval systems can't tell them this apart. They grab whichever text is close to your question and hand it over like a gospel. True story. Anonymized. I once watched a very expensive AI tool confidently quote a number from a group chat. Someone's rough guess texted 6 months earlier. The model loved the confident phrasing. The real audit number was 3 rows away in the actual filings. The AI just liked the group chat version better. It surrounded more enthusiastic. If your system can't tell an accountant under oath from a rumor in a group chat, it is not ready for real money. ### Numbers Agree **Shawn Chan** [11:23] Model 2. The numbers have to agree each other. Everywhere, every time, here's a memo that already died. Page 1 says revenue grows 18%. Page 11, in a little table nobody reads for fun, says 17.4. Nobody in the room cares about the missing 0.6. They care about what it means. If this person didn't check the easy mathematics, what did they not check on the hard stuff? That memo didn't pass. Not because of the number, because of what number you implied. And if you want the industrial strength version of this failure, remember the giant American real estate company that let algorithms buy houses at scale. The algorithm was extremely confident about house prices. The house disagreed. The company ended up writing off around half a billion dollars, shut the whole unit down, and let a quarter of the staff go. The model wasn't stupid. The model was unsupervised. Nobody built the boring machinery that forces the numbers to keep agreeing with reality after lunch day. Fluent and confident, remember, is not the same asright. Model 3. Surprises people. A contradiction is not a bug. A contradiction is a gift. If the CEO says one gross number on the earning call and the official filing says a different one, the gap is the single most interesting thing in the real story it hides. ### Contradictions **Shawn Chan** [13:45] Real diligence lives for that gap. AI does the opposite. It is trained to sound smooth and helpful. So when it hits, a conflict is quietly picked whichever version reads nicer and moves on. You never even learn there was a disagreement. I've sat through exactly these: CEO's number and the filing numbers. Meaningfully different. Nicer one. Nobody flagged it. Everyone just used the nicer one. We caught it because one person happened to have both documents opened at once. Pure luck. Luck is not a control. Your job as a builder isn't to resolve the argument. It's to make sure that the argument happens in front of a human instead of quietly alone inside a box. Model 4. Facts and guesses have to live in separate boxes. ### Facts vs Guesses **Shawn Chan** [15:08] And fluent AI loves melting them into one smooth sentence. Example: The company will likely receive approval next quarter. Reads like a fact. Sounds like a fact. It is a guess. Somebody biased the estimate wearing fact-shaped closing. A committee's entire job is to agree with the guesses while trusting the facts. If your system melts them together, the committee can't find them seams, and then all they can do is approve or reject the general vibe of the document. You should not spend $100 million on vibes. I watched approval accept soon. Turn across 3 drafts of a demo into approval received. Nobody lied. The guess just wrote its fact costume a little longer. Each rewrite until nobody remembered it started as a guess. The approval didn't arrive on schedule. That was an uncomfortable phone call. This fix almost embarrassingly cheap. Label your guesses. A tag. A color. Anything that survives being copy-paste into someone else's slide 3 weeks later. Model 5. If nobody can find where a claim comes from, it doesn't matter howright it is. ### Provenance **Shawn Chan** [16:59] You've all heard about the New York lawyer. He filed a legal brief written with a chatbot's help. The brief cited 6 court cases. Beautiful citations. Purple formatting. Very convincing. One small issue: the cases didn't exist. The AI invented all 6. And here, my favorite detail. The part that should be taught in school. Before filing, the lawyer got suspicious. So he asked the chatbot, "Are these cases real?" and the chatbot said, "Yes." That is like asking the guy who sold you the watch whether the watch is real. The judge fined them. The story went around the world. And my second favorite detail: the fake cases even had realistic surrounding names and page numbers. The AI didn't just lie. It's just the formatting. It cited itself beautifully. Wrong, but beautifully. The lesson is a 30-second test. When someone points at a sentence and says, "Show me where these come from," you either click once or land on the exact source paragraph, or you open 7 browser tabs and start swiping. I have personally been the guy with 7 tabs in a real meeting while the room full of people watched my scroll. 10 out of 10 would not recommend. If you remember only one sentence from this whole talk, the click-through is the product. Everything else is, well, written packaging. Model 6. My favorite. Because it's the most human someone has to sign. ### Accountability **Shawn Chan** [19:22] Here's the story you probably know. An airline website chatbot told a grieving customer he could book a full-price ticket now and claim a briefment discount afterward. That policy didn't exist. The chatbot made it up. Politely. Fluently. Confidently. The customer took the airline to a tribunal, and the airline's defense—"This is real"—was that the chatbot is, quote, "a separate legal entity responsible for its own action." That is the corporate version of "My dog ate my homework." The tribunal didn't buy it. The airline paid. And every one of us in a boardroom quietly took a note that day: you can't outsource accountability to your own software. At the bottom of every real decision are human signs. If your architecture doesn't have a fundable human at the end of it, you have not built a product. You have built an excuse generator. So build your AI around that accountable person, not instead of them. ### The Fix **Shawn Chan** [20:56] Okay. The fix. Five things. Each one is a direct cure for a story you have just heard. This is almost word-for-word what I'd demand from a vendor before lighting their system near a live deal. One: every claim comes with a receipt. Each sentence linked straight to its source paragraph with the source trust level attached. Not a citation tab at the end. Two: facts and guesses stay visibility separate. It lands at the page. I insistently see what's proven and what somebody best estimates. Three: numbers agree with each other automatically. The system refuses to shape a memo where the figures don't match. No human checking at 2 in the morning. Four: contradictions get surfaced. Never smoothed over. When sources disagree, the system raises its hand instead of picking the front-lier answer. Five: a real human approval gate. And it's logged who reviewed, what changed, when they signed. That log is audit trail. Notice what's not on the list. A smarter model. Not one of these is a bigger brain problem. All five: plumbing and honesty problems. The winners in this category won't win on benchmark points. They will win because a tired, skeptical finance person can't trust their output at 11 at night without opening 7 tabs. Now the part I promised: the money. Many of you are not just building AI products. You are rising for them. Or you will be. So let me tell you what actually happens after you leave the pitch meeting. Your deck becomes a memo. ### Pitch Memo **Shawn Chan** [23:36] Literally, someone like me sits down and writes an internal memo about you. Every number you said out loud gets checked. Gets checked against your data room. Which means everything I just told you about the documents applies to you personally. Full license. Okay. My time's up. Thank you. --- This library is powered by PodHood (https://podhood.com), the podcast website platform.