What HARDO Actually Is — And Why We Built It This Way

Most investment banking interview prep teaches you to recognize questions. It doesn't teach you to answer them while a VP is checking his phone, cutting you off at minute three, and asking why your terminal growth rate is 3% instead of 2.5%.

That gap — between knowing the answer and delivering it under pressure — is where most candidates lose offers. Vault guides, YouTube walkthroughs, and 400-page question banks are pattern libraries. They help you memorize the shape of an LBO question. They cannot simulate the moment an associate interrupts your sources and uses table to ask what happens to IRR if exit multiple compresses 2 turns.

HARDO is built for that moment. It's an AI mock interview platform with three difficulty tiers, twelve questions per session, and interviewers that push back when they sense weakness. This piece explains what it is, what it isn't, and why we built it the way we did.

The Gap Between Knowing and Performing

Ask a second-year student to explain accretion/dilution at a whiteboard with no time pressure and most of them will get it right. Ask the same student the same question over Zoom, with a banker who looks bored and three follow-ups queued up, and roughly half will fold.

The material didn't change. The conditions did.

Interview performance is a function of working memory under interpersonal stress, and working memory collapses faster than people expect. You can know the formula for the weighted cost of acquisition cold, then forget it the moment someone asks you to defend the assumption that the buyer's after-tax cost of debt is 4%. You start over. You hedge. You say "I think" three times in one sentence. The interviewer writes "shaky on technicals" and moves on.

The only fix is reps under conditions that approximate the real thing. Reading more does not solve this. Watching more does not solve this. You need someone, or something, sitting across from you asking "why" until you run out of answers, and you need to run out of answers in practice rather than in the interview that decides your summer.

How HARDO Is Structured

Three interviewer personas, each calibrated to a different stage of recruiting.

Intern is a patient VP. Encouraging tone, hint-style follow-ups, willing to accept a conceptual answer when you can't push to numbers. This is the simulation for first-round screens, HireVues, and the early networking conversations where the bar is "does this person know what a DCF is and can they explain it without panicking."

Analyst is a busy staffing VP. Uses IB shorthand, pushes on numerical edge cases, gets impatient with vague answers. "Walk me through how you'd value a private SaaS business with negative EBITDA" — and then ten seconds later, "okay but what multiple are you actually using." This is the simulation for superdays and lateral analyst screens, where the bar is "can this person hold a number in their head and defend it."

Associate is a tough MD. Simulates CFO, board, and investment committee pushback. The question isn't whether you know the answer. It's whether you can hold your position when someone with twenty years of experience tells you you're wrong. This is the simulation for associate-level interviews and for serious analysts preparing for client-facing reps.

Each session pulls twelve questions across eleven categories: Accounting, Valuation, Corporate Finance, Case Study, M&A, Private Equity / LBO, Behavioral / Fit, Business Acumen / Markets, Brainteaser, Curveball, and Due Diligence. The AI generates follow-ups when it detects weakness: vague phrasing, anchoring on a number you can't defend, skipping a step in a walkthrough. Case studies allow up to five follow-ups, because that's roughly how deep a real interviewer will go before they decide if you can structure a problem or not.

You can answer in voice or text. Voice is closer to the real interview and we recommend it, but text lowers the activation energy when you're squeezing in reps between classes.

Why Grades, Not Vibes

Most prep gives you one of two outputs: a thumbs-up, or a vague "good answer, but try to be more concise." Neither is useful, because neither tells you what specifically broke down.

HARDO grades each answer against a rubric for that category (Valuation answers are scored differently from Behavioral answers), and the letter is hidden until the end of the session. The hiding is intentional. If you see grades in real time you start playing for the grade rather than playing the interview, which is exactly the wrong muscle to build. At the end, you get a scorecard: letter per question, category averages, and a four-level hire recommendation (no hire, leaning no hire, leaning hire, hire).

The letter itself doesn't matter much. What matters is what sits next to it. A B− on a DCF walkthrough means nothing without the note that says "you anchored on a 12% growth rate for years 6 through 10 without justifying why growth would persist that long after the explicit forecast." That sentence is the rep. The letter just forces the feedback to be specific.

We grade hard. An A means you delivered an answer that would hold up against a real MD, not against an algorithm being polite. Most candidates start in the C/D range across most categories and that's normal. The point is to find the categories where you're weakest and run them until they aren't.

What HARDO Is Not

A few things, because positioning matters.

HARDO is not firm-specific. There is no "Goldman preset" or "JPM style." Bank culture varies more by group and MD than by logo on the door, and any product claiming to simulate a specific firm's interview style is selling a fiction. Restructuring at Lazard interviews differently from M&A at Lazard. Coverage at Goldman interviews differently from financing at Goldman. The actual variance is at the group level, and no platform — ours included — can credibly claim to mirror it.

HARDO is not a course. We don't teach accounting from scratch, we don't walk you through what EBITDA is, and we don't have a "fundamentals track." Bring the fundamentals. If you can't explain why depreciation flows through three statements without us prompting you, you're not ready for HARDO yet. You're ready for a textbook. Come back when you can.

HARDO is not a guarantee. Anyone selling interview prep who tells you they can guarantee offers is lying, and you should close the tab. Offers depend on your school, your GPA, your network, your timing, your interviewer's mood, and a dozen other variables we can't touch. What we can do is make sure that when you sit down for the interview, the technicals and the behavioral story aren't the thing that breaks.

HARDO is not a Vault guide replacement. The guides are good for context — what a typical day looks like at each bank, what the deal flow has been, how groups are structured. Use them for that. Use us for the reps.

Why We Built It This Way

The honest version: mock interviews with experienced bankers are the gold standard for prep, and they always will be. A real MD doing a forty-minute mock with you and then walking through every weakness is worth more than anything an AI can produce.

The problem is that those mocks don't scale. They cost $200–500 an hour through the established coaching services, more if you want someone senior. Most candidates manage two or three before recruiting starts, sometimes fewer. That's not enough reps to actually rewire how you respond under pressure. It's enough to learn a few specific lessons and hope they generalize.

AI interviewers are not better than a live MD. We're not going to pretend otherwise. What they are is available at 11pm the night before your Citi superday when you realize you've never said your "why investment banking" answer out loud to anyone who pushed back on it. They give you fifty reps in the time it would take to schedule two with a human, and the marginal cost of the fifty-first rep is zero.

The product isn't trying to replace the live mock. It's trying to replace the thing most candidates actually do, which is read another guide, watch another YouTube video, and walk into the interview with the same untested answers they've had for three months.

Try It

One free Intern interview, no card required. If it's useful, $14.99 a month, cancel anytime. Monthly only — we don't lock people into annual plans for a product they should ideally outgrow in a few months of focused prep.

The pitch is reps, not magic. Come do some.