Most candidates treat an "AI coach" as a chat window that grades a typed answer and says you did fine. That is not how interviewers score you, and it is not what I built. CoachNed's AI coach is a live voice case with Ned: you answer out loud, he follows up, and when the session ends you get a seven-score debrief that quotes what you said, names the leak, and points you at one drill. After this piece you will know what exists on the product, how a session actually runs, what those seven scores mean, and what the tool cannot do for you.
I am Edward "Ned" Calder. I built CoachNed. The voice in the room is simulated in my cadence; the grading logic is the one I use when I score cases for real. CoachNed is independent practice, not a firm product.
What it is (and what it is not)
The coach is two linked surfaces, not a floating chatbot:
- The live interview. On /interview, Ned runs CopperCup Coffee Margin Case out loud. About eighteen minutes. He reads the brief, asks, waits, and follows up. You speak, or type any turn you would rather not say. End early and you still get a debrief on what you covered.
- The debrief. After the case, you get seven dimension scores, short reasons tied to your words, the weakest leak named, and one drill assigned. Pricing on CoachNed describes that stack plainly: seven scores, quotes from what you said, the leak named, one drill assigned (CoachNed pricing).
What it is not: a human partner on a Zoom call, a guarantee of an offer, or a secret firm rubric. McKinsey, BCG, and Bain publish expectations for case interviews; none of them publish CoachNed's score sheet. The firms ask you to structure ambiguous problems, work data, calculate, and recommend clearly — see McKinsey interviewing, BCG case interview preparation, and Bain case interview preparation. My product maps those public expectations into seven practice dimensions so your feedback is specific enough to act on.
How a session runs
A typical first live rep looks like this.
You open /interview. CopperCup is the default voice case: a coffee-chain margin problem, firm-style BCG, easy difficulty, about eighteen minutes. Ned speaks the prompt. Numbers that matter stay on the table; you do not have to memorize the exhibit.
You answer the opening structure out loud. If you ramble, he will wait or nudge; if you freeze, ending early still produces a partial debrief rather than a fake full-case grade. Mid-case you may get an exhibit or a math ask. You can type a single turn if the mic fails or you need the number written. When you finish — or stop — the session closes into the debrief.
If you only need a taste before committing to eighteen minutes, /start is the free five-minute first rep: three typed turns of a case and instant scores, no account required. That is a warm-up, not a substitute for the voice interview.
The library behind the live room is 64 original cases. Pricing notes 63 of them as voice-ready; the rest still work in guided text. Model answers sit on every case once you are inside a plan. Between full cases sit 550+ drills across structure, math, exhibits, sizing, synthesis, and brainstorming, plus Casey, CCA, and PEI simulators for the assessment stack.
What the seven scores cover
Every full debrief grades the same seven dimensions. The labels in the product are:
| Dimension | What I am listening for |
|---|---|
| Structuring | A MECE tree that fits this prompt, not a memorized framework dump |
| Hypothesis | A testable lean early, with what would kill it |
| Math | Equation first, units, then a business so-what |
| Charts / exhibits | The swing in dollars or share, tied back to the decision |
| Synthesis | Recommendation first, supports, risk, next step |
| Communication | Clarity, pacing, recovery under follow-ups (voice surfaces this) |
| Business judgment | Whether a partner would actually run with the move |
Those seven are CoachNed's practice rubric. They are not "McKinsey's official scores" or anyone else's. Firms describe overlapping behaviors — BCG lists structuring, questions, data, calculations, clear communication, and business intuition (BCG careers); Bain stresses clarifying the objective, structuring, thinking aloud, and adapting to cues (Bain case prep); McKinsey frames the problem-solving interview as discussing a typical client scenario and evaluating analytical approach (McKinsey interviewing). I collapsed that into seven lines so a debrief can name one leak instead of handing you a vague "be more structured."
The debrief is supposed to quote you. The useful artifact is not the overall number; it is the line you said when the structure overlapped, when the math had no so-what, or when the close never answered the decision. Then one drill path: structure, brainstorming (for hypothesis habits), math, charts, market sizing, or synthesis.
Ned's rule. Treat the lowest dimension with a specific quote as the only assignment for the next twenty-four hours. If you try to fix all seven after every case, you will fix none of them by first round.
Worked example: one leak, one drill (not a product screenshot)
Invented candidate answers, same CopperCup-style margin prompt. I am not claiming these numbers came from a live CoachNed run; they illustrate how I want you to read a debrief.
Prompt fragment. CopperCup's contribution margin fell from 28% to 22% while revenue was roughly flat. What would you look at first?
Weak open (paraphrased). "I would look at the 4Ps, competitors, customers, and costs, and also brand and operations and digital."
What the sheet hears: overlapping branches, no prioritization, no link to a margin bridge. Structuring and hypothesis both leak. The drill is not "do another full case tomorrow." It is ten minutes on /drills/structure forcing a two-level tree with named, non-overlapping branches for a margin drop.
Stronger open. "I would split the six-point margin drop into price realization, mix, and unit cost, then ask which cell moved most in dollars. My working lean is mix or cost, because revenue flat with margin down is rarely pure volume. I would kill that lean if like-for-like price fell hard on the same SKUs."
What the sheet hears: a decision-shaped tree, a testable lean, a kill test. Math and exhibits still have to earn their scores later. The communication score will depend on whether you said that cleanly out loud or buried it in filler.
Reproduce the habit without the product if you want: write both opens for any margin prompt you invent, time yourself to ninety seconds spoken, and mark which open a partner could steal onto a whiteboard.
What it cannot do
Honesty matters more than marketing copy here.
- It cannot replace a human final-round coach for presence. A former interviewer still wins on reading the room, firm-specific warmth, and the last-mile calibration under partner pressure. Pricing on CoachNed says the same trade: use Ned for volume, a human for the one or two sessions where presence matters (pricing).
- It cannot score a dimension you never reached. End at turn two and you get a partial debrief, not a fictional full scorecard.
- It cannot invent firm secrets. There is no confidential McKinsey grid inside the model. Public careers pages are the ceiling of what we claim about firm process.
- It cannot make typed-only practice grade spoken communication the way voice does. Communication is in the seven; voice is where pacing and filler become visible.
- It cannot guarantee that a high practice score equals an offer. Practice scores measure consistency against this rubric. Real interviews add politics, fatigue, and interviewer variance.
If you need forty reps before a first round, the tool is built for that. If you need someone to tell you whether your handshake and eye contact survive a partner lunch, hire a human for an hour.
Access, plans, and what stays free
Everything on CoachNed is open for seven days with no card: live interviews, seven-score debriefs, drills, Casey, CCA, PEI, resume report. After the week, Quick math, sample walkthroughs, the AI lesson, and the first two courses stay free; your history and debriefs stay on the account either way (pricing FAQ).
Then you pick volume, not a cafeteria of packs:
| Plan | Price | Fit |
|---|---|---|
| Free week | $0, no card | Prove the debrief before you pay |
| Season Pass | $120 once, 90 days | Interviews in the next three months |
| Pro | $49 / month | Near-term first round or open-ended |
| Annual | $249 / year | Internship and full-time across cycles |
Fourteen-day refund, no questions asked. Every paid plan is the whole stack — not a drip of cases. That includes the live voice interviews, the 64-case library, 550+ drills, and the Casey / CCA / PEI simulators.
Practice this today
- Run /start once cold. Three typed turns, instant scores. Note which dimension felt weakest without reading the labels first.
- Same day, sit the live CopperCup interview on /interview. Answer the open out loud even if the rest of the case is messy.
- Read only the lowest score's quote. Run the assigned drill for twenty minutes. Do not open a second full case until that drill is done.
If you want guided text before voice, pick any case from /consulting-cases and use guided mode; the seven-score debrief still applies. For assessment-shaped reps after that, use the Casey and CCA simulators and /behavioral for PEI follow-ups.
CoachNed is not affiliated with McKinsey, BCG, Bain, Deloitte, or any firm named in a case prompt.
Frequently asked questions
What is CoachNed's AI coach?
It is the live voice case interview with Ned plus the seven-score debrief that follows: quotes from what you said, the weakest leak named, and one drill assigned. I built it as practice volume with consistent grading, not as a chat toy.
How long is the live interview?
The default CopperCup voice case on /interview is about eighteen minutes. You can end early and still receive a debrief on the turns you completed.
Is CoachNed free?
Everything is open for seven days with no card. After that, Season Pass is $120 for ninety days or Pro is $49 a month, with a fourteen-day no-questions refund. /start remains the free five-minute first rep without an account.
Can the AI coach replace a human coach?
For volume and consistent dimension feedback, yes enough to sit dozens of cases. For presence, firm-specific cadence under a real partner, and final-round confidence, no. Most candidates should do both: many reps with Ned, then one or two human sessions before the real round.
What does the seven-score debrief grade?
Structuring, hypothesis, math, charts/exhibits, synthesis, communication, and business judgment. Those are CoachNed practice dimensions aligned with publicly described firm expectations, not an official firm score sheet.
Sources
- Interviewing at McKinsey — problem-solving interview framed as a client scenario; analytical approach and sample cases. Checked 2026-09-24.
- BCG case interview preparation — structure, questions, data, calculations, communication, business intuition; no single right answer. Checked 2026-09-24.
- Bain case interview preparation — clarify objective, structure, think aloud, listen and adapt; sample cases. Checked 2026-09-24.