FDE Simulation II — Full Cycle
- Run a complete FDE engagement solo and timed: discovery → requirements → proposal → prototype plan → demo script
- Answer two written stakeholder objections without over-promising or capitulating
- Handle a mid-course scope change with an impact analysis and options, not a silent yes or a flat no
- Self-grade honestly against the FDE rubric and identify your weakest dimension for tomorrow
| Spaced-rep warm-up + rules of the game | 10 min |
| Milestone 1: brief + requirements (timed) | 25 min |
| Milestone 2: proposal + demo plan (timed) | 35 min |
| Milestone 3: objections + change request (timed) | 35 min |
| Practice: self-grade + quiz | 15 min |
Builds on: Day 168 — FDE Simulation I · Day 171 — Stakeholders & trade-offs · Day 173 — ROI & cost analysis
Flight schools end with a check ride: a full flight — preflight, takeoff, navigation, and the part that actually decides your license: the examiner pulls the throttle mid-flight and says "your engine just died." Nobody fails a check ride for imperfect straight-and-level flying. They fail for panicking when the plan changes.
Simulation I (Day 168) was the first half of the flight — discovery to proposal, with the air calm. Today the examiner rides along the whole way: a new prospect, the full sequence of deliverables, and then the throttle-pull — two stakeholders push back in writing, and midway through, the customer changes what they want. Real engagements are exactly this: the proposal is never the end, it is the thing people start arguing with. You are graded on the same thing as the pilot: not perfection, but composure with a method — impact analysis before commitment, options instead of yes/no, and deliverables that stay coherent after the turbulence.
FDE and solutions-engineering interviews are converging on exactly this format: a take-home or live simulation with a scenario, an objection, and a curveball — because it predicts job performance better than any whiteboard. More importantly, the first real engagement of your career will go exactly like today: the CISO email, the CFO skepticism, and the scope change WILL arrive, usually in the same week. Having produced this document package once under a timer means you will produce it the second time — with a real logo on it — without the panic.
Guided practice
Milestone 1 — The brief and discovery notes (read, then scope)
25 minStart your timer. Read the scenario, then produce a one-page problem statement + requirements doc (Day 164's template: functional, non-functional, constraints, out-of-scope, acceptance criteria).
CUSTOMER BRIEF — Northwind Freight International freight forwarder. 2,800 employees, 14 countries. Azure/Microsoft shop, conservative IT, one prior AI pilot (a chatbot) that "embarrassed everyone." Sponsor: Dana Okafor, VP Ops. Economic buyer: CFO. Wary party: CISO (the chatbot pilot leaked a rate sheet to a customer). From the discovery call (your notes): - Ops staff answer internal questions all day: customs documentation rules, carrier SLA terms, HS-code procedures, internal SOPs. "Our senior people are human search engines. New hires take 9 months to stop asking." - Knowledge lives in: SharePoint (SOPs, ~12k docs), a carrier-contract folder (PDFs, some scanned), and Maria in Rotterdam, who is retiring in March. - Dana's stated goal: "Cut time-to-answer for ops questions from ~20 minutes to under 2. New-hire ramp is the board-level story." - Volume guess: ~600 ops staff, "dozens of questions a day each, honestly." - Constraint from IT (relayed): nothing leaves the Azure tenant; Entra ID SSO mandatory; the CISO will personally review anything before it touches SOPs. - Prior-pilot scar tissue: any wrong answer shown to a customer is a fireable-offense-level fear. Internal-only is acceptable for v0. - Timeline pressure: Dana wants something visible before the April board meeting (10 weeks out).
Scoping notes: decide what v0 is (internal docs-QA over SharePoint SOPs + contracts, VPC/tenant deployment, SSO) and what is explicitly OUT (anything customer-facing, live shipment data, the scanned PDFs if OCR risk is high — your call, but write it down). Every requirement must trace to a line in the notes — invent nothing Dana didn't say.
Milestone 2 — Proposal, prototype plan & demo script
35 minProduce the proposal (2–3 pages max) and the demo plan.
- Approach & architecture: C4-context in text — Northwind users, Entra ID, SharePoint, your service in THEIR Azure tenant (Day 170's matrix says which model access path; justify it to a CISO who watched a chatbot leak a rate sheet). Name the guardrails and permission-aware retrieval explicitly (Days 132, 169) — this customer's scar tissue demands it in the proposal, not the appendix.
- Scope ladder: v0 (10 weeks, board-demo-able), v1, later. The out-of-scope list from Milestone 1 appears here verbatim.
- Plan with ranges: phases with week ranges, and the calendar-vs-engineering split (Day 169) — the Entra app registration and CISO review are on the critical path, so they start week 1.
- Risks: top 4 with mitigations. (If "scanned PDFs OCR poorly" and "adoption fails because Maria answers faster" aren't on your list, look harder.)
- Commercials: adapt yesterday's cost model to 600 ops users; pilot with pre-agreed metrics (time-to-answer sampled before/after, weekly active %, groundedness on a Northwind golden set).
- Demo script: 10-minute board-meeting demo, Day 167's arc, with the two prepared failure recoveries.
Milestone 3 — Objections and the throttle-pull (open only after M2)
35 minThree artifacts just landed. Respond to each in writing. Objection replies ≤ 250 words each; the change-request response is an impact analysis + options memo (≤ 1 page).
ARTIFACT 1 — Email from Priya Nair, CISO Subject: RE: Proposal — security concerns I'll be direct. The last "AI assistant" this company piloted sent a confidential rate sheet to a customer. I have read your proposal. Questions: (1) You say data stays in our tenant, but the model API is external. Which is it? Exactly what leaves, in what form, and is it retained or trained on? (2) Your assistant reads all of SharePoint with a service account. Our SharePoint permissions are, frankly, a mess. What stops an ops trainee from surfacing an M&A document a VP left in the wrong folder? (3) I want the audit log spec — every question, every document cited, every user — before I approve anything. Retention 7 years. I am not opposed. I am unconvinced. — P.
ARTIFACT 2 — Email from Marcus Webb, CFO (forwarded by Dana: "help?") Subject: FW: Proposal — the numbers Dana, your vendor quotes ~USD 9k/month all-in plus a 10-week build. Ops answered these questions for free last year. The "time saved" math assumes 600 people save 15 min/day, which I frankly do not believe — nobody returns saved minutes to the company. And Maria answered questions for 30 years without a line item. What am I actually buying, what happens if adoption is half the assumption, and why not the USD 6/seat tool IT found online?
ARTIFACT 3 — Change request from Dana Okafor (week 4 of the engagement, prototype in progress) Subject: One change before the board meeting Leadership loved the preview. One change: they want the CUSTOMER-FACING support team using it for the board demo — answering customer emails about shipment status and customs delays. Same corpus, right? IT says the shipment-tracking database has an API. Timeline stays: board meeting in 6 weeks. This is the difference between a pilot and a company-wide rollout, so I need a yes here. — D.
Response guidance — hold your ground where the ground is real: Priya gets exact data-flow facts (Day 169's addendum answers 1 and 3; permission-aware retrieval answers 2 — and concede honestly that permission hygiene is a shared risk, proposing the pre-launch ACL audit). Marcus gets the worst-case sensitivity you already computed, the deflection number (which does not depend on "returning minutes"), the buy-option comparison done straight, and the pilot as the honest "if I'm wrong, you find out cheap" offer. Dana's change request is the trap: customer-facing + live operational data + the SAME 6 weeks is the chatbot incident waiting to recur — her own board's biggest fear. Impact-analyze (new risk class, new data integration, new evals, the guardrail bar jumping from internal to external), then offer options: e.g. (A) board demo shows customer-facing flow on FROZEN data, clearly labeled, internal rollout ships as planned; (B) full customer-facing adds N weeks after v0; (C) scope swap. Recommend one, tied to HER stated fear of public wrong answers.
On your own
Self-grade against the seven dimensions
15 minTimer off. Grade the whole package against the seven-dimension rubric (discovery fidelity, scoping discipline, architecture soundness, communication register, commercial credibility, objection handling, change management), 1–4 each, with one sentence of evidence per score — quote your own deliverable.
Constraints: at least one dimension must score ≤ 2 (if everything is a 4, re-read Milestone 3 asking "would Priya actually be convinced?"); write down the single weakest dimension and WHY — Day 175 starts there. Then diff your package against Day 168's Simulation I output: name two concrete ways today's is better and one way it is still weak.
The engagement package
Assemble everything into portfolio/fde-sim-2/ in your practice repo: 01-requirements.md, 02-proposal.md (with demo script), 03-objection-ciso.md, 04-objection-cfo.md, 05-change-request-response.md, and 06-self-grade.md. This directory is a portfolio artifact — recruiters and interviewers can read an entire simulated engagement end-to-end, which is rarer and more convincing than another to-do app. Commit with a README fronting the scenario in three sentences.
Common mistakes & misconceptions
- Reusing Meridian's documents with the names swapped. Discovery fidelity is the first rubric dimension: Northwind's scar tissue (the leaked rate sheet), its retiring expert, and its board deadline must visibly shape YOUR scope and prose.
- Answering the CISO with reassurance instead of facts. "We take security very seriously" is what the last vendor said; exact data flows, retention terms, and the ACL-audit proposal are what changes her mind.
- Fighting the CFO on his own simplification. "Nobody returns saved minutes" is partially TRUE — concede it, then stand on deflection and the pilot, which don't depend on it.
- Saying yes to Dana because she said "I need a yes here." Sponsors under board pressure ask for exactly the thing that would burn them; the loyal move is the impact analysis and the labeled-demo option, not obedience.
- Blowing the timebox polishing Milestone 2. Real engagements ration attention; a good-enough proposal with strong objection replies beats a perfect proposal with rushed ones — and the rubric weights them equally.
- Grading yourself 4s. The self-grade is calibration training for Day 175 and for real retros; an all-4 grade means the grader failed, not the work succeeded.
Q1. The CISO asks "what exactly leaves our tenant?" The winning response style is:
Q2. Dana's change request (customer-facing + live shipment data, same timeline) is dangerous primarily because:
Q3. Your worst-case sensitivity (half adoption, half minutes-saved) shows the project barely breaks even. In the CFO reply you should:
Go deeper — curated resources
- Run Milestone 3 live — Have a friend (or the AI tutor in adversary mode) play Priya and Marcus on a call and push back on your written replies in real time — the spoken version is the interview format.
- All six package files committed under portfolio/fde-sim-2/
- Milestone 3 responses written without breaking the ≤ 250-word / 1-page limits
- Self-grade contains evidence quotes and one score ≤ 2
- Weakest dimension named in writing for tomorrow; quiz ≥ 2/3
← Back: This is Day 168's simulation with the second act added: Day 171's no-with-options carried the change request, Day 173's sensitivity number carried the CFO, and Day 169's addendum carried the CISO. Day 166's labeled-prototype honesty is what makes option A ethical.
Forward →: Tomorrow (Day 175) you rewrite today's weakest deliverable and fold the lessons into the capstone punch list. On Day 179, Northwind becomes an interview story; on Day 180, this package is a portfolio exhibit.