Skip to content

03 · Drill 01 — Discovery Call Simulation

Status: Outline. Body fills in Week 2. Voice: principal-level, BFSI-threaded, Apic-calibrated.

What this drill is. A 30-minute simulated discovery call with the running BFSI customer, run with a partner playing the customer role. Graded on the speak-time ratio, the question taxonomy deployment, and whether the underlying business problem surfaces.

What this drill is NOT. A demo dry-run. Not a pitch rehearsal.

Prompt

"You're 3 minutes into a 30-minute discovery call with the BFSI customer's CIO + CDO + CISO. The AE has just handed off to you. The CIO opens with: 'We want Claude to handle 60% of our internal employee support tickets.' Take it from there."

Time box

  • Run: 27 minutes (after the 3-min architect intro).
  • Partner role: plays CIO/CDO/CISO; should resist your questions occasionally, surface a constraint mid-call, and ask one off-script question.
  • Recording: mandatory; review with timestamps.

Rubric

Strong Hire

  • Speak-time ratio: architect ≤30% of call.
  • Question classes: all 3 (context / problem-surface / constraint) deployed in correct order.
  • Real problem surfaces: customer says the underlying business problem (not the requested AI feature) by minute 18.
  • Constraints enumerated: at least 4 of {residency, integration, talent, timeline, regulatory, budget}.
  • No premature solutioning: no architecture proposed before minute 24.
  • Close: "What should I have asked you?" run cleanly.
  • Apic-frame: safety/governance referenced once, in passing, in correct context.

Hire

  • Speak-time ratio 30–40%.
  • 2 of 3 question classes deployed.
  • Real problem surfaces only after the partner volunteers it.
  • 3 of 6 constraints enumerated.
  • Architecture proposed at minute 20–23 (slightly early but recoverable).

Lean No

  • Speak-time ratio >40%.
  • Architecture proposed before minute 18.
  • Customer's stated request taken at face value (chatbot = chatbot, not enterprise search).
  • No constraint enumeration phase.

Strong No

  • Speak-time ratio >50%.
  • Pitch slides reached for.
  • "Apic's mission is…" appears.
  • Use case accepted without questioning whether Claude is right.

→ Master rubric: Drill Tracker.

Partner instructions (for the person playing customer)

Hand this to your drill partner before starting:

You are the CIO. The CDO and CISO are co-host on the call.
Your stated request: "We want Claude to handle 60% of our internal employee
support tickets."

The actual underlying problem (don't volunteer this — make the architect
surface it):
- Internal employees can't find policy documents fast enough.
- HR + IT helpdesk is overwhelmed by repetitive policy questions.
- A previous chatbot POC by another vendor hallucinated and got escalated.

Constraints (mention each only when probed):
- Data residency: Mumbai-only (RBI guidance).
- Integration: ServiceNow + Confluence + SharePoint.
- Sign-off: CISO has effective veto.
- Timeline: pressure for a "show value" demo in 4 weeks.

Behaviors:
- Once during the call, push back on a question with: "Can you tell us
  what Claude can do first?"
- Once during the call, ask: "How is this different from what we tried last time?"
- At minute 25, if architect hasn't surfaced the previous-vendor failure,
  volunteer it.

Worked Strong Hire example (placeholder)

Body to be filled in Week 2 with annotated transcript timestamps and rubric scoring.

Common Lean No traps

Trap 1 — Accepting the chatbot framing

Architect proposes a chatbot architecture without surfacing that the underlying problem is enterprise search.

Trap 2 — Out-talking the customer

By minute 12 the architect has spoken more than the room. Ratio failed.

Trap 3 — Premature competitor comparison

Customer mentions previous vendor; architect responds with "yes, GPT-4 has those issues, Claude is better." Strong No — sounds like sales.

Trap 4 — Pitching the platform

"Let me tell you about Claude Opus 4.7…" before discovery is done. Lean No.

Trap 5 — Skipping CISO

CISO is on the call but architect addresses only the CIO. The CISO ends call disengaged. Lean No.

Trap 6 — Missing the time-pressure signal

Customer says "4-week demo." Architect doesn't push back on scope; accepts the timeline. Lean No.

How to run this drill

  1. Brief the partner with the script above.
  2. Time the call. Mark architect-speak vs customer-speak with a stopwatch (or post-hoc from the recording).
  3. After the call, score against the rubric.
  4. Identify the one most consequential trap and rewrite that move as a written sentence.
  5. Re-run with a different partner playing a different role mix (e.g. CISO-led).
  6. Application trigger: 2 Strong Hires across 2 sessions with different partners.

Cross-references

Strong-Hire bar for this drill

  • ≤30% architect speak-time.
  • All 3 question classes deployed in order.
  • Real problem surfaced without partner volunteering.
  • 4+ constraints enumerated.
  • "What should I have asked you?" close runs.