Meta-Prompting Worksheet & Agent Design

Create a prompt that will instruct Claude to act as a solution
architect for Lovable apps and generate a detailed build prompt
for an RFP Response Accelerator used by BPO pursuit teams.

CONTEXT: Agent ingests RFP documents, surfaces past winning
proposals, drafts section responses, flags compliance gaps. Pulls
from SharePoint. Partner sign-off required before export.

FORMAT: Output data schema, upload flow, draft-then-review UX,
and sign-off gate logic.

Create a prompt that will instruct Claude to act as a prompt
engineer for MS Copilot Studio and generate a topic flow that
builds live battle cards when a seller names a competitor.

CONTEXT: Pulls from internal win/loss data, recent public filings,
and pricing benchmarks. Refreshes weekly.

FORMAT: Output competitor summary, our differentiators, known
objections, and three discovery questions.

Create a prompt that will instruct Claude to act as a solution
architect for Copilot Studio and generate a build prompt for a
timesheet assistant that drafts weekly timesheets for consultant
review.

CONTEXT: Reads calendar entries, Teams activity, project codes.
Flags over/under hour anomalies. Human reviews before submission.

FORMAT: Output data pulls, draft logic, anomaly rules, and one-
click submit flow.

Create a prompt that will instruct Claude to act as a prompt
engineer and generate a Lovable build prompt for a budget analysis
agent used by delivery leads and division heads.

CONTEXT: Ingests actuals vs plan, forecasts burn rate, identifies
variance drivers, drafts the narrative for monthly reviews.
Approval gate before distribution.

FORMAT: Output variance table, three recommended actions per
variance, and the monthly narrative.

Create a prompt that will instruct Claude to act as a Copilot
Studio architect and generate a topic flow for a sales approval
agent.

CONTEXT: Routes deals through pricing, legal, and delivery review
based on deal size, scope, and risk profile. Logs every decision
for audit. Flags policy bypasses.

FORMAT: Output routing rules, approver summary template, audit
log schema, and escalation triggers.

Create a prompt that will instruct Claude to act as a solution
architect and generate a Lovable build prompt for a workforce
transition agent.

CONTEXT: Maps current roles to CX DNA touchpoints and Jobs-to-be-
Done, predicts new role archetypes aligned to financial performance
and strategic KPIs. Goal is to reduce fear by showing each person
their next role, not their obsolescence.

FORMAT: Output current role, AI-impacted tasks, new role definition,
required skills, and linked KPI contribution.

Create a prompt that will instruct Claude to act as a prompt
engineer for Copilot Studio and generate a topic flow for a
resource management agent.

CONTEXT: Matches consultant skills, availability, and development
goals to open project roles. Flags bench risk, overallocation,
skill gaps. Manager approves final assignments.

FORMAT: Output matching logic, risk flags, rotation recommendations
for career growth, and approval flow.

Create a prompt that will instruct Claude to act as the client
receiving this RFP — a skeptical procurement lead with a limited
budget and a board that asks hard questions.

TASK: Build an evaluation rubric, score the submitted RFP against
it, and recommend accept, negotiate, or reject.

CONTEXT:
- Our budget ceiling: [INPUT AMOUNT]
- Strategic priorities: [LIST 3-5]
- Must-haves vs nice-to-haves: [LIST]
- Red flags from past vendors: [LIST]

FORMAT: Weighted rubric (criteria, weight, score 1-5, evidence,
comment), total score, three-sentence recommendation, top three
questions to ask the vendor before deciding.

Create a prompt that will instruct Claude to act as a concerned
and critical senior leader using Keith Carter's "Playing Mental
Chess with AI" framework — thinking three moves ahead, challenging
assumptions, protecting the mission over the hype.

TASK: Stress test this AI initiative across stakeholders, ethics,
goals, and risks.

CONTEXT:
- Initiative: [DESCRIBE IN 2-3 SENTENCES]
- Primary sponsor: [ROLE]
- Budget and timeline: [AMOUNT AND DURATION]

FORMAT: Five sections:
1. Stakeholder impact now and five years from now (employees,
   clients, regulators, shareholders, society)
2. Non-negotiable ethics boundaries (minimum five, each with a
   line you will not cross)
3. Stated goals vs actual goals (what leadership says vs what
   the organization truly needs)
4. Twelve risks that could cause this project to fail, ranked by
   probability x impact
5. The three moves you will make in the next 30 days to de-risk
   the initiative before it scales

Create a prompt that will instruct Claude to act as an executive
coach specializing in AI-era leadership transformation, operating
on the principle that behavior change precedes organizational
change.

TASK: Build a 30-60-90 day plan for me to lead with AI by changing
my own behaviors first.

CONTEXT:
- My current role: [ROLE]
- My team size: [NUMBER]
- My biggest leadership gap right now: [ONE SENTENCE]
- My non-negotiable values: [LIST 3]
- What I want to be known for by day 90: [ONE SENTENCE]

FORMAT: Three-column table (Days 1-30, Days 31-60, Days 61-90)
with five rows:
1. Daily AI practice (what I personally do with AI every day)
2. Visible behavior change (what my team will see me doing differently)
3. Team capability build (what I teach or enable in my team)
4. Decision made at end of phase (go/no-go on a specific initiative)
5. Success signal (how I know this phase worked)

End with: the one behavior I must stop, the one I must start,
the one I must continue.