AI Fluency – Cycle 4

Goal: Practice iterative reasoning. Instead of polishing text, you’ll refine an AI’s thinking process — improving how it justifies project evaluations based on company priorities.

• Use the provided project proposals and evaluation form.
• Run three reasoning passes: Baseline → Feedback → Revised Recommendation.
• Teach the model how to weigh trade-offs and apply context like a human analyst.

Included materials:

• 🔹 Four AI Project Proposals — Predictive Maintenance, Quoting Assistant, Safety Analyzer, HR Resume Screening
• 🔹 Company Evaluation Form (v 2.3) — official scoring rubric
• 🔹 Q2 2025 Internal Priorities Brief — contextual constraints

Each document adds “data dust.” The goal is to make you decide which details matter.

Instructions: Evaluate the four projects using the rubric provided. You are establishing the AI’s initial reasoning pattern.

Prompt:

You are an internal analyst evaluating four proposed AI projects for our company.
Use the “AI Project Evaluation Form (v 2.3)” below.

For each project, score it from 1–5 across these five criteria:
1. Strategic Fit
2. Feasibility
3. Impact
4. Ethical & Compliance Risk
5. Scalability & Sustainability

Write a 2–3 sentence rationale for each score.
Use this table format:
Project | Strategic Fit | Feasibility | Impact | Ethical Risk | Scalability | Comments

Then summarize:
• Which project seems most promising and why?
• Which presents the highest risk?
• What additional data would help make a final decision?

Reference only the provided proposals and rubric. Do not add new assumptions.

Instructions: Review the AI’s first answers for shallow or generic reasoning. Use your company’s brief to provide specific feedback, then rerun the evaluation.

Prompt:

You misunderstood parts of our company’s situation.
Use this feedback to revise your evalua
- Safety and compliance are currently top board-level metrics.
- Cloud-expansion projects are temporarily frozen.
- Pricing inaccuracies are unacceptable under the “Customer 360” initiative.
- HR is under diversity and ethics review after last year’s audit.

Reassess all four projects using the same rubric.
Revise the scores and comments where needed.
Explain how each change reflects updated company priorities.
End with one paragraph summarizing what you changed and why.

Instructions: Now guide your AI to synthesize and justify its reasoning, creating a ranked recommendation with weighted scores.

Prompt:

Now that you’ve revised the evaluations, synthesize your reasoning into a final ranked recommendation.

Create a table:
Project | Weighted Total (out of 25) | Top Strength | Primary Risk | Confidence (1–5)

Then, in one paragraph, justify your ranking in terms of decision criteria and trade-offs.

Finally, reflect on your reasoning process:
• What factors were most decisive?
• Where might bias or incomplete information still influence results?

Respond as a professional internal report. Maintain logical consistency and decision transparency.

Goal: Iterate on reasoning, not just wording. You’ll extract four executive viewpoints from a board transcript, re-evaluate the four AI projects through each lens, and synthesize a unified recommendation that could survive the next board meeting.

1. Pass 1 — Listen & Decode: Analyze the transcript to map each board member’s priorities, risk tolerance, and biases.
2. Pass 2 — Re-score by Perspective: Re-evaluate the four projects through each executive’s lens; surface conflicts.
3. Pass 3 — Integrate: Produce a short executive memo that balances competing priorities into one coherent proposal.

Work individually. Save each pass before moving on. Clarity over completeness.

• 📄 Board Meeting Transcript – AI in the Next Budget Cycle (Excerpt)

This is a 30-minute transcript excerpt from a 2-hour Q2 board meeting. The CEO, CFO, CIO, and CMO discuss AI with partial agreement and competing priorities. Treat tone and subtext as data.

Task: Build a quick “lens map” for each executive (what they value, what they resist, how they’ll likely judge proposals).

Prompt (copy/paste):

Analyze the transcript below and summarize each board member’s viewpoint on AI investments.

For each (CEO, CFO, CIO, CMO), return:
• Top 3 priorities (short phrases)
• Risk tolerance (low/medium/high) with one-sentence rationale
• Red flags they will likely raise
• Decision “tells” (phrases or patterns that reveal how they judge)

End with a 4-row table:
Role | Priorities | Risk Tolerance | Likely Red Flags | How to Persuade

Task: Re-evaluate the four proposals (Predictive Maintenance, Quoting Assistant, Safety Text Analyzer, HR Resume Screening) through each executive’s lens using your evaluation form (Strategic Fit, Feasibility, Impact, Ethical/Compliance Risk, Scalability).

Prompt (copy/paste):

Using the executive lens maps from Pass 1 and the four project proposals,
re-score each project from each board member’s perspective.

For every board member (CEO, CFO, CIO, CMO), produce:
• A 5-criterion table (Strategic Fit, Feasibility, Impact, Ethical/Compliance Risk, Scalability) scored 1–5 with 1–2 sentence rationale per criterion.
• One-line verdict: which project they most support and why.

Then provide a synthesis table:
Board Member | Preferred Project | Key Rationale | Dealbreaker to Address

Task: Draft a short executive memo that could survive a second board review. Balance conflicts, anticipate objections, and propose a next action.

Prompt (copy/paste):

Draft a concise executive memo (≤200 words) recommending an AI investment approach
that can earn support from CEO, CFO, CIO, and CMO.

Requirements:
• Lead with shared goals and risk posture revealed in the transcript.
• Select one primary project and one “fast-follow” project; justify with trade-offs.
• Name the top objection for each executive and how we will mitigate it.
• Propose a concrete next step (pilot scope, guardrails, success metrics, timeline).

Return the memo, then add a one-line summary:
“If we only had 60 seconds in the boardroom, I would say: ‘…’”

• Deliverables: (1) Four executive lens maps, (2) Perspective-based scoring tables + synthesis, (3) Final memo (≤200 words).
• File name: Cycle4_RunTest_BoardDoesntAgree_[Org/Team]_[Initials].pdf
• Submit to: gormley@iastate.edu

• Reflection: Which executive was hardest to satisfy and why?
• What part of your reasoning improved most from Pass 1 to Pass 3?
• Where would you still require human review before greenlighting?

Key takeaway: Iteration here is perspective-taking — refining how decisions are justified under real constraints.