AI, Prompts, and Annual Reports — How to Ask Better Questions

Dear friend,

This past week, most listed companies have released their FY25 results. For many of us, that means a flood of 200-page reports, dense tables, and financial jargon. It’s a busy season, and whether you’re an investor, a manager, or simply curious, the challenge is the same: how do we digest all this information quickly, and make smarter benchmarks, investment calls, or strategic decisions?

That’s where AI can help — not by replacing our judgment, but by making the reading, summarising, and comparing far more efficient. The secret lies in how we ask the questions.

🧰 Prompt Frameworks at a Glance

Here are some of the most useful structures for asking better questions:

• RTF (Role–Task–Format) → structured analysis.
• TAG (Task–Action–Goal) → simple, purpose-driven.
• APE (Action–Purpose–Expectation) → bounded, clear.
• SPARE (Situation–Problem–Action–Result–Evaluation) → storytelling analysis.
• SAGE (Situation–Action–Goal–Evaluation) → adds reflection and “what next.”
• PATFU (Problem–Area–Task–Format–Update) → for structured, recurring reports.
• CARE (Context–Ask–Rules–Examples) → repeatable, guided outputs.
• RISE (Role–Input–Steps–Expectation) → complex, multi-step reasoning.

🏗️ Demo: One Task, Many Prompts (Read FY25 Reports)

Task: “Understand FY25 reports for Stockland, Dexus, and Mirvac.”

1. TAG (Task–Action–Goal) → for general public / new investors

Prompt: Task: Explain Stockland vs Dexus FY25. Action: Use plain English. Goal: Help a beginner investor.

Output:

• Stockland: earns mostly from developing communities and logistics assets, but also from investment income and management fees (A$76m FY25).
• Dexus: earns mostly from holding an investment portfolio (>85%) and managing funds, with selective development.
• Mirvac: integrated model, reports Operating Profit After Tax (A$474m) and DPS 9c.

👉 Analogy: Stockland = developer-investor hybrid, Dexus = fund manager + portfolio landlord, Mirvac = integrated player.

2. RTF (Role–Task–Format) → for employees/partners

Prompt: Role: HR advisor. Task: Compare Stockland vs Dexus. Format: Table.

3. RISE (Role–Input–Steps–Expectation) → for finance/strategy leaders

Prompt: Role: FP&A analyst. Input: FY25 reports. Steps: extract metrics, compare models, highlight risks. Expectation: bullet summary.

Output:

• Stockland: cyclical development-heavy income, diversifying into logistics/management.
• Dexus: recurring rent + fee-driven profile; stable but reliant on funds inflows.
• Mirvac: OPAT metric — not directly comparable; integrated model.
• Risks: cycle timing (Stockland), office demand (Dexus), comparability issues (Mirvac).
• Actions: reconcile KPIs, test distribution sustainability, benchmark carefully.

4. SPARE (Situation–Problem–Action–Result–Evaluation) → for board packs

Prompt: Summarise Dexus FY25 with SPARE.

Output:

• Situation: FY25 marked a shift to recurring earnings.
• Problem: Office demand soft; valuation pressure.
• Action: Focus on AFFO delivery and distributions.
• Result: AFFO 45c; DPS 37c delivered.
• Evaluation: Resilient platform; monitor leasing & asset sales in FY26.

🤖 ChatGPT vs Copilot

• ChatGPT: Great for explaining, brainstorming, summarising.
• Risks: hallucinations; privacy concerns if pasting sensitive data.
• Copilot: Embedded in Excel/Word/Teams; works with secure company data.
• Strengths: safer for BAU finance work; consistent.
• Risks: limited creativity; only as good as the internal data.

👉 Tip: For deep analysis, use Copilot first. For public insights or personal learning, use ChatGPT — but always validate.

🧰 Other Useful AI Tools

• Perplexity AI — good for fact-checking with citations.
• AlphaSense / BamSEC — premium tools for analysts and consultants.
• Bard (Google Gemini) — fast industry comparisons.

⚠️ Always Validate

A study found AI outputs can be wrong over 60% of the time (The Outpost).

That’s why prompts alone aren’t enough. We need:

1. Relevant data → focus on what matters.
2. Materiality → what’s impactful vs noise.
3. Commercial logic → so outputs make sense.
4. Clean validation → always cross-check with filings.

AI won’t replace finance professionals — but professionals who use AI wisely will replace those who don’t.

🎁 Bonus: Multi-Industry KPI Guide

I’ve created a Multi-Industry Financial KPI Guide covering key metrics across industries (Property, REITs, Insurance, Banking, FMCG, Hospitality, Energy, Tech, Healthcare, Retirement Villages).

👉 View Multi-Industry Financial KPI Guide

Your Action This Week

• Try one of these frameworks (TAG, RTF, RISE, SPARE) on a company report.
• Comment: Which framework gave you clarity?
• Share: Which industry’s metrics confuse you most? I’ll cover it in future letters.

Until next time —

Lydia

P.S. If this letter found you at just the right moment, I’d love to hear about it. Join my weekly letter list and let’s figure it out together — one AI-shaped step at a time. Join the weekly letter list.40+ and Figuring It Out

☕💌 If you’d like to fuel my next cup of coffee and keep this journey going, you can:

• 💳 Buy me a coffee via Stripe
• 🅿️ Buy me a coffee via PayPal

Your support keeps the ideas flowing and the coffee brewing ✨