AI FINANCE BRIEFHow finance pros actually use AI
WORKFLOWS PACK · v1 · 2026
The AI Finance Workflows Pack
7 paste-ready workflows you can run at your desk this week — the exact prompts, the pro move, and the compliance guardrail for each. Model-agnostic (Claude, ChatGPT, Copilot, Gemini).
How to use this pack. Each workflow is a copy-paste prompt with the parts you swap out in [BRACKETS]. Paste the transcript / filing / model / notes into the same chat, then send the prompt. Everything here is educational — it turns AI into a faster junior analyst, not a decision-maker. Read every output critically before it leaves your desk.
1The 4-minute earnings-call teardown
Analysts · PMs
The problem: 500+ transcripts in a 6-week season; you read 20–30 and miss the rest.
You are a buy-side equity analyst. Here is the earnings-call transcript for [TICKER], quarter [Q/FY].
Produce a one-page teardown:
1. Forward guidance changes vs the prior quarter (raised / cut / reaffirmed) with the exact quote.
2. Capex & R&D spend signals and any change in tone about them.
3. Management tone shift vs last call (defensive / confident / evasive) with evidence.
4. Anything management repeated or dodged.
5. Score divergence from current sell-side consensus 1–10 and say why.
6. Flag any language that historically precedes a guidance cut.
Output as a bulleted brief. Quote the transcript for each claim. If a section has no signal, say "no change."
Pro move: keep a saved "house style" line — "score divergence vs consensus, quote everything" — so every teardown is comparable across names.
Guardrail: only paste PUBLIC transcripts. Never paste MNPI, drafts, or internal notes into a consumer AI tool.
2DCF & model stress-test
Analysts · IB · Corp-dev
The problem: your model looks clean but you can't see which assumption is doing all the work.
Here is my DCF for [COMPANY] (assumptions pasted below).
1. List every assumption and rank them by how much the valuation moves per +/-1 unit change (sensitivity).
2. Flag the 3 assumptions where I'm most likely being optimistic vs history / peers.
3. Give a bear, base, and bull revenue-growth & margin set with the resulting fair value each.
4. Name one thing a skeptical IC member would attack first, and how I'd defend it.
Show your reasoning. Do not invent numbers I didn't give you — if you need a figure, ask.
Pro move: add "do not invent numbers — ask" to any modeling prompt. It's the single line that stops hallucinated inputs.
Guardrail: sanitize client/deal names to a codename before pasting a live deal model.
3IC memo / board deck first draft
PMs · Corp-dev · Strategy
The problem: the analysis is done; you lose an afternoon formatting it into a memo.
Turn the notes below into a 1-page investment-committee memo for [NAME].
Sections: Thesis (2 sentences) · What has to be true · Key risks & mitigants · Valuation range & catalysts · Recommendation & size.
Tone: senior, decisive, no hedging filler. Every claim tied to a number in my notes.
Then give me a 5-slide deck outline (title + 3 bullets per slide) covering the same.
Pro move: paste your firm's last approved memo first and say "match this structure and voice." Instant house-style.
Guardrail: AI drafts the structure — you own every number and the recommendation. Fact-check before it circulates.
4Macro synthesis in 5 minutes
Macro · Multi-asset · Wealth
The problem: Fed minutes, CPI, jobs, and 6 strategist notes — and a client asking "so what?"
Below are today's macro inputs (releases + notes pasted).
1. What actually changed vs last week, in 3 bullets — signal only, no recap.
2. Where do these inputs AGREE and where do they CONFLICT?
3. Give the base-case regime read (growth + inflation direction) and the one data point that would break it.
4. Write a 3-sentence "what this means for your portfolio" note in plain English for a client.
Be explicit about uncertainty. Separate what's known from what's a forecast.
Pro move: the "where do they conflict" question is where the edge is — consensus is priced; disagreement isn't.
Guardrail: the client-facing sentence is a draft. Compliance-review before it's sent externally.
55,000 tickers → 20 ideas
Analysts · Quant · Screening
The problem: screens spit out 400 names and you triage them by hand.
Here is a screen output (CSV pasted): [columns].
1. Cluster these names into 4–6 themes and label each.
2. Within each cluster, surface the 2–3 most interesting on the metrics I care about: [e.g. FCF yield, revenue accel, insider buys].
3. For each shortlisted name, one sentence on why it stands out and one risk.
4. Flag any name that looks like a data error or a value trap.
Give me a ranked 20-name shortlist with a one-line rationale each.
Pro move: ask it to "flag value traps" — the model is good at catching the too-cheap-for-a-reason names a fast screen hides.
Guardrail: the shortlist is a starting point for YOUR diligence, not a buy list.
6Instant comps & peer read
Analysts · IB
The problem: building a comp set and spotting the outlier eats an hour.
Here is a comp table for [COMPANY] and its peers (pasted below).
1. Is [COMPANY] cheap or rich vs the set, and on which multiple specifically?
2. Explain the gap: growth, margins, balance sheet, or sentiment? Use the numbers.
3. Which peer is the fairest comparable and which one is skewing the median?
4. If the market re-rated it to the peer median, what's the implied upside/downside?
No hand-waving — tie every point to a figure in the table.
Pro move: ask "which peer is skewing the median" — one bad comp quietly distorts most peer analyses.
Guardrail: verify the pasted multiples against your data source; AI won't catch a stale figure.
7The weekly client / desk update
Wealth · Advisors · IR
The problem: the same "what happened this week" note, rewritten from scratch every Friday.
Write this week's client update from the bullets below.
Audience: [e.g. HNW clients, non-technical]. Length: [200 words].
Tone: calm, plain-English, zero jargon, no alarmism. Lead with what matters to THEIR portfolio.
End with one forward-looking sentence and a soft "reach out with questions" line.
Give me 2 subject-line options, each under 45 characters.
Pro move: save one gold-standard past update and prepend "match this voice." The output stops sounding like AI immediately.
Guardrail: personalized financial communication is regulated — review before send, keep the audit trail.
This is one issue's worth. We ship a new one every week.
AI Finance Brief is the hands-on how-to for finance pros — tested prompts, workflows, and tools, free every week.
Get the free weekly brief →
This content was generated with AI assistance and reviewed by a human editor. For informational and educational purposes only — not financial advice. Verify AI output before acting. Do your own research (DYOR).