AI Prompts

Tested prompts for research, memos and modelling. Copy, adapt and run.

18 prompts · Structured for institutional tasks

What makes a finance prompt effective

Output quality reflects input quality. A prompt that produces a useful result for an investment task has a recognizable structure.

  1. 01

    Role

    “You are a senior equity analyst”

  2. 02

    Task

    “Build a one-page primer”

  3. 03

    Context

    Company, sector, time period

  4. 04

    Data

    The actual numbers to work from

  5. 05

    Format

    Table, bullet points, memo

  6. 06

    Constraints

    Word count, tone, audience

Browse all 18 prompts.

Filter by category or search by task. Each card includes a structured prompt you can copy and adapt for your workflow. The prompt engineering guide explains why these patterns work.

Company Research

3 prompts.

Prompt 01
You are an equity research analyst. Build a one-page company primer for [COMPANY NAME] covering business model, revenue segments, key geographies, competitive position, and the three financial metrics that matter most for this business. Keep each section to 3-4 sentences. Flag any areas where public disclosure is thin or where the business model is hard to assess from filings alone.
Company Research

Company Primer Builder

When you need a structured one-page overview of a company you are starting to cover.

Prompt 02
Map the competitive positioning for [COMPANY NAME]. Identify the top 5 direct competitors, their estimated market share, and the one structural advantage each holds. For [COMPANY NAME], identify the two biggest competitive threats and the one moat that is hardest to replicate. Explain your reasoning for each.
Company Research

Competitive Positioning Map

When you need to understand who the real competitors are and where the structural advantages sit.

Prompt 03
Summarize the leadership team and governance structure at [COMPANY NAME]. Cover CEO tenure and track record, board composition, insider ownership levels, and any recent executive departures or appointments. Flag governance concerns that would matter to an institutional investor, including related-party transactions, dual-class share structures, or compensation plans that seem misaligned with shareholder returns.
Company Research

Management & Governance Assessment

When you are assessing the quality and alignment of leadership before taking a position.

Earnings

3 prompts.

Prompt 04
Analyze the following earnings call transcript for [COMPANY NAME]. Extract the three most important management commentary points, changes in guidance, and any tone differences between the prepared remarks and the Q&A. Note where management was evasive or where the tone was notably different from prior quarters. Flag any question where the answer felt rehearsed or incomplete.
[INSERT TRANSCRIPT]
Earnings

Earnings Call Synthesis

When you need to extract the signal from a 60-minute earnings call in under five minutes.

Prompt 05
Using the following consensus estimates and management guidance for [COMPANY NAME], identify where guidance diverges from sell-side expectations. For each divergence, classify it as above, in line, or below consensus. Flag the line items where the gap is widest and suggest what would need to happen for the company to hit the guided range. Note whether guidance was raised, lowered, or maintained.

[INSERT CONSENSUS AND GUIDANCE DATA]
Earnings

Guidance & Estimate Revision Tracker

When you need to see where management guidance diverges from what the Street expects.

Prompt 06
From the following earnings release for [COMPANY NAME], extract the 8 key operating metrics that matter most for this business model. For each, compare to the prior quarter and prior year, calculate the change in absolute and percentage terms, and note whether the trend is accelerating, decelerating, or stable. Flag any metric where the company changed its definition or reporting methodology.

[INSERT EARNINGS RELEASE]
Earnings

Post-Earnings KPI Dashboard

When you need a structured view of how the key operating metrics trended quarter over quarter.

Modelling

3 prompts.

Prompt 07
You are a financial modelling specialist. For [COMPANY NAME], propose a DCF model structure including revenue build approach, margin assumptions, working capital drivers, and the three sensitivity variables that would have the largest impact on valuation. Group assumptions by category and note which ones require primary research versus public data. Suggest the output tabs and the key summary metrics to display on a dashboard.

[INSERT COMPANY DESCRIPTION AND FINANCIAL SUMMARY]
Modelling

Model Architecture & Assumption Framework

When you are starting a new model and need to align on structure and assumptions before writing formulas.

Prompt 08
Design a three-scenario framework for [COMPANY NAME] covering base, bull, and bear cases. For each scenario, specify what changes in revenue growth, margin trajectory, and capital intensity. Identify the two two-way sensitivity tables that would best illustrate where valuation risk concentrates. Explain why each sensitivity is worth running and what decision it would inform.

[INSERT BASE CASE ASSUMPTIONS]
Modelling

Scenario Design & Sensitivity Testing

When you need to stress-test a model and identify where valuation risk concentrates.

Prompt 09
Using the following trading comparables and precedent transactions for [COMPANY NAME], build a valuation football field. Identify which multiples are most appropriate for this business model, note any outliers in the comp set, and calculate the implied per-share value range. Flag where the current market price sits relative to this range and what that implies about market expectations.

[INSERT COMP DATA AND PRECEDENT TRANSACTIONS]
Modelling

Valuation Cross-Check

When you need to triangulate a DCF output against market-based valuation methods.

Comps

3 prompts.

Prompt 10
Identify 8-10 comparable companies for [COMPANY NAME] in the [SECTOR] sector. For each, note market cap, revenue, EBITDA margin, revenue growth rate, and the primary valuation multiple used in this sub-industry. Rank them by similarity to [COMPANY NAME] based on business model, end markets, and scale. Flag any that should be excluded as outliers and explain why.

[INSERT COMPANY DESCRIPTION]
Comps

Comparable Set Builder

When you need to build a comp set from scratch and decide what belongs and what does not.

Prompt 11
Using the following comp set, calculate EV/Revenue, EV/EBITDA, and P/E for each company. Identify the median, 25th percentile, and 75th percentile for each multiple. Flag companies trading at a meaningful premium or discount to the median and explain the likely operational reason for the divergence. Note which multiples are most reliable for this sub-industry and which are distorted by one-time factors.

[INSERT COMP SET DATA]
Comps

Trading Multiples Benchmark

When you need to see where a company trades relative to its peers and understand why.

Prompt 12
From the following M&A transactions in [SECTOR] over the past 3 years, extract deal value, target revenue, target EBITDA, and the implied transaction multiples. Group by deal type (strategic, financial sponsor, take-private). Identify whether takeout premiums are trending up or down. Flag any transactions that are not truly comparable and explain why. Calculate the median and range for each multiple type.

[INSERT TRANSACTION DATA]
Comps

Precedent Transaction Analysis

When you need to understand what acquirers have actually paid for similar businesses.

IC Memos

3 prompts.

Prompt 13
Structure an investment thesis for [COMPANY NAME] as a [LONG/SHORT] idea. Cover the core argument in three sentences, the three catalysts that will drive the stock toward your target, the two principal risks, and the expected holding period. Write it for an IC audience that has 90 seconds to form a view. Avoid jargon and lead with the single most important point.

[INSERT YOUR RESEARCH NOTES]
IC Memos

Investment Thesis Structuring

When you have a view forming and need to structure it into something an IC can evaluate in 90 seconds.

Prompt 14
Draft an IC memo for a [LONG/SHORT] position in [COMPANY NAME]. Structure it as investment thesis, key drivers, financial summary, valuation approach, risk factors, and position sizing recommendation. Keep it to 800 words. Write in the house style of a fundamentals-driven long/short fund. Lead each section with the conclusion and follow with supporting detail. Do not include any recommendation the data does not support.

[INSERT YOUR RESEARCH AND ANALYSIS]
IC Memos

IC Memo First Draft

When your research is done and you need to turn it into a structured memo your committee can act on.

Prompt 15
For the following investment thesis on [COMPANY NAME], identify the five things that could go wrong. For each risk, classify it as company-specific, sector-level, or macro. Rate each on probability (high, medium, low) and impact (high, medium, low). Suggest what monitoring signals would tell you the risk is materializing, and what the response would be if it did.

[INSERT INVESTMENT THESIS]
IC Memos

Risk Factor Decomposition

When you have a thesis and need to pressure-test it before presenting to the committee.

Portfolio

2 prompts.

Prompt 16
Using the following portfolio holdings and current positions, identify the three positions where the investment thesis is most at risk. For each, note what has changed since initiation, whether the catalysts are still intact, and whether the position sizing still makes sense relative to conviction. Flag any position where the downside case has become more likely than at initiation.

[INSERT PORTFOLIO HOLDINGS AND THESIS NOTES]
Portfolio

Position Monitor & Catalyst Tracker

When you need to review existing positions and identify where the thesis is under pressure.

Prompt 17
For the following portfolio companies reporting in the next two weeks, identify the three metrics to watch for each, the consensus expectations, and what would constitute a positive versus negative surprise. Flag any companies where the setup into the print looks risky based on recent estimate revisions, price action, or sector-specific headwinds. Suggest one question to investigate for each name before the print.

[INSERT PORTFOLIO COMPANIES AND POSITIONS]
Portfolio

Earnings Season Prep

When earnings season is approaching and you need to prepare for every portfolio company reporting.

Communication

1 prompt.

Prompt 18
Draft a quarterly investor letter covering portfolio performance, attribution by sector, two wins and two losses from the quarter, and the outlook for the next 90 days. Keep it under 400 words and write in the voice of a fundamentals-driven portfolio manager addressing limited partners. Lead with the most important takeaway, not with performance numbers. Avoid jargon your LPs would not use themselves.

[INSERT PERFORMANCE DATA AND ATTRIBUTION]
Communication

Investor Letter Drafting

When you need to communicate quarterly performance and positioning to limited partners.

Common mistakes to avoid

Vague task description

“Analyze this company” gives too little direction. Specify the type of analysis, the audience, and the format you need.

Missing assumptions

If your forecast relies on specific growth rates or margin targets, include them. The model cannot infer your assumptions from context.

Unclear output expectations

Not specifying format, length or tone produces outputs that need significant reformatting before they are usable.

Treating AI output as a conclusion

AI can draft the commentary, but it cannot know what actually drove the variance. Always validate against real operational data.

Chain-of-thought prompting

For complex analytical tasks, structure your prompt to ask the model to reason step by step. Ask it to show assumptions, logic and conclusions separately. This improves output quality and makes errors easier to catch. The model that shows its work is easier to verify than the one that hands you a number.

Boundaries

When AI should not make the call.

AI accelerates the analytical process, but investment judgment stays with the analyst. There are specific scenarios where AI should not be relied on for final outputs or decisions.

Final investment decisions

The decision to commit capital is a judgment call that weighs asymmetric information, portfolio context and risk tolerance. AI can inform the decision. It cannot make it.

Regulated financial judgments

Accounting policy elections, impairment assessments and revenue recognition determinations require documented professional judgment. These cannot be delegated to a model.

Confidential deal analysis

Transaction-related forecasts carry strict confidentiality obligations. Do not input deal-specific data into any tool without an enterprise agreement that covers the use case.

Material non-public information

Never input MNPI into a consumer AI tool. If your firm has an enterprise deployment with appropriate data controls, confirm what is and is not permitted before using it.

FAQ

Frequently asked questions.

Yes. AI can support drafting company primers, structuring earnings analysis, building model frameworks, and generating first drafts of IC memos. The investment judgment and assumption ownership must remain with the analyst, and human review is essential before any AI-assisted output informs a decision.

Use only AI tools approved by your firm's compliance and IT teams. Avoid inputting material non-public information into any tool without an enterprise data agreement. Treat every AI output as a first draft requiring verification against source data.

Include a clear role, a specific task, relevant context such as the company and time period, the data to work from, the desired output format, and any constraints on length or tone. Company-specific context is what separates effective finance prompts from generic ones.

Avoid AI for regulated financial judgments, final investment decisions, confidential deal analysis, and any output that will inform a high-stakes decision without independent verification. AI accelerates the analytical process but does not replace professional judgment.

Specificity. A prompt that names the company, the audience, the time period, the data source, and the output format will almost always outperform a generic instruction. The more context you provide, the more useful the output.

Workshops

Join Pascal Academy

Learn directly from the Pascal team through live workshops.

Sign Up

USA

228 Park Ave S PMB 55152, New York

UK

86-90 Paul Street, EC2A 4NE, London

India

Indiqube Edge, ORR, Bangalore

Copyright © 2026 | All Rights Reserved

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.