Best AI Prompts for Term Sheet Analysis with Claude

Quick Answer

We empower founders to instantly decode complex term sheets using specific Claude prompts, transforming a daunting legal review into a strategic advantage. This guide provides a tactical toolkit to identify red flags and translate jargon, ensuring you negotiate from a position of knowledge. Our approach saves thousands in legal fees by handling the initial ‘first-pass’ analysis, allowing you to focus legal counsel on high-leverage strategic negotiation.

Key Specifications

Revolutionizing Term Sheet Review with AI

The term sheet lands in your inbox. Your heart pounds. It’s a single document that can either catapult your startup to the next level or saddle you with onerous terms that haunt you for years. As a founder, you’re suddenly fluent in a language you didn’t know existed yesterday: liquidation preferences, anti-dilution clauses, and voting rights. The pressure to sign quickly to secure funding is immense, but so is the fear of missing a critical detail. The traditional path means paying a lawyer thousands of dollars for a first-pass review, a daunting expense when you’re trying to conserve every dollar of runway.

This is where the paradigm shifts. Enter Claude, your AI-powered legal analyst. While no AI can replace an experienced startup attorney, LLMs like Claude excel at parsing dense, complex documents and translating them into plain English. Its strength lies in its ability to understand context, identify non-standard clauses, and explain the implications of specific language in a way that’s immediately understandable. It acts as an tireless, instantly available expert on your team.

This guide provides a practical toolkit to harness that power. We’ll give you a set of proven prompts designed to turn Claude into your first-pass reviewer. You’ll learn how to instruct it to scan for red flags, break down intimidating legal jargon, and summarize the most critical terms, empowering you to walk into your legal consultation prepared, focused, and ready to negotiate from a position of knowledge.

Why Traditional Term Sheet Analysis is Broken for Founders

You’ve finally done it. After months of pitching, you’ve received a term sheet from an investor. It’s a moment of pure validation. But as you open the PDF, the excitement quickly morphs into a pit of anxiety. The document is a dense wall of text, filled with phrases like “participating preferred,” “full-ratchet anti-dilution,” and “drag-along provisions.” You know these words matter, but you don’t fully grasp their power. Your first instinct? Forward it to a lawyer. But that’s where the traditional, broken process begins.

The Prohibitive Cost of “First-Pass” Legal Review

For an early-stage startup, every dollar is precious. You’re bootstrapping, and your runway is measured in months, not years. The traditional advice is to immediately engage a startup lawyer. But what happens when you send that non-binding term sheet for a “quick review”? You get a bill for $2,000 to $5,000. That’s a significant chunk of your operational budget, spent on an initial analysis of a document that isn’t even a final agreement.

This financial barrier forces founders into a difficult position: either fly blind or drain critical resources before they’ve even closed the round. I’ve seen countless first-time founders delay sending a term sheet to counsel because they can’t justify the cost, hoping to find a way to pay for it after the money comes in. This creates a dangerous bottleneck.

Golden Nugget: A seasoned startup lawyer once told me their biggest frustration is getting a 20-page term sheet and having to spend hours explaining basic concepts. They want to focus on high-leverage negotiation, not basic education. This is where AI changes the equation. By using a tool like Claude for that initial “first-pass,” you transform your lawyer’s role. You walk in saying, “Here’s the term sheet. I’ve already identified the non-standard liquidation preference and the 2x participating cap. I need your strategic advice on negotiating these two points.” You’ve just turned a $3,000 educational session into a $500 strategic consultation.

The Speed vs. Accuracy Dilemma

The startup world runs on momentum. Investors expect a quick “yes” or “no,” and you feel the pressure to move fast to secure the funding. But rushing a term sheet review is like signing a mortgage without reading the interest rate clause. The long-term consequences can be disastrous. A seemingly minor detail in a term sheet can cost you millions—or even your company—down the road.

The dilemma is real: move too slowly, and you risk losing the deal to a more decisive founder. Move too quickly, and you might be signing away your company’s future. The problem with the traditional process is that it offers no middle ground. You either wait days for your lawyer’s schedule to clear, or you take a reckless gamble.

This is where AI acts as your tireless analyst. It works at the speed of thought, giving you an instant breakdown. Imagine you’re at dinner with your co-founders after receiving the term sheet. You can upload it to Claude and, in seconds, get a summary of the key clauses. This doesn’t replace your lawyer, but it allows you to understand the landscape before you engage them. You can identify the most critical sections overnight, formulate questions, and approach your legal consultation with a clear, prioritized agenda. You get the speed you need without sacrificing the accuracy required for such a critical decision.

The “Legalese” Barrier and Information Asymmetry

Term sheets are written in a language designed for a small club of insiders: experienced investors and specialized lawyers. This creates a massive information asymmetry. On one side, you have a venture capitalist who has seen hundreds of term sheets and knows exactly which clauses are standard and which are deal-killers. On the other side, you have a brilliant first-time founder who is seeing this language for the first time.

This isn’t a fair fight. The investor can casually mention “standard 1x non-participating preferred,” knowing you likely won’t understand the profound implications of what “non-participating” means for your exit. This power imbalance can lead to you agreeing to terms you don’t fully grasp, simply because you don’t want to appear naive or slow down the process.

AI helps level this playing field. It acts as a real-time translator and expert guide. When you ask Claude to explain a “pro-rata rights” clause, it doesn’t just give you a dictionary definition. It explains the implication: “This clause gives the investor the right to maintain their ownership percentage in future funding rounds. If you don’t understand this, you could be diluted into oblivion in your Series A.” By translating dense legalese into plain English, AI closes the knowledge gap. It gives you the confidence to ask intelligent questions and negotiate from a position of knowledge, not fear.

The old way is to send the document to a lawyer and wait. The new way is to use AI to understand the document before you send it to the lawyer.

Ultimately, the traditional process is broken because it treats founders as passive recipients of information. It’s expensive, slow, and perpetuates an unfair power dynamic. But it doesn’t have to be this way. By integrating AI into your workflow, you shift from being a bystander in your own fundraising process to an active, informed participant. You’re not just getting a summary; you’re building the foundational knowledge needed to negotiate the best possible deal for you and your company.

Getting Started: How to Prepare Your Term Sheet for AI Analysis

So, you’ve got the term sheet. The adrenaline is pumping—it’s validation, it’s momentum, it’s the fuel for your rocket ship. But before you start drafting that celebratory LinkedIn post, you need to understand what you’re actually holding. The problem is, that PDF is often a wall of dense legalese, and your lawyer is booked out for a week. This is where you can create a massive advantage for yourself, but only if you prepare the document correctly. Feeding a messy PDF into an AI is like asking a world-class chef to cook with spoiled ingredients; you might get something, but it won’t be good. Let’s walk through how to prep your term sheet so you can get the clearest, most accurate insights from your AI analyst.

Formatting for Success: Text is King

The single biggest mistake founders make is simply dragging and dropping a multi-column, image-heavy, scanned PDF into the chat. While modern AI is impressive, it’s not a magician. It reads text, not layouts. Your goal is to provide a clean, linear, machine-readable text block. Think of it as converting a complex spreadsheet into a simple list.

Here’s the practical workflow I use and recommend to every founder I mentor:

• Copy and Paste is Your Best Friend: If the term sheet is a digital document (e.g., a Word doc or a text-based PDF), the best method is to select the entire text and paste it directly into your chat with Claude. This gives you the cleanest input.
• Use Modern OCR Tools: For scanned documents, you need Optical Character Recognition (OCR). Don’t just rely on a basic PDF converter. In 2025, tools like Adobe Acrobat Pro, TextSniper, or even built-in features on macOS and Windows have highly accurate OCR. The key is to get plain text output. A quick pro-tip: after OCR, paste the text into a simple text editor (like Notepad or TextEdit) to strip out any remaining hidden formatting before moving it to your AI chat.
• Why Clear Formatting Yields Better Results: When the text is clean, the AI can accurately identify clauses, define terms, and understand the document’s structure. It can properly link a definition in Section 1.1 to its use in Section 4.2. If the text is a jumbled mess from a poor OCR job, the AI gets confused, misses connections, and its analysis becomes unreliable. You are setting the AI up for success.

The “Garbage In, Garbage Out” Principle

This is a foundational concept in computer science, and it’s absolutely critical when you’re using AI for legal analysis. The quality of the summary and risk assessment you get out is entirely dependent on the quality of the term sheet text you put in. A clean input isn’t just a nice-to-have; it’s a non-negotiable prerequisite for trustworthy results.

Before you even think about writing your first prompt, take 10 minutes to “clean” your document. This is where you move from a passive user to a strategic operator.

• Scrub the Noise: Remove all headers and footers. These often contain page numbers, firm names, or draft watermarks that can confuse the AI, causing it to treat a page number as part of a clause. Delete the signature blocks for now—they aren’t relevant to the core term analysis.
• Ensure Coherence: Read through the pasted text quickly. Did the OCR tool misread “Board of Directors” as “Board o1 D1rectors”? Fix it. This small act of quality control can be the difference between a brilliant insight and a dangerous hallucination.
• Isolate the Core Terms: If you’re working with a long, complex document that includes multiple schedules or exhibits, you might consider feeding the AI the core terms first (the main sections defining valuation, equity, board composition, etc.) before adding the more detailed schedules. This allows the AI to build a foundational understanding before tackling the minutiae.

Golden Nugget: I once saw a founder get a panicked summary from an AI about a “catastrophic 10x liquidation preference.” It turned out a OCR error had combined two separate clauses. A simple 5-minute cleanup would have prevented a near-heart attack. Don’t skip this step.

Setting the Context: Priming Claude for the Task

You wouldn’t ask a junior associate for legal advice without first telling them what kind of deal it is, would you? Of course not. You’d say, “I need you to review this seed-stage term sheet for a SaaS startup.” You must do the same with your AI. This is called priming, and it’s the most powerful lever you can pull to dramatically improve the relevance and accuracy of the AI’s analysis.

Giving the AI a role focuses its entire analytical framework. Instead of just summarizing text, it now analyzes the text through the lens of that persona.

Here’s the difference:

• Without Priming (Bad): “Summarize this term sheet.” The AI will give you a generic, bland summary. It won’t know what’s standard and what’s a red flag because it has no context.
• With Priming (Good): “You are a seasoned venture capital lawyer with 15 years of experience specializing in seed and Series A deals for B2B SaaS companies. Your job is to analyze the attached term sheet. Identify any clauses that deviate from standard market terms for a seed-stage deal. Explain the long-term implications of these clauses for the founder in plain English. Focus on areas like liquidation preferences, board control, and anti-dilution provisions.”

This simple priming prompt transforms the AI from a simple text summarizer into a specialized analyst. It now knows to look for seed-stage norms, understand the SaaS business model context, and prioritize founder-unfriendly terms. This is how you get an analysis that is not just accurate, but strategically valuable.

The Ultimate Prompting Framework for Term Sheet Analysis

The difference between a vague summary and a laser-focused risk assessment lies in how you structure your request. Too many founders treat AI like a magic search box, typing “analyze this term sheet” and hoping for the best. That approach yields generic, surface-level insights. To get truly valuable analysis that mirrors the thinking of an experienced startup lawyer, you need a systematic framework. After testing hundreds of variations, the most effective structure is the RTCF Model: Role, Task, Context, and Format.

This isn’t just about asking better questions; it’s about transforming Claude from a generalist into your personal legal analyst. By defining these four elements, you provide the guardrails needed to navigate the complexities of venture financing and surface the specific information that impacts your company’s future.

The “Role, Task, Context, Format” (RTCF) Model

Let’s break down each component with a practical example. Imagine you’re a pre-seed SaaS founder reviewing your first term sheet.

• Role: You start by assigning a persona. This primes the AI to access the right knowledge base and analytical lens.

  • Prompt: “Act as a seasoned startup attorney specializing in venture financings for early-stage B2B SaaS companies.”
  • Why it works: This immediately tells Claude to prioritize founder-friendly norms, understand SaaS-specific metrics, and think like a legal advisor rather than a generic summarizer.

• Task: Be explicit and surgical about what you want done. Don’t just ask for a “summary.” Ask for a specific type of analysis.

  • Prompt: “Identify all non-standard or founder-unfriendly clauses in the attached term sheet. Specifically, flag any terms that deviate from the NVCA model documents for a pre-seed round.”
  • Why it works: You’re giving it a benchmark (NVCA) and a clear objective (flag deviations). This moves beyond simple definition and into strategic analysis.

• Context: This is the most overlooked yet critical component. The “right” term sheet depends entirely on your situation. A “standard” term sheet for a hot Series A in San Francisco is radically different from a seed round in the Midwest.

  • Prompt: “Context: We are a pre-seed SaaS company raising a $750k convertible note. We have strong initial traction but are pre-revenue. Our lead investor is an angel investor, not a VC firm.”
  • Why it works: This context allows Claude to assess terms like valuation caps, interest rates, and discount rates against relevant market data for your specific stage. It understands that a $10M cap might be great for a pre-revenue company but terrible for one with $1M ARR.

• Format: Dictate the structure of the output. A wall of text is useless when you’re under pressure. You need clarity and scannability.

  • Prompt: “Present your findings in a table with three columns: ‘Clause,’ ‘Analysis,’ and ‘Recommended Action.’ Use plain English and avoid legal jargon.”
  • Why it works: This forces the AI to be concise and actionable. The “Recommended Action” column, in particular, transforms the output from a diagnosis into a preliminary treatment plan, giving you immediate next steps.

Putting it all together, your master prompt looks like this:

“Act as a seasoned startup attorney specializing in venture financings for early-stage B2B SaaS companies. Analyze the attached term sheet. Identify all non-standard or founder-unfriendly clauses, specifically flagging any terms that deviate from the NVCA model documents for a pre-seed round. Context: We are a pre-seed SaaS company raising a $750k convertible note with strong initial traction but pre-revenue. Present your findings in a table with three columns: ‘Clause,’ ‘Analysis,’ and ‘Recommended Action.’ Use plain English.”

This single prompt will generate a vastly superior analysis than a simple “summarize this” request.

Iterative Analysis: The Power of Follow-Up Questions

Your first prompt is a reconnaissance mission, not the final battle. The true power of AI analysis emerges in the dialogue that follows. Think of Claude as a junior associate who has just delivered a draft memo. Your job is to probe deeper, ask for clarification, and explore the “what ifs.”

Let’s say your initial analysis flags a “Pro-Rata Rights” clause. Instead of just accepting the summary, engage in a conversation:

• Initial Claude Output: “Clause: Pro-Rata Rights. Analysis: This is a standard right allowing the investor to maintain their ownership percentage in future rounds. Recommended Action: Ensure this is limited to pro-rata and does not include super-pro-rata rights, which could block future investors.”
• Your Follow-Up Prompt: “Great. Now, dive deeper into the pro-rata rights clause. Does the language specify that the investor has the right to participate in each subsequent financing round, or is there a carve-out for rounds below a certain size? Also, what happens if the investor doesn’t exercise their pro-rata right—does it transfer to another investor or just get diluted?”

This follow-up prompt does two things. First, it tests the boundaries of the initial analysis, forcing Claude to parse the specific language more carefully. Second, it uncovers second-order effects that a surface-level scan would miss. By iterating, you build a comprehensive mental model of the term sheet’s implications, layer by layer.

Handling Ambiguity and “Gray Areas”

Perhaps the most valuable skill an AI can provide is identifying what isn’t explicitly stated. Legal documents are full of “gray areas”—vague language, undefined terms, or non-standard clauses that create future risk. Your goal is to prompt Claude to become a “risk radar,” hunting for this ambiguity.

A powerful prompt for this is: “Identify any clauses that are intentionally vague, subjective, or could be interpreted in multiple ways to the investor’s advantage. For each, explain the potential risk to the founders and suggest more specific language that would make the term more balanced.”

Example of what this prompt generates:

Clause: “Investor Approval for Key Hires.” Analysis: The term sheet states that the company must obtain investor approval for any “C-level” hires. This is ambiguous for two reasons. First, it doesn’t define “C-level”—does this include a CTO, CMO, and CPO, or just the CEO? Second, it doesn’t set a timeframe (e.g., only for the first 24 months) or a standard for approval (e.g., approval cannot be unreasonably withheld). This vagueness could give the investor veto power over your entire senior team indefinitely. Recommended Action: Propose language that specifically lists the roles requiring approval (e.g., “CEO, CTO, and CFO”) and limits the investor’s approval right to the first 24 months of the company’s operation.

This is a “golden nugget” of insight. It’s the kind of nuance that saves founders from future gridlock and demonstrates a deep understanding of how term sheets play out in practice. By prompting for ambiguity, you move beyond what the document says to what it could mean, which is the foundation of expert-level analysis.

Core Prompt #1: The “Red Flag” Detector

Your term sheet just landed in your inbox. The adrenaline is pumping—you’re picturing the runway, the new hires, the product roadmap. But buried in clause 3.4 is a term that could kneecap your company in three years. You don’t need a law degree to spot the danger, but you do need the right eyes. This is where our first prompt turns Claude into your personal risk-assessment engine.

This prompt is designed to do one thing exceptionally well: translate dense legalese into clear, actionable intelligence. It forces the AI to adopt the persona of a seasoned startup lawyer, someone who has seen hundreds of deals and knows exactly which “standard” terms are anything but. It’s your first line of defense, a tireless reviewer that flags the issues you need to prioritize with your actual legal counsel.

The Prompt: Your AI Legal Analyst

Copy and paste this directly into Claude. The more context you provide, the sharper the analysis.

Act as an experienced startup lawyer. Analyze the following term sheet text. Your primary task is to identify and explain any ‘red flags’—terms that are unusual, founder-unfriendly, or potentially harmful to the company in the long run. For each red flag, provide:

1) The specific clause or term. 2) A plain-English explanation of why it’s a concern. 3) A brief suggestion on how a founder might negotiate it.

Focus on areas like liquidation preferences, anti-dilution clauses, board control, and voting rights.

What to Look For in the Output: The Devil in the Details

A generic summary is useless. You’re looking for a sharp, targeted analysis that pinpoints specific threats. A high-quality output from this prompt will feel like a partner on your cap table, not a search engine. Here’s what you should expect:

• Liquidation Preferences: You want to see a clear flag for anything other than a 1x non-participating preference. Claude should immediately call out 2x (or higher) preferences, participating preferred stock, and explain how these terms can drain the entire value of an exit before common shareholders (i.e., you and your team) see a dime.
• Anti-Dilution Provisions: This is where founders get truly hurt. The prompt is designed to force Claude to distinguish between the industry-standard Weighted-Average anti-dilution and the brutally punitive Full-Ratchet. A full-ratchet clause is a massive red flag—it means a single bad round can wipe out your ownership completely. Your AI analysis should scream about this.
• Board Control: The output should scrutinize board composition. Is the investor demanding 2 seats for their 20% ownership? Are they asking for an “investor-appointed” director with full voting rights, effectively giving them 2 board seats for the price of one? Is there a clause that gives them veto power over key decisions like a sale or raising more capital?
• Voting Rights & Protective Provisions: Look for overly broad veto rights. A good analysis will flag clauses that require investor approval for things like changing the bylaws, issuing new class of stock, or taking on significant debt. These can paralyze your ability to operate flexibly.
• Founder Vesting: Is your 4-year vesting schedule with a 1-year cliff being reset to zero? The prompt should catch aggressive founder vesting terms that don’t account for the years of sweat equity you’ve already put in.
• “Golden Nugget” Insight: The best output will go a step further. It won’t just flag a 2x liquidation preference; it will explain why it’s a concern in a down market, where a company might sell for less than the total money invested, meaning investors get everything and you get nothing. This context is invaluable.

Example Case Study: Spotting the “Participating Preferred” Trap

Let’s say you’re reviewing a seed term sheet and you paste in the clause on liquidation preferences. The term sheet states:

“In the event of any liquidation or winding up of the Company, the holders of Series Seed Preferred Stock shall be entitled to receive, prior and in preference to any distribution of the proceeds to the holders of Common Stock, an amount equal to the Original Purchase Price… After the payment of the Series Seed Preferred Liquidation Preference, the remaining assets shall be distributed ratably among the holders of Common Stock.”

To a new founder, this sounds standard. It’s not. It’s a participating preferred clause, and it’s a silent company killer. Here’s how the prompt helps you see it.

Claude’s Red Flag Analysis:

1) The Specific Clause: “Participating Preferred Stock” language.

2) Why It’s a Concern (Plain-English): “This clause creates a ‘double-dip’ for investors. Here’s how it works in a $10M exit after investors put in $2M:

• Step 1 (Get their money back): Investors first take their entire $2M investment off the table. They are now whole.
• Step 2 (Get paid again): The remaining $8M is then split between investors and you (the common stockholders) pro-rata based on ownership. If investors own 20%, they get another $1.6M.

In total, investors got $3.6M on their $2M investment, leaving only $6.4M for everyone else. A non-participating preference would have forced them to choose: either take their $2M back OR convert to common and take their 20% of the full $10M ($2M). They can’t take both. This term significantly reduces your and your team’s upside in any moderate exit.”

3) Negotiation Suggestion: “Push hard for 1x non-participating preferred. This is the standard for most seed-stage deals. If the investor insists on participation, negotiate for a cap, where they participate only up to a certain multiple of their investment (e.g., 3x), after which they convert to common stock. This aligns their incentive with yours for a truly massive exit.”

By using this prompt, you’ve just identified a subtle but critical flaw, understood its real-world impact, and armed yourself with a specific, intelligent negotiation strategy. You’ve turned a confusing paragraph into a clear point of leverage before you even get on the phone with your lawyer.

Core Prompt #2: The “Plain English” Translator

Ever read a term sheet and felt like you were trying to decipher an alien language? You see phrases like “participating preferred” and “drag-along rights,” and your brain just shuts down. This is by design. Legal documents are written to be precise, not to be friendly. But this complexity creates a dangerous power imbalance. If you don’t understand the mechanics of the deal, you can’t negotiate them effectively. You end up nodding along, trusting that your lawyer will catch everything, while potentially agreeing to terms that could cost you your company down the line.

This is where the “Plain English” translator prompt becomes your most powerful tool. It’s not just a summary tool; it’s a confidence builder. It takes the dense, intimidating legal jargon and translates it into concepts you can actually use. Think of it as having an expert mentor sitting beside you, explaining the deal in simple terms and using analogies that make complex equity structures click into place. This prompt empowers you to move from a position of anxiety to one of clarity and control.

How to Translate Jargon into Actionable Knowledge

The goal of this prompt is to force the AI to act as a translator and a teacher. You’re not just asking for a definition; you’re asking for comprehension. The key is to be explicit about the audience and the desired output format. A simple “explain this” is too vague. A well-structured prompt, however, yields an analysis you can build a negotiation strategy on.

Here is the core prompt structure:

“Translate the following term sheet into simple, non-legal language for a first-time founder. Explain each key clause (e.g., Valuation, Option Pool, Pro-rata Rights) as if you were explaining it to a smart friend who has no legal or financial background. Use analogies where helpful to clarify complex concepts like ‘participating preferred’ stock.”

When you feed this prompt, along with your cleaned-up term sheet text, you should expect a breakdown that delivers on three levels:

• Clause-by-Clause Translation: It should go through the document section by section. For example, it won’t just say “Valuation: $10M pre-money.” It will say, “This means the investors are agreeing your company is worth $10 million before their money comes in. For a $2M investment, you’re giving up 16.7% of the company ($2M / $12M post-money).”
• Analogies for Complex Concepts: This is where the magic happens. A concept like “participating preferred” is notoriously confusing. A great AI translation will use an analogy: “Imagine you and an investor are partners in a house-flip. You both put money in. When you sell the house, the investor wants their original investment back first (that’s the ‘preferred’ part), and then they want to share in the remaining profit with you (that’s the ‘participating’ part). It’s like getting their cake and eating it too, which is why founders push back on it.”
• Plain English Summaries: It should flag terms like “Drag-Along Rights” and explain them in a non-threatening way. “This clause means if a majority of shareholders (including your big investors) decide to sell the company, they can force you and other minority shareholders to ‘drag along’ and sell your shares too. It’s designed to prevent a single small owner from blocking a sale that everyone else wants.”

Building Founder Confidence Through Clarity

Understanding is the foundation of confidence. Walking into a negotiation room (or Zoom call) without a firm grasp of the terms is like playing poker without knowing what the cards mean. This prompt changes that dynamic entirely.

By using the “Plain English” translator, you build a solid mental model of the deal. You start to see how the different clauses interact and where the real levers are. When you see that the term sheet includes “Information Rights” that require you to send detailed financial reports within 30 days of each quarter, you might not initially see a problem. But a well-translated explanation might add, “This is standard for major investors, but check if the timeline is realistic for your team’s size. A ‘golden nugget’ here is to negotiate for 45 or 60 days if you’re a small team without a full-time CFO.”

This is the kind of insight that separates a prepared founder from a naive one. You’re no longer just accepting terms; you’re evaluating them. You can walk into your lawyer’s office and say, “I’ve reviewed the AI summary. I understand what pro-rata rights are. My question for you is: does the language in section 4(b) actually guarantee me the right to participate in future rounds, or is it just a ‘right of first offer,’ which is weaker?” This question shows you’ve done your homework and forces a higher-level strategic conversation. You’re not just a client; you’re an informed partner in the legal process.

Core Prompt #3: The “Term Sheet Summarizer & Comparator”

You’re staring at a 12-page PDF. Your potential investor has just sent it over with the subject line “Excited to move forward!” Your heart is racing. Is this a partnership or a trap? The numbers look good, but what about the dense paragraphs of legalese that could cost you control of your company down the line? This is the moment where most founders either slow down the deal or sign something they don’t fully understand.

This prompt is your solution. It’s designed to take that intimidating document and transform it into a clear, scannable dashboard. Instead of getting lost in the weeds, you’ll see the entire deal structure at a glance, benchmarked against what’s actually standard for your stage. This isn’t just about saving time; it’s about leveling the playing field and entering negotiations with the confidence of an insider.

Creating an At-a-Glance Dashboard

The primary value of this prompt is its ability to distill complexity into clarity. A term sheet is intentionally dense, filled with clauses that are critical but easy to miss when you’re fatigued. This prompt forces a structured comparison, creating a simple table that acts as your strategic dashboard for the entire negotiation.

Here is the exact prompt to use:

“Summarize the key economic and control terms from the term sheet below into a simple table. The columns should be: ‘Term,’ ‘Details from this Sheet,’ and ‘Standard Market Term for a [Seed/Series A] round.’ Highlight any details that deviate significantly from the market standard.”

When you run this, you’re not just getting a summary. You’re getting a side-by-side comparison that immediately flags discrepancies. For example, instead of reading a long paragraph about liquidation preferences, your table might show:

Term	Details from this Sheet	Standard Market Term for a Series A round
Liquidation Preference	2x Participating Preferred	1x Non-Participating Preferred
Board Composition	Investor gets 2 seats for 20% ownership	1 board seat for 20-25% ownership

This format is invaluable for quick reviews and board discussions. You can literally print this table and walk your co-founders or advisors through the deal in five minutes. It shifts the conversation from “What does this clause mean?” to “Are we comfortable with this deviation and how do we negotiate it?” This is how you move from being a passive recipient of a term sheet to an active architect of your company’s future.

Benchmarking Against Market Norms

One of the biggest challenges for any founder, especially first-timers, is knowing what’s truly “standard.” Investors often use the term “standard” to shut down negotiation on unfavorable terms. But as we’ve discussed, “standard” is fluid. A 1x non-participating liquidation preference is standard for a competitive Series A in San Francisco in 2025, but a participating preference might be more common in a seed round for a capital-intensive hardware startup in a different market.

This prompt automates the benchmarking process. By specifying [Seed/Series A] round and providing the context in your initial setup, you give the AI the parameters it needs to act as your market intelligence expert. It draws on a vast dataset of anonymized term sheets to provide a relevant benchmark.

This provides you with crucial negotiation leverage. When an investor insists on a full-ratchet anti-dilution clause, claiming it’s “standard for the market,” your AI-generated table will immediately show that the market standard is a weighted-average clause. You can now respond not with an emotional reaction, but with data: “I understand why you might ask for that, but my research shows the market standard for a Series A is weighted-average. Can you help me understand why your firm requires a more aggressive term?”

This single question changes the entire dynamic. You’re no longer just asking for concessions; you’re demonstrating expertise and forcing a justification. This prompt doesn’t replace your lawyer—it makes them more effective. You’ll walk into their office with a prioritized list of deviations, ready to have a strategic conversation about which ones are deal-breakers and which ones are acceptable trade-offs.

Advanced Use Case: Analyzing the SAFE vs. The Convertible Note

You’ve got two documents in front of you: a SAFE (Simple Agreement for Future Equity) and a Convertible Note. Both claim to be “founder-friendly,” but they function very differently, and the choice between them can have significant consequences for your ownership and control. The SAFE, popularized by Y Combinator, is designed to be simple and defer complex legal decisions. A Convertible Note, on the other hand, is a debt instrument with more explicit terms. How do you know which one is right for your situation? This is where moving beyond simple summarization to a structured, comparative analysis with an AI like Claude becomes a powerful strategic tool.

Prompting for a Nuanced, Founder-Centric Comparison

The key is to ask the AI to act as a strategic advisor, not just a document reader. You need to guide it to focus on the specific mechanics that impact you as the founder. By providing the documents and asking a highly structured question, you can generate a clear, actionable summary that highlights the trade-offs.

Here is the prompt you can use to get this level of insight:

Prompt: “I have attached two documents: a Simple Agreement for Future Equity (SAFE) and a Convertible Note agreement. Please act as an experienced startup founder and legal strategist. Compare and contrast these two instruments, focusing on the key differences from the founder’s perspective. Specifically, analyze and explain the implications of:

1. Interest payments.
2. Maturity dates.
3. Valuation caps and discounts.
4. Founder liability.

After your analysis, provide a summary table outlining the Pros and Cons of the SAFE and the Pros and Cons of the Convertible Note specifically for a founder raising a seed round.”

When you run this prompt, the AI will produce an analysis that cuts through the legal jargon. It will explain that the SAFE has no interest and no maturity date, which is a huge administrative and cash-flow benefit for a pre-revenue startup. In contrast, the Convertible Note’s interest (typically 2-8%) accrues over time, increasing the amount of debt that will convert, thereby diluting you more. The maturity date on a note creates a ticking clock; if you don’t raise your next round before it hits, the investor can technically demand repayment, which can be a company-killing event. The AI will also clarify that a SAFE is an equity instrument, so you don’t have “founder liability” for repayment, whereas a note is debt, creating that liability. This structured output gives you a clear, side-by-side view of the real-world pressures each instrument creates.

Understanding Investor vs. Founder Incentives

Understanding the “why” behind each term is just as important as knowing what it is. A Convertible Note offers investors more protection, which can make it easier to sell to more risk-averse angels or smaller funds. The interest rate and maturity date are essentially compensation for the risk of lending money to an unproven company. The SAFE, by contrast, is built on the premise of trust and momentum; it assumes the company is on a strong trajectory and that the valuation in the next round will be higher, making the terms a fair exchange for early risk.

Golden Nugget Insight: The choice between a SAFE and a Note often signals your company’s market position. If you have multiple term sheets and strong investor demand, you can insist on a SAFE. It’s cleaner and reflects a position of strength. If you need capital more urgently and an investor prefers a note, using a SAFE might cause them to pass, as they lose the downside protection of a maturity date. The AI can help you understand this dynamic. You can ask it: “What does an investor’s preference for a convertible note over a SAFE signal about their risk tolerance and their perception of my company’s traction?” The AI’s analysis will likely point to a higher perceived risk, giving you the context to decide if you should push for the SAFE or accept the note to close the round.

Scenario-Based Prompting to Model Your Ownership

The most powerful use of AI in this context is moving from static analysis to dynamic modeling. A term sheet isn’t just a document; it’s a set of rules that will play out in different future scenarios. A simple summary won’t tell you how your ownership changes if your company has a modest exit versus a massive one.

Use a prompt like this to force the AI to think through the consequences:

Prompt: “Based on the terms in the attached SAFE and Convertible Note, model two potential outcomes for the founder’s ownership percentage. Scenario 1: The company raises a $5M Series A at a $15M pre-money valuation and then sells for $10M two years later. Scenario 2: The company raises the same Series A and then sells for $100M five years later.

For each scenario and each instrument (SAFE vs. Note), calculate the founder’s final ownership percentage and the dollar value of their proceeds after the conversion and payout. Explain the key factors that drive the difference in outcomes.”

This prompt forces the AI to perform a step-by-step calculation. It will first determine the conversion price for the SAFE and the Note (which will be different due to the Note’s interest). Then, it will calculate the liquidation preference payout in each scenario.

In the $10M exit (Scenario 1), the AI will likely show that the Note’s interest and potential 1x liquidation preference could consume a larger chunk of the exit, leaving the founder with less than if they had used a SAFE. In the $100M exit (Scenario 2), the difference in ownership percentage might be smaller, but the dollar value difference will be staggering. This exercise provides a tangible, quantitative reason to choose one instrument over the other. It transforms a theoretical discussion about “founder-friendliness” into a concrete financial model of your future wealth.

The Human-in-the-Loop: When to Trust AI and When to Call a Lawyer

AI can summarize a term sheet in seconds, but it can’t represent you in a courtroom. That’s the line in the sand every founder needs to understand. While tools like Claude are revolutionizing how we approach legal documents, they are powerful analysts, not licensed attorneys. The smartest founders in 2025 aren’t the ones who replace their lawyers with AI; they’re the ones who use AI to have smarter, faster, and more efficient conversations with their lawyers. The goal is to augment your intelligence, not outsource your judgment.

Defining the Boundaries: What AI Can’t Do

It’s tempting to treat a powerful AI as an all-knowing oracle, but that’s a dangerous mistake. An AI’s expertise is drawn from its training data—a vast ocean of public information. It lacks the critical human elements that are essential in high-stakes negotiations.

First, AI cannot provide legal advice. It can explain what a “full-ratchet” anti-dilution clause is and why it’s dangerous, but it cannot tell you whether it’s legally enforceable in your specific jurisdiction or advise you on the best counter-proposal. That requires a licensed attorney who understands the nuances of corporate law and has a duty of care to you.

Second, AI has no memory of your specific situation. It doesn’t know the verbal promises an investor made over coffee, the history of your relationship with a co-founder, or the strategic goals you have for the company beyond this round. It analyzes the text in a vacuum, completely disconnected from the human context of the deal. A real advisor would ask, “What’s your relationship with this investor? Are they known for being supportive or aggressive in board meetings?”

Finally, AI cannot predict how a court will interpret a clause. Legal language is famously ambiguous, and its interpretation often depends on precedent, jurisdiction, and the specific judge. AI can flag a non-standard clause, but it can’t predict the outcome of a future dispute based on that clause. It’s a pattern-matching engine, not a prophet.

The “Sanity Check” Workflow: A 3-Step Founder’s Framework

To harness the power of AI without falling into its traps, you need a disciplined workflow. This isn’t about replacing your lawyer; it’s about supercharging your preparation so you get maximum value from every billable hour.

1. AI-Powered Initial Review & Education: The first pass is for you and the AI. Upload the term sheet and use prompts like the “Plain English Translator” to break down complex jargon. Your goal here is comprehension. You should walk away understanding every single clause at a high level and having flagged anything that seems unusual or concerning. This step turns a confusing 10-page document into a clear, understandable framework.

2. Strategic Preparation & Question Formulation: Now, you translate the AI’s output into a strategic agenda. Don’t just forward the AI summary to your lawyer. Instead, use the insights to prepare a prioritized list of questions and negotiation points. For example, the AI might flag a 2x participating liquidation preference. Your job is to formulate the strategic question: “I see the investor is asking for 2x participating preferred. Given our strong traction and the current market, is this a standard request for a Series A, or is this a point we should push back on to preserve our common stock value?”

3. Final Review & Official Counsel: With your prepared agenda, you now engage your lawyer. You’re no longer a passive client asking them to “just review this.” You are an informed partner driving a strategic conversation. You can say, “The AI flagged three major issues: the liquidation preference, the board control clause, and the full-ratchet anti-dilution. I’ve done some research, and I believe these are aggressive. What’s our best counter-strategy for each?” This approach saves your lawyer’s time (and your money) and empowers them to provide higher-level strategic advice instead of just basic document explanation.

Red Flags That Demand Immediate Legal Counsel

While AI is great for a first pass, certain terms are so complex or risky that they should immediately trigger a call to your lawyer. These are non-negotiables where the nuances can have devastating, long-term consequences.

• Complex Founder Vesting Schedules: If the term sheet proposes anything other than a standard 4-year vesting with a 1-year cliff, especially if it includes a “single trigger” acceleration or resets your vesting clock to zero, stop. This directly impacts your personal ownership and can be incredibly difficult to untangle later.
• Personal Guarantees: Never sign a term sheet or related documents that include a personal guarantee. Your liability in the company should be limited to your equity. A personal guarantee puts your personal assets—your house, your savings—on the line. This is an absolute red line.
• Ambiguous or Overreaching IP Clauses: Any clause related to intellectual property needs meticulous review. This includes “IP assignment” language that could claim ownership of work you did before founding the company, or any vague terms about “improvements” to IP. Your company’s core IP is its most valuable asset; its ownership must be crystal clear.
• “Full-Ratchet” Anti-Dilution: As mentioned in the prompt library, this is a deal-killer term. It’s so punitive that it signals a fundamental lack of trust from the investor. AI can flag it, but you need a lawyer to explain the full, devastating implications and help you negotiate it out or walk away.
• Vague or Unusual “Founder-Friendly” Clauses: Be wary of terms that sound good on the surface but are poorly defined. For example, a “founder’s right to a reasonable salary” is a trap. What’s “reasonable”? Without a clear definition, it’s a potential source of conflict with the board. Your lawyer needs to define this with precision.

Your AI is a tireless analyst that works for you 24/7. Use it to educate yourself, prepare your strategy, and sharpen your questions. But when it comes to the final, binding agreement that will define your company’s future, your lawyer remains your most critical partner. The AI hands you the map; the lawyer helps you navigate the treacherous terrain.

Conclusion: Empowering Founders with AI-Powered Diligence

The fundraising landscape has fundamentally shifted. The days of founders passively accepting term sheets and hoping for the best are over. By integrating AI into your diligence process, you’ve gained a strategic advantage that was once the exclusive domain of well-funded VCs with large associate teams. You’ve seen how a simple prompt can transform a dense, 20-page legal document into a clear, actionable summary of risks and opportunities.

The key benefits are tangible and immediate:

• Speed: What used to take a lawyer days to summarize, you can now understand in minutes.
• Cost Savings: You enter legal discussions prepared, focusing your lawyer’s time on strategic negotiation rather than basic education, saving you thousands in billable hours.
• Enhanced Understanding: You move from being intimidated by terms like “full ratchet” to confidently modeling their impact on your future ownership.
• Improved Leverage: When you can pinpoint a non-standard liquidation preference and ask, “Can you help me understand the rationale behind this deviation from the market standard?” you shift the power dynamic. You’re no longer just a petitioner; you’re a well-informed partner.

The future of fundraising is augmented. AI isn’t replacing the essential role of experienced lawyers or VCs; it’s democratizing access to the high-level analytical capabilities that empower you to engage with them on equal footing.

This isn’t just about spotting red flags; it’s about building a stronger, more resilient company from day one. The knowledge you’ve gained is your most powerful asset. Don’t try to implement every advanced technique at once. Start with the “Plain English Translator” prompt. Get comfortable with it. Build your confidence. As you do, you’ll naturally begin asking sharper questions and integrating the more advanced scenario-modeling techniques.

Remember, in a fundraising round, knowledge is power. You now have the tools to wield that power wisely. Use it to build the strongest possible foundation for your company’s future.

Expert Insight

Frequently Asked Questions

Q: Can Claude’s analysis replace a startup lawyer

No, Claude is a powerful first-pass tool designed to translate complex legal jargon and identify potential red flags, but it cannot replace the strategic advice and fiduciary duty of a qualified attorney

Q: How do I paste a term sheet into Claude securely

Always copy and paste the text directly into the chat, avoiding file uploads if possible. For highly sensitive documents, use the API with proper data handling policies or an enterprise account, and never use public, non-enterprise versions for confidential information

Q: What is the best prompt for understanding liquidation preferences

A highly effective prompt is: ‘Explain the liquidation preference in this term sheet as if I were a non-lawyer founder. Detail the payout order and provide a simple numerical example based on a $10M acquisition scenario.’