How to Draft a Demand Letter With AI: In-House Guide

To figure out how to draft a demand letter with AI, start with the business decision the letter needs to support. The company may want payment, a corrected performance, or a negotiated exit. That choice determines the facts, legal position, remedy, tone, and deadline the draft needs.

A demand letter is a formal written request that states the claim, the relief the sender seeks, and a deadline for responding. In-house teams use one to move a documented dispute toward a business decision, whether that means payment, performance, or escalation.

A useful AI draft connects the business record to a request counsel can verify.

GC AI is our enterprise legal AI platform for in-house teams. As of October 2026, 2,200+ legal teams use it, including 300+ public companies.

GC AI helps legal teams turn those records into a draft they can inspect against the source. Counsel still sets the objective, tests the legal position, and approves the letter.

In our Intro to AI for In-House Legal class, legal professionals practice this workflow. They set the objective, give the AI the right records, and review the draft against the record before it goes out.

During the class, Cecilia Ziniti, our CEO and a three-time general counsel, described a customer team that turns a recurring escalation into a starting draft from a Zendesk ticket:

One of our customers has a team that handles a certain type of escalation, and they write basically demand letters. They get the Zendesk ticket from it, and they get essentially a great letter that they can start from. And it used to be about a three-hour task, and now it's about a 10-minute task.

The example shows how an intake record can start the drafting process. Counsel still reconciles the facts, chooses the remedy, and approves the outbound letter. The useful handoff is clear: business records become a decision-ready first draft, and attorney review keeps the company from sending a claim it cannot support.

The sections below show how to make that handoff work, from setting the objective and four inputs to checking sources, adapting the prompt, and approving the letter.

Set the Objective and Prepare Four Inputs for an AI Demand Letter

Before asking AI to draft, agree with the business owner on what the letter should accomplish. Then separate confirmed facts from disputed facts and open questions. Prepare these four inputs:

1. Start with the facts and supporting records: Write a short chronology, identify the parties and amounts, and attach the contract, invoices, SOWs, correspondence, account records, and business timeline that support each material fact. Include adverse facts, credits, disputed performance, and open questions.
2. Establish the applicable law and process: Identify the governing law clause, forum or arbitration clause, notice requirements, delivery method, and any contractual cure period. If the applicable law remains unsettled, ask the AI to flag the question for counsel instead of choosing a default rule.
3. State the legal position: Name the claim counsel is prepared to assert, the contract provision or statute that supports it, the elements that must be shown, and material defenses or weaknesses. If the legal theory is unsettled, request a separate analysis before asking for an outbound letter.
4. Set the outcome and communication instructions: Choose the relief, amount or calculation, response date, recipient, delivery method, tone, and approved next step. Keep settlement limits and negotiation strategy out of the proposed letter, and include only matter information your organization's policy permits the AI to see.

These inputs control the draft's boundaries. A missing jurisdiction can lead to generic law, an unreconciled balance can create false precision, and an unstated remedy can produce a letter that asks for options the business has not approved.

Draft and Check a Demand Letter in GC AI

Once the objective and record are ready, use GC AI to turn them into a draft you can inspect against the source documents.

Consider a hypothetical commercial dispute: a business team asks legal to send a demand after a vendor misses a contractual obligation. Operations supplies the agreement, SOW, acceptance records, correspondence, and a timeline.

Before prompting, counsel confirms the obligation, the performance gap, any contrary account from the vendor, the notice process, and the outcome the business will support.

Before uploading records, confirm that your organization's policy permits the proposed use and that the access and data-handling settings suit the matter.

Our security information describes the relevant controls. We are SOC 2 Type II and SOC 3 certified, GDPR compliant, and encrypt data with AES-256. Our AI providers do not train on your data, and we maintain zero-data-retention agreements with LLM providers wherever feasible. For the full list of third parties that process Customer Data, see our public subprocessor list.

In GC AI, place the agreement, SOW, correspondence, timeline, and related records in a Files collection for the matter. Keeping the sources together gives counsel one place to check the facts, contract language, calculations, and final draft.

Three GC AI Features for the Demand Letter Drafting Workflow

The three features below each answer a different review question:

• Easy Prompt: This feature turns a plain-language description of the dispute and desired output into a structured legal prompt. Use it to organize the matter and expose missing instructions before you request the draft.
• Exact Quote: This feature lets you inspect verbatim language from a source document. Use it to compare the draft's description of an obligation, payment term, cure period, or notice requirement with the contract itself.
• US Case Law: This feature links cited cases to the full opinion. When the legal position depends on authority, read the holding and confirm that it applies to the jurisdiction and proposition in the letter.

The figures provide context about adoption. The quality of a demand letter still turns on the matter record and counsel's review.

Example: A General Counsel Tests the Workflow

Imagine a general counsel at a mid-sized consumer platform handling a vendor dispute. The business needs a clear answer about what obligation was missed, what records prove it, and what outcome it will support.

GC AI can map each alleged failure to a supporting document before it drafts. It can flag contrary evidence and open questions for counsel to resolve.

The general counsel then checks the draft against the agreement, the timeline, and the business objective.

The lawyer starts with a source-backed draft and spends time on the legal judgment, remedy, and next step.

To try this workflow in GC AI, start your 14-day trial. Then adapt the prompt to the dispute in front of you using the structures below.

Adapt the AI Demand Letter Prompt to Your Dispute

The structure stays consistent across these examples. The evidence and legal decisions change. Treat the prompts as starting points, replace the brackets, remove instructions that do not fit, and ask the AI to surface questions before it drafts.

Use the supplied records and counsel's instructions below.

First identify material gaps, conflicting evidence, and unresolved legal questions. If a material fact is missing or disputed, flag it in review notes and ask for clarification before treating it as settled.

Do not invent facts, dates, amounts, quotations, contract terms, statutes, or cases. Keep proposed changes to the legal theory or remedy in separate review notes.

Return:

1. A proposed demand letter stating the parties, supported facts, approved legal position, requested relief, response instructions, and approved next step.

2. Separate internal review notes identifying the source for each material assertion by document and location, any calculations, contrary evidence, and unresolved issues.

Keep internal strategy, settlement limits, and review notes out of the proposed letter. Include legal authorities in the letter only when counsel requests them and verifies their relevance.

Unpaid Invoice or Overdue Payment

Draft a demand letter for an unpaid invoice using the attached contract, invoice records, and correspondence.

Parties and agreement: [Company] and [Counterparty] under [contract or PO].

Facts and records: invoices [numbers] issued [dates], due [dates]; payments, credits, disputes, and follow-up communications appear in [source names].

Applicable law and process: [governing law], [forum or arbitration], and [notice requirements].

Legal position: breach of contract for nonpayment under [section], with [interest or fees] only if the contract and applicable law support them.

Requested outcome: [counsel-approved amount or remedy] by [counsel-selected response date].

Communication: [recipient, delivery method, tone, approved next step].

Review notes: reconcile the balance, identify contrary evidence, and flag any missing record before drafting.

The accounting line makes the amount auditable. It also gives the reviewer a place to check payments, credits, disputed charges, interest, and fees before the letter states a number.

Breach of Contract

Draft a demand letter for a contract breach using the attached agreement, SOW, performance records, and correspondence.

Obligation: [specific promise] in Section [X], due or measured by [date or standard].

Failure: [specific performance gap], supported by [source names].

Evidence and contrary facts: [acceptance record, notice, excuse, cure, or dispute].

Applicable law and process: [governing law], [forum or arbitration], [notice and cure requirements].

Legal position: [claim counsel approves] and the elements or contract language that support it.

Requested outcome: [counsel-approved cure, payment, service credit, termination, or other relief] by [counsel-selected response date].

Communication: [recipient, delivery method, tone, approved next step].

Review notes: identify missing evidence, defenses, and any remedy that needs separate legal analysis.

The obligation and cure lines keep the draft tied to the contract's language and the process the contract requires. The reviewer can then decide whether the record supports a default, a cure demand, or a different remedy.

Trademark or IP Demand

Draft a demand letter concerning potential trademark infringement using the attached registration records, examples of use, and correspondence.

Rights and use: [mark], owned by [Company], with [registration status and number if any], used for [goods/services] since [date].

Respondent's use: [similar mark or material], used for [goods/services], first observed on [date] at [URL or location].

Applicable law and process: [authority counsel selects], [forum], and [notice or response requirements].

Legal position: assess the supported infringement and likelihood-of-confusion theories from the supplied facts. Separate factual gaps and legal research from the proposed letter.

Requested outcome: [counsel-approved cease-use, removal, written confirmation, or other relief] by [counsel-selected response date].

Communication: [recipient, delivery method, tone, approved next step].

Review notes: distinguish registered rights, common-law rights, defenses, and remedies supported by the record.

For a trademark matter, state the registration status accurately and provide facts about the parties' goods, services, and uses. The Lanham Act may be relevant, but registration status alone does not resolve likelihood of confusion or the available remedy.

Vendor Nonperformance

Draft a demand letter for vendor nonperformance under the attached MSA or SOW.

Obligation and acceptance: [deliverable or service] due [date] under Section [X], measured by [acceptance criteria or SLA].

Failure and record: [missed deadline or failed requirement], supported by [source names], with [vendor response or disputed facts].

Business objective: [cure, service credit, termination, renegotiation, or other approved outcome].

Applicable law and process: [governing law], [forum or arbitration], [notice and cure requirements].

Legal position: [claim counsel approves] under [contract section or authority].

Requested outcome: [specific relief] by [counsel-selected response date].

Communication: [recipient, delivery method, tone, approved next step].

Review notes: check dependencies, change orders, limitation of liability, exclusions, cure rights, and termination conditions.

For vendor disputes, the business objective determines whether the letter should seek a cure, a credit, termination, or another approved outcome. The contract's limits and conditions determine which remedy the letter can support.

Review the Letter Before Sending

Review the draft against the company's evidence, the counterparty's likely response, and the business objective. A polished letter can still overstate the facts, demand an unavailable remedy, or miss a notice requirement.

Unsupported Facts, Amounts, or Dates

An AI can turn an incomplete record into a clean sentence. Check the chronology, account balance, performance record, calculations, and every factual assertion against the source document. Include contrary evidence in the review, even when the business team's preferred account is clear.

Incorrect or Misused Authority

In Mata v. Avianca, the court sanctioned lawyers after a court filing included fabricated case citations generated by an AI tool. Apply that verification discipline to a demand letter: open each cited case, read the holding, and check the proposition, relevant facts, procedural posture, and treatment. US Case Law can make the cited opinion easier to inspect in GC AI.

Wrong Law or Notice Requirements

A demand letter that cites the wrong statute of limitations, interest rule, or notice requirement can weaken an otherwise supported claim. Check the governing-law clause, forum or arbitration provision, contract notice method, delivery record, response deadline, and any law that controls the remedy. A governing-law clause answers one part of the question. Counsel also checks the forum, notice, delivery, deadlines, and remedy rules.

Unsupported Remedy or Business Position

A draft can select a remedy the contract restricts, overlook a cure period, or state a position the business has not approved. Check the elements and defenses, limitation of liability, exclusions, termination rights, calculations, settlement posture, and the next step the company will take if the recipient responds or stays silent.

Attorney Review Checklist Before You Send

Before approval, confirm:

1. The sender, recipient, entity names, and delivery method are correct.
2. Each material fact, amount, and date traces to a source, with contrary evidence recorded.
3. Each legal claim and authority supports the proposition stated in the letter.
4. The requested amount, interest, fees, or other relief matches the records and calculations.
5. The contract's caps, exclusions, cure rights, notice terms, and remedy restrictions have been checked.
6. The response date and delivery method comply with the contract and applicable law.
7. The business objective, requested remedy, and next step have approval from the responsible business owner and legal reviewer.
8. Internal strategy, settlement limits, privileged analysis, and confidential material stay out of the outbound letter.
9. Assign an owner for the delivery record, response tracking, and any separate filing or contractual deadlines.

The AI draft saves blank-page time. Counsel still owns the legal position, commercial objective, and final approval.

Learn AI for Legal Work With GC AI Courses

The demand-letter workflow is one example of a broader legal AI skill. Our courses teach it through live legal examples and exercises.

• Intro to AI for In-House Legal: Our 75-minute live class is designed for legal professionals with no prior AI knowledge.
• Advanced AI for In-House Legal: Our 60-minute live class is for legal professionals who already have some AI familiarity.
• Custom AI Class for Legal Teams: Our 60-minute private course adapts the content and workflow exercises to your team's needs.

After class, test the workflow on a completed matter your organization permits you to use.

Compare the AI draft with the approved letter, record each correction, and turn recurring corrections into instructions your team can reuse. For the broader application, read our guide to using AI for legal document drafting.