How to Write a Pre-Approval Letter with AI in 2026

A strong mortgage pre-approval letter does three things: it gives the buyer a credible commitment they can use in offers, it accurately reflects what the lender has actually pre-approved (no more, no less), and it includes the conditions and expiration that protect both the lender and the borrower if circumstances change. Writing them by hand for every borrower is repetitive structured work that has real compliance edges — overpromising in a pre-approval letter is the kind of thing that surfaces months later when the loan can't close at the terms the letter implied. AI is excellent at producing the structural part in two minutes. The pre-approval decision itself comes from your LOS through your lender's approval workflow — never from the AI.

This is a practical walkthrough for writing a mortgage pre-approval letter with AI that holds up when the buyer's agent reviews it.

What a strong pre-approval letter contains

Before you can use AI well, you need to know what good looks like:

• Header block — lender name, NMLS company ID, your NMLS ID, license states, contact info, letter date
• Borrower identification — borrower name (and co-borrower if applicable)
• Pre-approval amount — the maximum loan amount the borrower is pre-approved for at current terms
• Loan product — specific product (Conventional 30-year fixed, FHA, VA, jumbo), with key terms (LTV, DTI thresholds satisfied)
• Property scope — whether the pre-approval is for a specific property or general purchase amount
• Documentation reviewed — credit, income, asset, employment verification status (without revealing borrower-protected specifics)
• Conditions — what still needs to happen before final approval (appraisal, title, final underwriting, satisfactory property selection)
• Expiration date — when the pre-approval expires (typically 60–120 days)
• Equal Housing Lender language — required disclosure
• Signature block — your name, NMLS ID, lender, contact

Loan officers whose pre-approval letters convert into closed loans are the ones whose letters accurately reflect the lender's actual position — not aspirational language that creates an expectation the loan can't meet. AI handles the structural and language layer; the pre-approval decision and the specific dollar amount come from your LOS.

The right prompt structure

The mistake most LOs make on first try is asking AI for "a pre-approval letter" with just a name and amount. The prompt that actually works gives the AI the lender-issued decision and the conditions:

<task>Write a mortgage pre-approval letter.</task>

<context>
Lender: ABC Mortgage Company, NMLS Company ID #12345
Loan officer: [LO NAME], NMLS ID #987654
License states: TX, OK, NM
Letter date: May 20, 2026

Borrower: [BORROWER NAME] (first + last; placeholder for actual letter)
Co-borrower: N/A

Pre-approved loan amount: up to $385,000
Loan product: Conventional 30-year fixed
LTV considered: up to 90% (10% minimum down assumed)
Term: 30-year fixed

Documentation reviewed (status):
- Credit: pulled and reviewed; meets product requirements
- Income: W-2 income verified through current employer (no income amounts disclosed in letter)
- Assets: down payment and reserve funds documented
- Employment: 2+ year employment history verified

Conditions for final approval:
- Satisfactory property appraisal at or above purchase price
- Clear title with acceptable title insurance
- Final underwriting approval upon receipt of fully-executed sales contract
- No material change to borrower's credit, income, or employment between
  pre-approval and closing
- Property meets product eligibility requirements (e.g., condo approval if applicable)

Property scope: General pre-approval — borrower has not identified specific
property. Pre-approval can be re-issued at specific property and price upon
contract acceptance.

Pre-approval valid through: August 18, 2026 (90 days from date of issuance)

Required disclosures:
- Equal Housing Lender (with logo or text disclosure)
- Lender NMLS ID
- LO NMLS ID
</context>

<instructions>
- Tone: professional, lender-neutral, not warm or congratulatory
- Sections: borrower identification, pre-approved amount and product, documentation reviewed status, conditions for final approval, expiration, signature block with NMLS IDs
- Use placeholders [BORROWER NAME], [LO NAME], [LO TITLE], [DATE]
- Equal Housing Lender language in footer
- 350 words maximum
</instructions>

<avoid>
- Stating specific borrower income, employment details, or credit score
- Implying the loan is approved (it's pre-approved with conditions)
- Promising specific interest rates or APR (rates are not locked at pre-approval)
- Open-ended pre-approval ("good until further notice") — use specific expiration
- Language that could be read as a binding commitment without the stated conditions met
- Generic congratulatory or sales language
</avoid>

The structure: the lender's actual decision, the conditions, the expiration, and explicit instructions about what NOT to include (no rates, no income specifics, no over-commitment). The AI produces the letter; the LO and the LOS produce the underlying decision.

What to never let AI do

State the pre-approval amount. The maximum loan amount comes from your LOS based on the actual borrower file. AI doesn't have access to the file. Provide the amount; the AI puts it in the letter.

Imply the loan is approved. A pre-approval is conditional. Letters that read like the loan is approved create expectations the loan can't meet if conditions aren't satisfied. The conditions section is mandatory.

Quote an interest rate or APR. Rates aren't locked at pre-approval. Letters that quote specific rates create exposure if rates move between pre-approval and lock. Most lender compliance teams prohibit rate quoting in pre-approval letters; the AI should follow.

Reveal borrower-protected specifics. The buyer's agent and listing agent are seeing this letter. Specifics about borrower income, employment, or credit score don't belong in the letter — they're irrelevant to the agent's purpose and they reveal information that should stay between the borrower and the lender.

Make commitments beyond what the LOS authorized. "We are also able to consider higher amounts upon request" goes beyond the actual pre-approval. Stick to what the LOS approved. Higher amounts get a new pre-approval after the LOS re-evaluates.

Common mistakes

No expiration date. Pre-approval letters without expiration dates create stale-data risk. Use a specific expiration (60–120 days is standard).

Missing conditions. Conditions are what make pre-approval "pre" and not "final." Letters without explicit conditions read like binding commitments — which exposes the lender if the loan can't close.

Stating specific borrower information. Income, employer, credit score, social — none of these should be in the letter. The letter says "documentation reviewed and meets product requirements," not "borrower earns $X at Y company."

Wrong product on the letter. If the borrower is pre-approved for Conventional but the agent's offer requires FHA, the letter needs to reflect what's actually been approved. Don't change product types on the letter unless the LOS has re-run the underwriting.

Missing NMLS IDs or EHL disclosure. These are compliance requirements. NMLS Company ID + LO NMLS ID + Equal Housing Lender disclosure go on every letter.

What to never put in a pre-approval letter

• A binding commitment to lend specific terms without the stated conditions met
• An interest rate, APR, or monthly payment estimate
• Borrower-protected information (income, credit score, employer specifics)
• Open-ended validity ("until you find a home")
• Language that contradicts your lender's actual conditions

These aren't AI-specific risks — they apply to any pre-approval letter. AI can produce them quickly without flagging the risk; the LO's compliance review step is where they get caught.

The free tool that handles this for you

If you don't want to engineer the prompt every time, the Pre-Approval Letter Drafter on AI Career Lab is pre-configured for the structure that holds up to agent and underwriter review. It produces letters with the elements above, in the professional lender-neutral tone that satisfies compliance expectations when paired with your LOS-issued decision.

Pair it with the Rate Comparison Sheet Generator for the borrower-education layer, the Refinance Analysis Generator for the refi pipeline, and the Client Update Generator for borrower communication through underwriting.

Free with an AI Career Lab account, capped at five runs per day on the free tier.

Working as a loan officer rather than a broker?

If you're an NMLS-licensed LO at a single lender (vs an independent broker), see our Loan Officer profession hub for LO-specific workflows — multi-audience pipeline updates, LOE drafting, LE/CD explainers, and realtor co-marketing with TRID/ECOA/RESPA guardrails. The Loan Officer AI Cowork Vault ($19, one-time) ships 52 LO-specific skills for Claude Cowork and Microsoft 365 Copilot Cowork.

Try it on your next borrower

Pick the next pre-approval you need to issue this week. Get the LOS-approved amount and product. Run the inputs through the tool above. Compare to the letter you'd issue by default — note the difference in clarity of conditions and absence of over-commitment.

Create your free AI Career Lab account and try the mortgage tools today. No credit card.

---

This article is general guidance for NMLS-licensed mortgage brokers and loan officers. AI-generated pre-approval letters are starting drafts requiring LO review for accuracy of pre-approval amount, product, conditions, expiration, and required disclosures. TRID, TILA, RESPA, ECOA, state-specific licensing rules, NMLS requirements, and your lender's compliance policies govern actual practice. The pre-approval decision itself comes from the lender's LOS, not from AI.