The underwriter who counts the defects everyone argues about
$211,350top of the range in New York · middle $76,690 / yr
AI is transforming this role
Underwriters in the United States earn a median of $76,690 a year. Pay starts near $39,430. Pay reaches $211,350 at the top of the range in New York, the best-paying state for this work among those with at least 500 people in the job.
Source: U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025 (Loan Officers, SOC 13-2072). Last checked 9 September 2026.
Entry level
$39,430
Top of the range · New York
$211,350
Education
Bachelor's degree in Finance
Wages — U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025 (Loan Officers). Top of the range is the highest state-level figure among states with at least 500 people in the job. AI-impact rating is PayCrunch's editorial assessment. Updated September 2026.
🆕 New & Trending AI Tools for UnderwriterReviewed September 2026
We track new AI-tool launches every week and refresh this list — here’s what’s gaining traction for Underwriter work right now.
RunwayNEWFree tier / $12 mo
AI video generator and editor for short cinematic clips.
How an Underwriter uses it: generate and edit video b-roll and effects without a full production
GammaNEWFree / $9 mo
Generates polished slide decks and one-pagers from a prompt.
How an Underwriter uses it: turn an outline into a designed presentation instantly
NotebookLMNEWFree / $7.99 mo
Google tool that answers questions grounded only in the documents you give it — with citations.
How an Underwriter uses it: load your own manuals, policies, or PDFs and ask questions that stay accurate to the source
Canva AIFree / $13 mo
Design platform with AI text-to-image, writing, and one-click layouts.
How an Underwriter uses it: produce on-brand graphics, social posts, and decks without a designer
Adobe FireflyFree credits / paid
Adobe's commercially-safe AI image and video generation, built into Creative Cloud.
How an Underwriter uses it: generate and edit images and video safe for commercial use
Midjourney$10+ mo
High-end AI image generator known for striking visuals.
How an Underwriter uses it: create original concept art, mockups, and hero images from a prompt
DescriptFree / $16 mo
Edit video and podcasts by editing the transcript like a doc.
How an Underwriter uses it: cut and polish video/audio by editing text, and remove filler words automatically
ElevenLabsFree / $5+ mo
AI voice generation with hundreds of natural voices.
How an Underwriter uses it: produce voiceovers and narration in minutes
ChatGPTFree / $20 mo
The most-used AI assistant — writing, analysis, research, and images from a plain-language chat.
How an Underwriter uses it: draft emails and documents, summarize long files, and get instant answers to on-the-job questions
The file hits your queue with a name, a requested amount, and a stack of documents that may or may not agree with each other. Income on the application has to meet income on the pay records. The credit history has to be read, not glanced at. Collateral, if any, has to be the collateral the policy accepts. An underwriter's job is to decide whether this request fits the lender's rules, to set the conditions that would make it fit, or to decline it in language a salesperson and a borrower can understand. The pressure is always toward yes. The signature is yours when the file was a no. That tension is the occupation.
This page is about credit underwriting: consumer loans, mortgages, and commercial credits, depending on the desk. Finding borrowers and taking applications is a different day, even though a wage survey may group that day with yours. You may talk to the people who do that work. You may send a file back because a document is missing. You do not "make the deal work" by ignoring a fact the policy treats as fatal. Lenders who want that kind of help are asking you to stop being an underwriter. The career worth having is the one where a clean no is as respectable as a clean yes.
What you actually decide
A consumer desk moves fast. Auto loans, personal loans, and credit cards arrive in piles, and a policy has already sorted the easy ones. Your time goes to the exceptions: a thin file, a recent delinquency, income that does not line up, an identity detail that needs a second look. You approve, you counter, or you decline, and you write the reason in the code the system and the regulation expect. Speed matters. Sloppiness matters more. A fast wrong yes is how a portfolio rots. A fast wrong no is how a fair applicant gets a letter the lender cannot defend.
A mortgage desk is slower and more documentary. The property, the occupancy the borrower claims, the income, the assets, and the program the loan is supposed to fit all have to be tied to papers. Conditions are the ordinary tool: you will approve if this document arrives and says what the file implies. You are not a coach for stretching a story. If the documents and the application disagree, the disagreement is the finding. Escalation to a senior underwriter or a credit committee is a feature of the job, not a failure. The cases that should reach that table are the ones the policy did not fully anticipate, not the ones you hoped nobody would read.
Commercial underwriting looks at a business, not only a person. Financial statements, cash flow, guarantors, and the purpose of the loan sit beside the policy for that product. You may visit nothing and still need to understand how the business gets paid. A small-business request and a large corporate credit are different depths of the same instinct. Write so a credit officer who was not in the room can see why you recommended the line. Memory of a phone call is not a credit file.
Across all three desks, the tools are the policy, the documents, and a system of record. Your notes should let an auditor, a capital-markets buyer, or your own manager replay the decision. If you would be uncomfortable reading your note aloud to a compliance reviewer, rewrite it before you approve. That habit feels slow in the first month and protective for the rest of a career. Underwriters are paid for judgment that can be shown.
Mortgage desks and NMLS registration
No single government licence is required in every state for every kind of underwriting. A bank's internal credit authority is often the permission that matters on a consumer or commercial desk. Mortgage work is the exception people must ask about directly. Mortgage work may require NMLS registration through the state. NMLS is the Nationwide Multistate Licensing System, the system states use for mortgage licensing and registration. Whether your particular seat must be registered depends on the state and on whether you are performing activities that state covers. Ask the employer, and check the system, before you assume a back-office title exempts you.
The official home for that system is nationwidelicensingsystem.org. Registration, where it applies, shows that you are known to the state in the mortgage channel and that you met the conditions that state set for the people it tracks. It does not prove you are a skilled underwriter. A clean file proves that, over time. People prepare by working in mortgage operations or underwriting and by completing whatever education and background steps the state and the system list for the registration they need. Read the current instructions there. Do not rely on a coworker's memory of how it worked at a prior lender.
Ask which desk before you celebrate the title
Consumer, mortgage, and commercial underwriting share a title and not a rulebook. If the loans are mortgages, ask whether NMLS registration through the state applies to your seat. If the loans are not mortgages, do not spend the interview pretending a mortgage registration is the center of the job.
How lenders choose who may sign
Hiring managers want evidence you can say no. The resume that works names the products, the authority you held, and the kinds of exceptions you referred upward. "Reviewed loans" is weak. "Underwrote conventional mortgages to a published program, with authority to approve within policy and a requirement to refer exceptions" is a picture. If you came from processing, servicing, or credit analysis, say which pieces you already owned: conditions, income calculation, or the memo a committee read. Own the piece. Do not borrow the closer's signature.
Interviews use files, sometimes sanitized, sometimes described aloud. Walk the decision in order. What did you read first. What conflicted. What condition would fix it. What would make you decline. What would make you take it to a senior. Mention policy by how you used it, not by performing a speech about being tough. Managers also listen for how you treat the salesperson. Contempt for the people who bring loans is a culture problem. Capitulation to them is a credit problem. The workable answer respects both the relationship and the rule.
Volume shops and relationship shops hire for different temperaments. A consumer queue rewards consistency and calm speed. A commercial or private-bank desk rewards a longer memo and the ability to sit with a borrower's story without being captured by it. Ask how many files a full day holds, who covers you when you are out, and how quality is checked after you approve. A desk with no review is not a compliment. It is a place where your name will stand alone when a loan fails. Know that before you take the chair.
Junior file, senior credit, then a policy seat
The path often starts as a junior underwriter, a credit analyst, or a processor who moves into decisions. Juniors work inside a tight authority and a thick review. You graduate when your conditions are clean and your declines hold up. Senior underwriter means harder files, mentoring, and a larger share of the exceptions. Credit manager or chief credit officer is a different altitude: portfolio trends, policy changes, and the argument with sales leadership about what the lender is willing to be. Some people prefer to stay senior and excellent on files. Say so. A lender that only promotes managers will otherwise read your preference as a lack of drive.
Moves between consumer, mortgage, and commercial are possible and costly. The instinct transfers. The documents do not. A mortgage underwriter who wants commercial credit should expect to be junior again for a while. A commercial analyst who wants a mortgage desk should expect the registration question and a different pace. Side doors include quality control, where you review files after the decision, and investor delivery, where a buyer's conditions come back to the lender. Both teach you how a yes can still fail. That lesson makes you a better underwriter if you return to the queue.
Careers stall when approvals are casual, when conditions are waived in a chat message, or when you become the person sales calls because you can be leaned on. Careers move when your files are boring to audit and your exceptions are well chosen. Keep a record of products, authority levels, and any registration you hold, with the state named if a state issued it. Leave customer names out. The next lender needs the shape of your judgment, not a borrower's private story.
Pay in the loan-officer survey, read for this desk
These wages are Occupational Employment and Wage Statistics for May 2025. The survey bucket is named once, here: Loan Officers. This page is credit underwriting. The dollars come from that broader lending series, so they mix people who originate loans with people who decide them. Use the figures as the published map, and keep the job you actually do in the same sentence as the title of the series.
Early-career pay in the release sits at $39,430. The national midpoint is $76,690. The distance from the early figure to the midpoint is $37,260. A junior underwriter still inside heavy review can use $39,430 as the lower reference. An underwriter who already approves within policy has a reason to look toward $76,690 and to ask why the offer has not moved with the authority. From the national midpoint to the high end of the published range is $134,660. That span is the width of the range. It is a poor opening demand for a first full authority.
New York holds the high end of the published range, at $211,350. Massachusetts holds the highest median, at $101,600. Those are different statistics, and they belong to different states. The New York figure of $211,350 is the top of the published range. The Massachusetts figure of $101,600 is a median, the middle of pay in that state. New York also has a median, and that median is a different statistic from the New York high end. From the national midpoint to the Massachusetts median is $24,910. Use $24,910 when you mean the lift to the highest state middle. Use $211,350 only when you mean the New York high end, labeled as the high end.
State medians, in this order, are Massachusetts at $101,600, Connecticut at $95,730, New York at $95,710, Minnesota at $95,170, and Colorado at $94,520. Each is a middle. Massachusetts leads. Connecticut at $95,730 and New York at $95,710 sit near each other, and New York's median must not be swapped for New York's high end of $211,350. Minnesota at $95,170 and Colorado at $94,520 continue the list, all above the national midpoint of $76,690, all still medians. None of them is the high end. Keep this order when you lay the states out. Massachusetts first is the highest middle, not an invitation to start the story in New York because the high end lives there.
The lowest median in the release is Puerto Rico at $35,730. The gap between that median and the Massachusetts median is $65,870. The gap compares two middles. It leaves the New York high end as a separate fact. Puerto Rico's $35,730 sits below the national entry figure of $39,430. Read a Puerto Rico offer against $35,730. Read Colorado against $94,520. Read Minnesota against $95,170. Read New York's middle against $95,710. Read Connecticut against $95,730. Read Massachusetts against $101,600. Read $211,350 only as the high end of the published range in New York.
Bring the statistic that matches the chair
Put the offer beside the right anchor. A trainee seat can sit near $39,430 if someone else still signs. A working underwriter can put $76,690 on the table as the national middle and then name the state. In Massachusetts the middle is $101,600, which is $24,910 above the national midpoint. In Connecticut it is $95,730. In New York the middle is $95,710, and the high end of $211,350 has to stay in another sentence. In Minnesota the middle is $95,170. In Colorado it is $94,520. Quoting Massachusetts because you like $101,600, while the desk is in Colorado, ends the meeting. Quoting New York's high end as if it were New York's typical pay ends it faster.
The high end sits $134,660 above the national midpoint. Mention $211,350 only when the seat, the market, and the record are genuinely about the top of the published range, and say that you are talking about a high end in New York, not about a median and not about Massachusetts. If a manager waves at $211,350 to suggest you are already overpaid at $76,690, separate the figures. One is a national middle for a broad lending series. The other is the top of the published range in one state. Your underwriting authority is a reason to discuss where you sit between $39,430 and the relevant median. It is not a reason to invent a third New York number between $95,710 and $211,350.
Mortgage registration, a commercial book, or a lead role can explain a place inside the range. They cannot replace the map. Write the offer, write $76,690, write the one state median, and write whether anyone has confused New York's high end with Massachusetts' median or with New York's own median of $95,710. Then stop talking. An underwriter who can keep two statistics apart is demonstrating the same care the lender wants on a file. That resemblance is the close. The policy, after you are hired, will ask for it every day.
The top of Underwriter pay — and how to get there with AI
$211,350what Underwriter pay reaches in New York
Highest state-level top-of-range annual wage for Loan Officers, among states with at least 500 people in the job. U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025.
And the role it leads to — Personal Financial Advisors — reaches $459,050 in Oregon.
$39,430entry$76,690middle$211,350top end
Underwriters clustered in the middle of this range judge files one at a time; the ones at the top can tell you which conditions cause the most rework, how long each stage really takes, and what changed after they fixed it.
Reviewing loan agreements for completeness and accuracy against policy, compiling credit histories and corporate financial statements, computing payment schedules and sending applications on for verification are all measurable, and almost nowhere are they measured. Teams argue from anecdote about which processor is slow or which document keeps coming back wrong. A model can now read a stack of closed files and summarise the exceptions in them, which puts real quality data within reach of one determined person. Whoever produces that data stops being a set of hands on a queue and starts being the person management asks before changing a policy.
Your playbook, by where you are now
Just startingKnow the file better than the checklist does
Work the 1003 Uniform Residential Loan Application line by line until you can spot an inconsistency without the system flagging it.
Compute a payment schedule by hand once for each product type so the amortization loan software never surprises you.
Keep your own tally of why your files come back: missing income documentation, a stale credit report, a condition worded so the borrower could not act on it.
Learn what a credit analyst actually needs from you, and ask two of them what your submissions are missing.
What proves it: A month of your own files with a written reason attached to every exception.
Realistic span: the first eighteen months on the desk
A few years inTurn your tally into the team's numbers
Extend the exception log to the whole team with a fixed list of defect categories, since two people describing the same problem differently makes the data useless.
Build the reporting in Microsoft Excel or Microsoft Access against dates you can trust rather than dates somebody typed.
Use Excel Copilot to draft the pivot and the chart, then check the totals against the raw records yourself before anyone sees them.
Set up a Power Automate step so the log fills from the handoffs instead of from goodwill.
Present one monthly page: exceptions by category, rework hours, and where files sat waiting.
What proves it: A monthly file-quality report your manager circulates without editing.
Realistic span: years two through five
ExperiencedSet the policy the numbers justify
Rewrite the conditions that cause the most rework, in plain wording a borrower can act on the first time.
Run a quality review of closed files, sampled rather than complete, and report findings without naming individuals.
Keep abreast of new loan products and price the operational cost of each one before it launches, not after.
Have Claude summarise a policy bulletin into what changes for your desk, then verify each change against the bulletin itself.
Look at where the higher-paid work sits, since advising clients directly is the next step from this seat and New York pays this field the most.
What proves it: A policy or condition rewritten on your evidence, with the defect rate before and after.
Realistic span: six years and up
The next 90 days
For the next ninety days, log every file of yours that comes back, and log it the same way each time: what was wrong, who found it, which stage it returned to, and how long the fix took. Twenty files is enough to see a shape. You will almost certainly find that a small number of causes account for most of the rework, and that at least one of them is a condition your own team words badly. Write that up as one page, three categories, a count against each, and one suggested fix. Take it to your manager as a question rather than a complaint. Underwriters who bring the count instead of the anecdote get put on the process work, and process work is what separates a queue seat from the desk that decides how the queue is run.
Wage figures: BLS OEWS, May 2025. The playbook is PayCrunch editorial guidance, not a guarantee of pay or placement.
Every figure is the national median from the U.S. Bureau of Labor Statistics (OEWS) shown on that role’s own page.
Never used AI before? Start here (2 minutes).
Start with document analysis - it's where underwriting time disappears. Tools like Ocrolus read bank statements, pay stubs, and tax returns and turn them into verified income and cash-flow data with fraud flags, instead of you keying and cross-checking by hand. If your shop runs it (or Hyperscience), turn it on for your next file and verify its output against the source documents.
For learning and guideline questions (never borrower data), use ChatGPT or Perplexity to get fluent on agency rules, a loan program, or a credit concept in plain language. Keep every file with a borrower's name, SSN, or financials inside your lender's approved systems. AI is the analyst that preps the file; you are the underwriter who owns the credit decision.
The one rule, forever: Lending is governed by fair-lending law (ECOA, Fair Housing Act) and privacy law (GLBA). AI must never rely on prohibited factors or produce disparate impact, and every denial needs an accurate, specific adverse-action reason you can defend. Never paste borrower PII, SSNs, or financials into a consumer AI tool. Models are subject to model-risk governance (SR 11-7); you own the credit decision, not the algorithm.
The plays — exact steps, exact prompts
Do these in order. Each one is copy-paste ready. You do not need to know anything about AI going in.
1
Underwrite more files with AI income and document analysis
Why this pays: Underwriter productivity is files cleared per day at quality. AI that verifies income and assets from raw documents lets you decision more loans accurately - the throughput that earns higher authority and the top pay band.
OcrolusHyperscienceBlend
1
Use Ocrolus or Hyperscience to auto-extract and verify income, employment, and assets from pay stubs, bank statements, and tax returns - with anomaly and fraud flags - instead of manual calculation.
2
Sanity-check a complex income scenario.
Copy-paste this prompt
Walk me through how to calculate qualifying income for a self-employed borrower with these de-identified figures: [Schedule C net income], [add-backs for depreciation and depletion], [two-year trend]. Show the standard agency methodology step by step and note what documentation supports each add-back. General methodology only.
Use de-identified numbers only. Verify against Fannie Mae, Freddie Mac, or your investor guidelines - you sign off on the income calculation.
3
Reconcile every AI-extracted figure against the actual document before you rely on it.
What you'll haveMore files decisioned accurately per day - the productivity that earns authority and moves comp toward $211,350.
2
Spread and analyze commercial credit with AI
Why this pays: Commercial and complex credit pays far more than clearing conforming mortgages. AI that spreads financials and models cash flow lets you step up to commercial underwriting - the higher-value work behind a senior credit salary.
OakNorthMoody's CreditLensNumerated
1
Use OakNorth or Moody's CreditLens to auto-spread borrower financials and run scenario and stress analysis on a commercial credit instead of building spreads by hand.
2
Pressure-test the borrower's story.
Copy-paste this prompt
For a commercial loan to a [HVAC contractor] with these de-identified financials - [revenue], [EBITDA], [existing debt service], [requested loan] - calculate DSCR and leverage, identify the three biggest credit risks, and list the covenants and conditions a prudent credit officer would require. Analysis only, not an approval.
De-identify the borrower. The credit decision, structure, and covenants are yours to set within your institution's policy.
3
Use Numerated to streamline data collection and speed the commercial workflow front to back.
What you'll haveThe ability to underwrite complex commercial credit - the skill that lifts an underwriter into the top pay band.
3
Catch fraud and red flags before they become losses
Why this pays: Fraud losses and buybacks destroy an underwriter's reputation; catching them builds it. AI fraud-detection tools flag synthetic identities and doctored documents you'd miss - protecting the loan quality that earns trust and authority.
SentiLinkSocureOcrolus
1
Use SentiLink or Socure to screen for synthetic-identity and first-party fraud signals, and Ocrolus's document-integrity checks for altered statements.
2
Turn a red flag into an investigation checklist.
Copy-paste this prompt
An underwriting file shows these de-identified inconsistencies: [employer not verifiable], [deposit patterns inconsistent with stated income], [recently issued SSN]. List the specific fraud red flags these suggest, the additional documentation I should require, and the verification steps to resolve or decline. General fraud-review guidance only.
Never paste identifying data. AI flags are leads to investigate, not proof; follow your institution's fraud and adverse-action procedures.
3
Document every red flag and its resolution so the file is defensible on audit or repurchase review.
What you'll haveFewer fraud losses and buybacks - the clean loan quality that builds the reputation behind higher authority.
4
Master guidelines with instant AI lookup
Why this pays: Guideline errors cause conditions, delays, and buybacks. An AI you can query on agency and investor rules means fewer mistakes and faster clean decisions - the accuracy and speed that get you promoted.
NotebookLMChatGPTPerplexity
1
Load your investor and agency guideline PDFs into NotebookLM so you can ask questions and get answers cited to the exact section.
2
Resolve a guideline question fast.
Copy-paste this prompt
Explain the general agency requirements for [using rental income from a departing residence to qualify]: what documentation is required, how the income is calculated, and the common reasons this gets rejected in underwriting. Summarize the rule and tell me exactly which guideline section to verify. General education only.
AI can misstate or outdate a rule - always confirm against the current, primary agency or investor guideline before you condition or decision.
3
Keep a personal, AI-organized cheat sheet of the guideline traps you hit most often.
What you'll haveFewer guideline errors and faster clean approvals - the reliability that earns higher authority and pay.
5
Specialize in complex or non-QM credit
Why this pays: Jumbo, non-QM, self-employed, and construction lending pay premiums because they demand judgment automated engines lack. AI accelerates the expertise that lets you own this higher-margin niche.
NotebookLMChatGPTPerplexity
1
Pick a complex niche (non-QM, jumbo, construction, self-employed) and use NotebookLM to master its programs, documentation, and edge cases.
2
Build a structured learning path.
Copy-paste this prompt
Act as a mentor to a loan underwriter moving into [non-QM and bank-statement loans]. Build a 90-day plan to become competent: how these programs qualify borrowers, the documentation and income-calculation methods, the biggest risk and compliance pitfalls, and five authoritative resources. Educational only.
Verify all program and documentation specifics against current investor guidelines. AI teaches the concepts; the sign-off authority is earned.
3
Use Perplexity to track rate, program, and regulatory changes in your niche.
What you'll haveCommand of a high-margin lending niche - the specialty that commands senior-underwriter pay.
6
Improve the credit box with AI analytics
Why this pays: The underwriter who thinks about portfolio performance, not just single files, gets pulled into credit policy and leadership. AI analytics reveal where the credit box is too tight or too loose - the strategic contribution that leads to credit-officer pay.
Zest AIPower BIExcel Copilot
1
If your lender uses Zest AI or similar credit-model analytics, learn how the scorecard weighs factors and where manual-underwriting overrides add value.
2
Analyze portfolio patterns.
Copy-paste this prompt
Given this de-identified loan-performance summary by segment (approval rate, early-payment default, loss rate by credit tier): [paste], identify where credit policy may be too conservative (leaving good loans on the table) or too loose (driving defaults), and what policy changes a prudent credit team should test. Frame it as a memo to a credit committee.
Use aggregated, de-identified data. Any policy change must pass fair-lending and model-governance review; the analysis supports, it doesn't decide.
3
Bring the findings to your credit team as a proactive policy proposal.
What you'll haveA voice in credit strategy and a leadership reputation - the path to credit-officer and management pay.
Your 12-month sequence to the top of the range
How the plays above stack into a path from median pay toward the $211,350 tier.
Month 1
Turn on AI document and income analysis (Ocrolus or Hyperscience); verify every figure and clear more files.
Months 2-3
Add AI fraud screening (SentiLink, Socure) and build an AI guideline-lookup notebook.
Months 3-6
Start spreading commercial credit with AI (OakNorth, Moody's CreditLens) to step up in complexity.
Months 6-9
Commit to a complex or non-QM niche and get genuinely expert with AI's help.
Months 9-12
Use portfolio analytics to contribute to credit policy - the route to credit-officer authority.
Next steps for an Underwriter
Some links below are affiliate or partner links. PayCrunch may earn a commission if you enroll or subscribe through them, at no extra cost to you. Wage figures on this page still come from the Bureau of Labor Statistics, not from these programs.
Underwriter work is specific enough that a stamped 'check out these courses' block would be noise. BLS files this work as Loan Officers (SOC 13-2072). O*NET Job Zone 4 is typical: a bachelor's degree, so the honest next credential is a professional certificate or bachelor's-level coursework — not a random catalog dump.
The occupation's listed knowledge areas include Economics and Accounting and Sales and Marketing; the links search those subjects, not a generic 'career courses' list.
Underwriters in this dataset list Microsoft Dynamics among the tools in use, so a program that names that stack is a better fit than a survey course.
Coursera search for writing — a professional certificate or bachelor's-level coursework that lines up with business and finance, not a generic professional-development aisle.
FlexJobs screens remote, hybrid, freelance, and flexible listings so you are not wading through unverified ads. This is a job-board search for Underwriter work, not a claim that they list a counted SOC 13-2072 inventory.
Write an Underwriter resume, or one aimed at Personal Financial Advisors, instead of a blank template. Resume Now is a resume builder; we are not claiming a counted template set for this SOC.
An Underwriter resume that names the actual tasks on this page, or the step-up title Personal Financial Advisors, beats a blank template when you apply.
What Underwriters earn by state
These are the Bureau of Labor Statistics’ own figures for Loan Officers, state by state — not a cost-of-living adjustment applied to the national number. Only states employing at least 500 people in the occupation are shown, because a state median drawn from a handful of workers is noise rather than a signal.
Massachusetts
$101,600
highest of them · +32% vs the national median
Puerto Rico
$35,730
lowest of the 49 states and territories that qualify · -53% vs the national median
The same job pays $65,870 more a year at the median in Massachusetts than in Puerto Rico — 184% higher. That gap is what the Bureau measured, before any question of what it costs to live in either place. The top-of-range figure quoted at the head of this page, $211,350, is a different statistic in a different place: it is the 90th-percentile wage in New York. The state that pays the typical worker most and the state where the best-paid go highest are not always the same one.
Source: U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025, SOC 13-2072. 49 states and territories clear the 500-employee reporting floor for this occupation; those below it are left out rather than shown with a wide error band.
Free data. Use any of it.
PayCrunch publishes verified, BLS-sourced salary + AI-playbook data on 1,000+ professions — free, no signup.
Automated underwriting engines already approve clean, conforming files, and that automation keeps growing - so the underwriter who only clears vanilla loans is exposed. What automation can't do is judge messy income, thin or non-QM files, complex commercial credit, and edge cases, or defend an adverse-action decision to an examiner. Underwriters who move up the complexity curve and use AI on the routine work are more valuable, not less.
Is it safe to use AI on real loan files?
Only inside your lender's approved, secured systems - never by pasting borrower PII, SSNs, or financials into a consumer tool (GLBA). And lending AI carries strict fair-lending duties: no prohibited factors, no disparate impact, and accurate adverse-action reasons you can defend. Treat AI as decision support you verify, not a verdict you inherit.
How does AI move an underwriter toward the top of the pay band?
Comp rises with authority and complexity. AI clears routine files fast (more volume at quality), catches fraud (fewer losses and buybacks), and helps you step into commercial and non-QM credit (higher-value work). Throughput plus clean loan quality plus complex-credit skill is exactly what earns higher signing authority and the $211,350 top of the range.
Can I trust AI income calculations and fraud flags?
As leads and drafts, never as final answers. AI document analysis is fast and often catches what you'd miss, but you must reconcile every income figure against the source documents and treat every fraud flag as something to investigate, not proof. On audit or repurchase, the decision is yours - so verify and document.
Which AI capability should I learn first?
Document and income analysis, because verifying pay stubs, bank statements, and tax returns is where underwriting time disappears. Once that's habitual, add fraud screening and an AI guideline-lookup notebook, then move up into commercial credit spreading. Start with document analysis - it frees the hours you'll reinvest in complex, higher-paying credit.
Methodology & sources
Salary (median, 10th, top of the range) — U.S. Bureau of Labor Statistics, OEWS.
By state — the Bureau of Labor Statistics’ own state medians, limited to states employing at least 500 people in the occupation. No cost-of-living arithmetic is applied to a wage anywhere on this page.
The plays — PayCrunch's own step-by-step guidance using publicly available AI tools. Tool names/URLs are real and current as of August 2026; prompts written to work as-is. Verify any professional output before relying on it.