Loan Officer Assistant Salary: The Real Math on LOA Cost

You're staring at another stack of unworked leads and files stalled on a missing pay stub, and the obvious fix is to hire another loan officer assistant. Before you post that job, it's worth knowing what the role actually costs loaded, and which of its duties a system now does reliably instead of a person.
Quick answer: A mortgage loan officer assistant salary averages roughly $48,700 a year (about $23/hour) per early-2026 ZipRecruiter and Salary.com data, and lands near $69,600 fully loaded once benefits, payroll taxes, and overhead are added. Most of an LOA's duties — chasing documents, updating clients and referring agents on status, maintaining the database — are the remember-to-do-it work a system can run; licensed guidance and the relationship stay human.
What a loan officer assistant salary actually is
A mortgage-specific loan officer assistant earns about $48,700 a year, roughly $23 an hour, per early-2026 ZipRecruiter and Salary.com figures compiled by BrokerVA. Payscale lands close, putting the median base at ~$38,660 with a $30K-$56K base range and $29K-$75K total pay once commission and bonus are counted.
That spread is where owners get tripped up. Glassdoor reports ~$73,379 average total pay for "Loan Officer Assistant" nationally, with a $60K-$92K range, but that headline blends non-mortgage job titles like IT roles that happen to share the words. BrokerVA's breakdown shows Glassdoor's own financial-services-specific figure, ~$50,891, lines up with the ZipRecruiter and Salary.com mortgage numbers. If you plan around the $73K headline, you'll overestimate the salary line and still miss the real cost.
Plan around roughly $48,700 as the mortgage LOA salary. Then add the part the salary sites leave out.
The cost of a loan officer assistant, fully loaded
The salary is maybe 70% of what the hire costs. Benefits, payroll taxes, and overhead typically add about 30% on top of wages, which pushes a ~$48,700 LOA to roughly $69,600 a year fully loaded, per BrokerVA's math. That's before you count the equipment, software seats, and the time your team spends training.
A widely cited rule of thumb from MIT's Joseph Hadzima puts an employee's true cost at 1.25 to 1.4 times base salary once payroll taxes and benefits are added. EconKit runs a $70,000 salary all the way out — payroll taxes, health insurance, 401(k), PTO, equipment, recruiting, onboarding — and gets about $99,132, a 1.42x multiplier. The pattern holds: whatever the offer letter says, budget 1.3 to 1.4 times that.
On a mortgage LOA, the numbers stack up like this:
| Line | Amount |
|---|---|
| Base salary (mortgage LOA) | ~$48,700 |
| +30% benefits, taxes, overhead | ~$14,600 |
| Fully loaded, year one | ~$69,600 |
Then there's the risk you don't price until it happens. If the hire doesn't work out, a common industry estimate puts turnover and replacement cost at six to nine months of salary — $24K to $37K on a $48.7K role — once you count the lost ramp time, the re-recruiting, and the re-training. That's the number that makes "just hire another one" expensive even when the salary looks reasonable.
Loan officer assistant duties, sorted by what a system can run
Read any LOA job description and the duties are overwhelmingly the boring, remember-to-do-it work. AvaHR's template centers on collecting and organizing documents, communicating with clients to gather paperwork and give status updates, and maintaining referral-partner connections. A real Wood Group / Fairway listing centers on updating customers, real estate agents, and title companies on loan progress, coordinating closings, and maintaining the loan officer's customer database.
That list sorts cleanly. The rule-based, information-processing parts of a support role — reviewing documents against a checklist, following up on conditions, routing files — are exactly what AI tools already do, in several cases faster and with fewer errors, per Mortgage Professional America. The parts that stay human are the licensed guidance and the relationship.
The honest split:
- Runs reliably on a system: chasing a borrower for a missing pay stub or stale bank statement, sending status updates to the client and the referring agent on a schedule, nudging a file stalled on a pending document, keeping the CRM and the loan origination system (LOS) in sync so nobody re-keys data, and maintaining the customer database.
- Stays with a licensed human: mortgage advice, product and rate discussions, qualification and pricing calls, approval guidance, and the judgment calls and relationship work a borrower actually wants a person for. RealtyCTL makes the same cut: even with AI handling admin and communication, licensed professionals should stay responsible for advice, product, qualification, pricing, and relationships.
Notice which column is bigger. The database maintenance, the document chasing, the status updates — the work you'd hire an LOA primarily to do — sits almost entirely in the first column. The second column is real and non-negotiable, but it's a smaller slice of a $69,600 role than the job description implies.
This is where SalesOS fits, and it fits one job: it takes the remembering-and-chasing load. SalesOS keeps your CRM and your LOS in sync, tracks what's owed at each stage, chases the missing document, and updates the client and the referring agent on status — following the texting and contact rules your team sets, which your compliance people decide, not us. It doesn't replace the LOA you keep; it hands that person the work only a human can do.
Hire, outsource, or build a system: which decision you're actually making
Owners often collapse three different decisions into one. Hiring in-house buys you a person for both columns above and the full loaded cost. Outsourcing to a mortgage VA buys you offshore human labor for the boring column at a lower rate — Emapta reports 80% of lenders using outsourcing said it converted fixed costs into variable ones. Building a system buys you reliable coverage of the boring column with no turnover risk, and leaves the human column to the people you already employ.
The market backdrop matters here. Loan officer headcount nationally has fallen from a 2021 peak of 124,805 to 86,192 by Q1 2026, per NMLS data cited by HousingWire. The team is thinning, not growing. The lever most owners can actually pull is getting more out of the people and the leads already in the building — which is the whole argument we made in AI sales assistant vs hiring another loan officer assistant.
A rough way to decide:
- List the duties your next LOA would own. Put each in the system column or the human column using the split above.
- If the human column is thin — mostly document chasing, status updates, data entry — a system covers it without the $69,600 and the turnover risk.
- If the human column is heavy — a lot of borrower conversations, judgment, or licensed work — you need a person, and a system makes that person's day about those conversations instead of re-sending reminders.
- Most teams land in the middle: keep one person, give the chasing and syncing to a system, and stop paying a salaried human to copy screenshots between the CRM, the LOS, and the inbox.
Common pitfalls when you run the math
The mistake that costs the most is planning around the inflated salary number. If you budget the $73K Glassdoor headline, you'll either overpay or walk away from a hire that actually pencils at $48.7K base.
A few others worth naming:
- Comparing a salary to a software price. The honest comparison is the loaded cost — $69,600 and the turnover risk — against what a system costs. We broke down the system side in what AI sales automation actually costs a small team.
- Assuming a tool will "just do" the human column. It won't, and anyone promising that is overselling. Rate talk, qualification, and the relationship stay with your licensed people.
- Expecting automation to ignore your texting rules. It should follow the rules your team and compliance set, not run around them. The leak you're fixing is the forgotten follow-up, not the one your compliance officer flagged.
- Buying another dashboard. The database and the follow-up leak out because a human had to remember, not because you lacked a fourth place to look at status. A tool that adds a screen without closing the loop solves nothing. That's the leak we keep coming back to in CRM for mortgage brokers: start with the leak, not the tool.
FAQ
How much does a loan officer assistant make?
A mortgage-specific loan officer assistant makes about $48,700 a year, roughly $23 an hour, per early-2026 ZipRecruiter and Salary.com data. Payscale puts the median base lower at ~$38,660 with total pay reaching $75K once commission and bonus are included. Glassdoor's ~$73K figure runs high because it blends in non-mortgage job titles.
What does a loan officer assistant do all day?
A loan officer assistant mostly collects and organizes borrower documents, chases missing paperwork, updates clients and referring real estate agents on loan progress, coordinates closings, and maintains the loan officer's customer database. A smaller share of the role is licensed or judgment-heavy work like qualification and relationship management, which has to stay with a human.
What does a loan officer assistant actually cost to hire?
A loan officer assistant costs roughly $69,600 a year fully loaded, once you add about 30% for benefits, payroll taxes, and overhead on top of the ~$48,700 base. If the hire doesn't work out, replacement runs an estimated six to nine months of salary on top of that.
Can AI replace a loan officer assistant?
AI can reliably run the boring, rule-based part of the loan officer assistant role — document chasing, status updates to clients and agents, keeping the CRM and LOS in sync, and maintaining the database. It should not handle mortgage advice, product and rate discussions, qualification, or the relationship, which stay with a licensed person. So a system replaces the remembering-and-chasing load, not the whole role.
If you want to see which of your next hire's duties a system would carry and which you'd keep a person for, the fastest path is to walk us through your actual process. Book a demo and we'll sort the duties against your own pipeline before you decide whether the next lever is another hire or a system.
About Terrell Gentry
Founder at 6omb
Terrell is the founder of 6omb and runs Claude Community, the #1 Skool community for Voice AI agents. Over 16 months his team has built 100+ AI agent systems delivering $10M+ in business value, including voice agents like Emily, which booked 453 new clients for a law firm in 8 months. He is a Y Combinator Startup School alum (SUS20) and a Gold Retell partner.
You might also like

AI sales assistant vs hiring another loan officer assistant
AI sales assistant vs hiring a loan officer assistant: the real mortgage math on LOA cost, what's automatable, and how one extra closed deal a month pays for it.

What the bi-merge credit change means for my loan team's process
Bi-merge credit mortgage change explained for loan teams: what's decided, what's still open, and how to prepare your credit-pull step without a scramble.

Working 50+ Refi Candidates When Rates Drop (Without an Army)
Refinance outreach at scale: a triage-and-outreach playbook for working 50+ refi candidates fast when rates drop, without hiring an army of loan officers.
Close more of the deals you already have
SalesOS follows up with every lead in your own voice, keeps your CRM in sync, and shows you where every deal stands. See it working on a pipeline like yours.
Book a Demo