Every loan officer has had the file where the most profitable product and the right product were not the same thing. What you do in that moment is the whole of this subject, and it is worth being concrete about rather than writing a values statement.
Here is what putting the borrower first looks like in practice, where the legal obligations actually sit, and the one claim about automation that gets lenders into trouble.
Recommend the loan, not the margin
Start by asking what the borrower is trying to do before you price anything. How long they expect to stay in the house, what the payment has to fit around, whether there is a date they are working toward. Those answers change the recommendation more often than the rate sheet does.
Then present more than one option, including the one that pays you less, and say plainly which you would choose and why. Borrowers are better at detecting a steered recommendation than loan officers assume, and the ones who detect it tell people.
Be equally clear about the trade-offs. A lower payment achieved by extending the term costs more over the life of the loan. Cash-out converts equity into debt secured on the home. A borrower who hears that from you trusts the rest of what you say.
Explain eligibility before the application, not after
A lot of what reads as sharp practice is just a borrower finding out late. Tell them early what the file will need, what might cause a problem, and what would change the answer.
Say when something will not work. Declining to take an application you do not believe will close costs you a file and buys you the referral, and it spares the borrower a credit inquiry and several weeks of hope. The loans you turn away are what make the recommendations you do give worth something.
Where this is a legal obligation rather than a preference
Fair lending is not an ethics topic. The Equal Credit Opportunity Act prohibits discrimination in credit transactions on the basis of race, color, religion, national origin, sex, marital status, age, and several other protected characteristics. The Consumer Financial Protection Bureau publishes guidance on how it applies, and reading it at the source is better than reading about it.
Two practical consequences. Apply the same process to every applicant, so that the questions asked, the options presented and the documentation requested do not vary by who is in front of you. And keep records of why decisions were made, because consistency you cannot evidence is hard to defend.
The claim about automation that is wrong
It is often said that automating decisions removes human bias, because software screens on financial data alone. That claim is not true and repeating it is risky.
Automated systems inherit the assumptions of the people who build them and the patterns in the data they learn from. A rule built on historical lending outcomes can reproduce historical disparities precisely because it is consistent. This is why fair lending obligations apply to automated and algorithmic processes, why lenders are expected to be able to explain adverse decisions, and why regulators have been explicit that technology does not reduce the obligation.
What automation genuinely does well is consistency of process. Everyone gets the same questions, the same disclosures and the same follow-up, nothing is skipped because the day was busy, and there is a record of what happened. That is worth a lot. It is a different claim from the one about bias, and the difference matters.
Measure something other than volume
If the only number reviewed each month is production, the behavior that number rewards is the behavior you get.
Worth tracking alongside it: how many borrowers were given more than one option, how many files were declined early and why, what past borrowers say when asked some months after closing, and how much of your volume comes from people you have served before. That last number is the one that is hardest to manufacture and hardest to lose once you have it.
Where a CRM helps
Most of this is consistency, and consistency is what a system is for. In BNTouch, the borrower record holds what was discussed, what was presented and what was agreed, so the file carries its own history rather than depending on memory. Campaigns and tasks run from that record, which is what makes the same process repeatable across every borrower rather than only the ones who happened to get your attention that week.
Campaigns can be translated, and the language is held on the borrower, so the automation sends in the language that borrower reads. Serving borrowers in their own language is a straightforward way to widen who you can actually help.
Questions loan officers ask
How do I recommend the right product when a different one pays more? Present both, state the difference in what each costs the borrower and what each pays you if your policy allows it, and give your recommendation with the reason attached.
Is it worth declining business? Early and clearly, yes. A file that fails late costs the borrower more than a conversation that ends in week one.
Does automation reduce fair lending risk? It can improve consistency and documentation. It does not remove bias, and the obligations apply to automated processes. Treat any automated decision as something you must be able to explain.
