Before I moved to the technology side of mortgage, I spent years working inside origination teams at regional banks and independent brokerages. The question I heard most often from loan officers was not "how do I close more loans?" It was "how do I get through the paperwork faster?"
That distinction matters. Closing more loans requires finding more borrowers, qualifying more applicants, navigating tighter credit boxes. Those are sales and underwriting problems. Getting through paperwork faster is a different problem entirely, and it is one that sits entirely within the pre-underwriting preparation stage. It is also the one that Maestro is built to address.
What a loan officer actually does before a file goes to underwriting
Most people who have not worked inside a mortgage origination team think of the loan officer's job as finding borrowers and qualifying them: taking the application, running pre-qualification numbers, locking the rate. That is the front end of the process. But between the application and the file hitting an underwriter's queue, there is a substantial block of preparation work that loan officers either do themselves or supervise a processor to do.
In the teams I worked with, this preparation work breaks into four categories. First, document collection: chasing the borrower (and sometimes the borrower's employer) to provide all required documents, confirming receipt, following up on missing items. Second, document organization: taking the collection of PDFs, photos, and scanned documents that arrive in no particular order and organizing them into a coherent loan file package. Third, data entry: reading the income figures from the W-2 and paystubs and keying them into the LOS, reading the asset balances from bank statements, pulling the liability data from the credit report and reconciling it against the 1003. Fourth, pre-check: doing a rough DTI calculation to make sure the file is in the right range before it goes to underwriting, flagging obvious gaps.
All four of these steps happen before underwriting. They are not underwriting tasks. They are preparation tasks that determine how long it takes a file to get ready for underwriting and how many conditions it triggers when it gets there.
Where the time actually goes: a realistic breakdown
Based on what I have seen working with origination teams, the breakdown on a fairly typical purchase application with a salaried borrower runs roughly as follows. Document collection takes the most calendar time but least active staff time: sending requests, waiting, following up. The active work of collecting and confirming documents takes 30 to 60 minutes spread across 1 to 3 days of back-and-forth.
Document organization takes 20 to 40 minutes on a reasonably clean file. Longer on files where the borrower has submitted documents in non-standard formats (phone photos of paystubs, bank statement pages uploaded out of order, multiple W-2 years submitted when only one was requested).
Data entry is where the hidden time cost lives. Keying income figures from multiple documents into the LOS, reconciling them, entering asset balances, building out the liabilities section from the credit report: this takes 45 to 90 minutes on a clean salaried file and easily 2 hours or more on a self-employed borrower with multiple income sources.
Pre-check takes 20 to 30 minutes for an experienced loan officer who knows what they are looking for. Less experienced processors may take longer or skip it, increasing the likelihood of conditions coming back from underwriting.
Add it up and a clean file requires roughly 2 to 3 hours of active staff time before it is ready for underwriting, spread over 2 to 5 calendar days of document-chasing. A more complex file can easily reach 4 to 6 hours of active prep time.
The cost at pipeline scale
A loan officer managing 15 active files has somewhere between 30 and 45 hours of preparation work in flight at any given time. For a loan officer working without a dedicated processor, that preparation work competes directly with their core job: talking to borrowers, building referral relationships, originating new loans.
The math is straightforward. If preparation work takes 2.5 hours per file and an LO has capacity for 50 hours of work per week, then every file that takes 2.5 hours of prep is consuming 5 percent of their weekly capacity. Cut that prep time by half and the same LO can handle more files with the same weekly capacity, or maintain the same file count with meaningfully less stress and fewer hours.
The challenge is that preparation work is also the category most likely to produce errors that extend the timeline further. A data entry error in the income field does not get caught until the underwriter reviews the file, which might be a week after the prep work was done. The underwriter issues a condition. The LO re-opens the file, re-keys the field, and re-submits. That cycle can add 3 to 5 calendar days to the origination timeline per condition, and most files come back with at least one condition.
What changes when document prep is automated
When Maestro processes an incoming loan file, the data entry step largely moves from manual to automated. Income figures are extracted from the W-2 and paystubs, asset balances from bank statements, employment tenure from paystub headers. The structured output is available in minutes rather than the 45 to 90 minutes a processor would spend on the same task manually.
More importantly, the pre-check step becomes more reliable. Maestro cross-validates extracted figures across documents, flags inconsistencies, and checks completeness against a standard document checklist before the file goes to underwriting. That means the conditions that would otherwise come back from the underwriter ("W-2 income does not match 1003 income" or "only 2 months of bank statements provided") get surfaced earlier, while the borrower is still in regular contact with the loan officer.
This is not a wholesale elimination of preparation work. Document collection still requires human coordination. Document organization from an unstructured upload still benefits from human oversight. The review step after Maestro's extraction still requires a human to confirm flagged fields and verify the summary against the source documents. What changes is the shape of the preparation work: less time on data entry and manual cross-referencing, more time on the judgment-intensive review of flagged items.
The capacity question for independent brokers and lean teams
The impact is sharpest for independent mortgage brokers and loan officers at smaller shops who are doing their own processing. For an LO who manages 20 active files without a dedicated processor, the preparation overhead is a direct tax on their origination capacity. Every hour spent on data entry is an hour not spent on borrower conversations, referral calls, or file reviews that require their specific expertise.
We built Maestro specifically with this profile in mind. The most common complaint we heard from loan officers was not that underwriting was too slow or that rates were too high. It was that they spent too much of their time on tasks that any capable person could do, while the tasks that actually require their experience and judgment sat in the queue waiting. Document preparation is precisely that category: time-consuming, error-prone when done manually, and not requiring the credit judgment that a loan officer spends years developing.
That is the throughput problem we are solving. Not closing more loans per se, but reducing the prep overhead on each file so that the same team can handle more volume, or handle the same volume with less administrative overhead competing with their real work.