System of Record Enables Parallel Cutovers
Mike Yu, CEO of Vesta, on building a new system of record for the mortgage industry
The key implication is that a mortgage LOS is easier to replace than most core banking software, because most of the data inside it expires as loans close. In practice, a lender can keep old in flight files in the legacy LOS, route every new application into the new LOS, and avoid a risky full data conversion on day one. That turns a scary rip and replace into a staged cutover tied to normal loan cycle timing.
-
This works because origination is a short workflow, not a permanent ledger. The loan file is used to collect documents, run underwriting, coordinate processors and closers, then the loan moves into servicing, investors, and securitization systems. The LOS does not need to remain the long term home for every record.
-
The hard part is not the final data move, it is operating two workflows at once and reconnecting the rest of the stack. Pricing engines, borrower intake, compliance tools, and closing vendors all have to know which LOS owns each file during the transition. That is why modern API based systems have an advantage over older walled gardens.
-
This is also why Vesta is positioned as a system of record and workflow engine, not a consumer front end. Front end tools like Blend and point solutions like pricing engines can plug into whichever LOS is active, but the real switching decision sits in the back office system that assigns tasks, stores the loan file, and drives underwriting work.
The next phase of mortgage software will be won by platforms that make cutovers feel operationally boring. As more lenders buy modular tools instead of one vendor suite, the valuable LOS will be the one that can take new volume fast, connect cleanly to surrounding systems, and let the legacy book drain away without freezing the business.