For years, the mortgage industry wore integrations like a badge of honor. Legacy loan origination systems have hundreds of them, connecting to every vendor, service provider and data source imaginable. The more, the merrier.
Sounds impressive? Look closer.
Think about the incentives. For a service provider, being integrated into an LOS means easier access to every lender on that platform. For the LOS, more integrations mean a stickier and more marketable product; and some charge their vendors integration fees on top of that. Everyone wins, right? Except, often, the end user.
The dependency problem
The top priority should be whether the integration actually saves the ops team time. Whether it makes their day easier. Whether it works. But that’s not always what’s being optimized for. I hear it constantly at conferences: “Yeah, we have that integration, but we don’t actually use it.” That’s not a technical problem. That’s a priorities problem.
Even when integrations work exactly as designed, they create a dependency problem few people talk about. A few weeks ago, a technology partner notified us of a technical issue and apologized for the inconvenience. Last week, three of our technology partners sent emails letting us know they’d upgraded their APIs, and we need to adhere. Our timeline is now their timeline. Our roadmap shifts. We’re not building, we’re maintaining.
That’s the best-case scenario.
The worst case? An integration breaks mid-transaction. The LOS user calls support, who contacts the service provider, who reaches out to the third-party contractor who actually built the thing, who responds back up the chain. In mortgage lending, where deals are time-sensitive and borrowers are anxious, that chain of telephone is inefficient and expensive.
Many service providers in this industry are not technology companies. They’re excellent at what they actually do: law, title, appraisals, escrow. But they’ve outsourced their tech to vendors building on their behalf. When something breaks, nobody truly owns the problem.
So what’s the alternative?
The alternative is true partnerships, where each party does what they’re actually good at, on a single platform. When a service provider’s expertise is delivered directly through the LOS – no middleware, no third-party contractors, no API versioning surprises – that’s a win for everyone. When something needs to change, there’s one conversation, not a game of telephone.
Take closing documents. The data in your LOS must map accurately to the legal closing docs, and the language in those docs must reflect the latest applicable regulations. Simple enough – until something changes. A regulation updates, a document gets revised, and now you have a discrepancy between what’s in your LOS and what’s in the closing docs. Who catches that? Who fixes that? How fast? In a true partnership model, there’s one team that owns it.
Which brings me to a question you should ask about your current tech setup: when something changes – a regulation, a document, a data field, who’s responsible and how long does it actually take to implement the change? Can you initiate that change yourself, or are you dependent on a vendor’s busy timeline? Do you have full visibility into the data being exchanged between systems?
And as AI makes data exchange smarter and faster, which platforms will be positioned to take advantage of it, and which will be too tangled in their own integrations to move? If the answers aren’t clear, you already have a problem.
The future of mortgage technology
With AI becoming more prevalent, we might see a different breed of integrations soon – agents responsible for orchestrating data transfer intelligently, without the overhead we deal with today. But it’s not quite there yet in the mortgage industry. With that said, the principle still stands: keep the technology with the tech companies, and the expertise with the experts. Build on that foundation, and you’ll be ready for whatever comes next.
Not all integrations are created equal. Some are seamless. Many are duct tape dressed up as infrastructure. The next time one fails you mid-workflow, the problem isn’t unique to you – and a better model exists.
The future of mortgage technology isn’t whoever has the most integrations. It’s whoever builds the most coherent platform.
Daniel Gottesmann is a Co-Founder of Elphi.
This column does not necessarily reflect the opinion of HousingWire’s editorial department and its owners. To contact the editor responsible for this piece: [email protected].
English (US) · 