The End of 30-Year Mortgage Servicing | Valon
MTS
October 11, 2026
Andrew Wang and Linda Du explain why Valon sold its mortgage servicing business to focus entirely on software, how AI agents require operational context rather than raw intelligence to manage complex financial contracts, and how low-friction servicing will enable new products like income-sharing...
Linda Du
A mortgage can last up to 30 years. And over the course of 30 years, just so much stuff can happen. It's like a homeowner can lose their job, can go through bankruptcy, can get married, can get divorced, like in the process of all of these things. And every single one of these events has an implication on this long running financial contract. And so if you want an AI agent to be able to handle this very heavily regulated transaction correctly, it needs to have all of this regulatory context, operational context, like how does the company want to handle certain situations in order to be able to do the right thing consistently.
SPEAKER_2
Hello everyone and welcome back to MTS. Today I am joined by Andrew Wang and Linda Du, the co-founders of Valon. Oh my gosh, we were just talking about it before and I was like, wait.
Linda Du
Valen, Valen.
SPEAKER_2
I got too obsessed with saying it in a French accent. Valen, well welcome. I know last time you guys were on MTS, Valen was doing a few things, running that mortgage servicing business, was building this technology underneath that mortgage servicing business. And since then, Carrington acquired the Servington business, and Valen had now has gone all in on the technology. And yesterday, you guys even announced that $150 million Series D at a $2.3 billion valuation. So really excited to have you guys back on the show.
Andrew Wang
Thank you for having us.
SPEAKER_2
Amazing. So let's talk a little bit about this earlier version, and why did you guys actually end up selling that business?
Andrew Wang
Yeah, so we originally started Morgan Service because of two main reasons.
The first thing was, for the complexity of such a regulated space, we needed to make sure we not only just like generally understood the problem, we were building generally the right product, but we need to understand all of the intricate details in this space so that we made sure that we were building a tailor-made system for this ecosystem. So that's number one. Number two is actually in this ecosystem, it is so regular and people are so traumatized by the great financial crisis that they really care to not only know that there was a great product, but it's been battle tested that someone had spent the years and the effort necessary to go through regulatory tasks, different audits, working with the same sort of clients. Those two things were really important for us to be able to build the right software and really drive the right level of comfort from the industry. We finally got there about a year ago, where the industry finally told us, hey, you guys have been spending six to seven years, not only building this great servicer on this great technology, but you guys have been competing with us so much that we know you're real.
SPEAKER_4
We know you're-
Linda Du
They're like, hey, so we're ready to do the software thing now.
Andrew Wang
You guys proved your point. You told us you're going to do this to prove that the software work, and finally you guys got there. We're ready to go. Can you guys please stop annoying us if you're a servicer? Please sell us software. Now, obviously, if you're going to sell them software, you can't be competing with them. And so for us, the biggest thing for us going into the next stage was, we really wanted to make sure that we could be fully focused on software. And so we sort of said, we're going to now burn the boats, focus on software, because that's what the original intent of the company was. And now we're here to actually get it done.
SPEAKER_2
Congratulations. And even in this announcement today, you guys said something pretty interesting. You say that the big bottleneck for agents, especially when it comes to regulated industries, because you guys are experts in these very highly regulated industries, is context, not intelligence. Linda, I'd love to get your perspective. What does that actually mean in practice?
Linda Du
Yeah. Basically, if you think about it, the models are smart enough at this point to handle basically any servicing activity. And so you're not intelligence constrained anymore. It's really understanding. Like think of it as, you know, at these mortgage servicers, there's someone there who has just done it for 30 years. And they know, like, when this specific edge case happens, you know, we talked to an auditor for New York three years ago, and they told us to do this.
And humans just aren't very good at listing out edge cases proactively. And so for us, it's like the way that you extract that context, that you can build the right thing, is you actually have to go and hit all of those edge cases yourself, capture that context, and then you can kind of build the right system, build the AI on top as well.
Read the other 12 minutes of this episode here.
$15 gets you this episode and 99 more. Credits never expire. After checkout, one link brings you back here to read it.
Not what you expected? Full refund within 14 days.
Thousands of transcripts fetched by people building searchable podcast archives
Fetch transcripts with code
The same credits work from a script or an AI agent. The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000795815149