Skip to content

A mortgage office, mid-morning

Their leads were arriving by text. They were dying in inboxes.

Catalyst Mortgage

Hands typing on a smartphone with a text message visible on screen
Where Catalyst's leads were actually arriving — one text thread at a time.

Catalyst Mortgage's leads came in through every channel — calls, web forms, and a steady drip of inbound SMS. The text leads were the problem. They arrived as raw conversations on a loan officer's phone, and the path from "interesting conversation" to "structured CRM record" was manual, inconsistent, and slow.

By the end of the day, half the leads existed only in someone's text thread. Some never got entered at all.

A dedicated number, an AI in the middle

We set up a single phone number for all inbound text leads, paired with a custom CRM built for the way Catalyst actually works. An AI sits in the middle. It reads each text conversation as it happens, extracts the borrower's information — name, scenario, contact, intent — and writes a structured record into the CRM with no manual entry.

If a conversation is missing something the CRM needs, the AI asks the lead the right follow-up question by text rather than guessing.

Image placeholder
Placeholder — drop in a side-by-side of the SMS thread and the auto-created CRM record.

What it removed

The daily ritual of copy-pasting text threads into a CRM disappeared. Loan officers stopped triaging their inbox at the end of the day. Lead response time tightened because the structured record existed by the time anyone went looking for it.

The platform doesn't replace the loan officer. It removes the layer of administrative work that was eating their attention.

We stopped losing leads to data entry.

Catalyst Mortgage operations

The platform became the floor of the operation — one number, one CRM, no manual seam between them.

Client
Catalyst Mortgage
Surface
Custom CRM + dedicated SMS intake
Stack
AI parsing, custom CRM, telephony
Custom CRMSMS intakeLead capture

Want a system like this for your team?

Tell us the operational problem you want AI to actually solve. We'll come back with a path from pilot to production.