Application to clear-to-close: it collects, verifies, and chases every document, drafts Letters of Explanation, and works underwriter conditions. Inside your LOS, with a human where it counts.
Encompass
live read / write
24/7
works every file
Human
clears the risky calls
Unlike back-office roles, processing stays expensive and manual. Your team spends its days creating doc-request tasks, chasing borrowers, checking documents, and re-requesting the wrong ones. The funnel leaks at every step, and every leak is a chase-the-borrower problem.
After the app link
45% of borrowers never submit a single document.
Past the first LO call
Only ~25% of leads that pass the initial call get preapproved.
Preapproved → clear-to-close
Only ~49% of preapproved borrowers reach Clear-to-Close.
Pre-approval → under contract
Takes 5 to 6 weeks, most of it spent waiting on paperwork.
of borrowers never submit a single document after getting the application link. Every one of those is exactly the work the AI processor takes over.
Industry pain point: the work Loandock is built to fix.
Real files from a production deployment, the messy ones a team would dread.
35
Conditions in a weekend
Thirty-five underwriter conditions cleared over a single weekend, work that normally takes about two weeks.
2 weeks
Suspended file to clear-to-close
A suspended file taken all the way to clear-to-close in two weeks, versus roughly a month by hand.
7 days
120+ flag file, cleared
A file with 120+ conditions and flags cleared in 7 days, vs ~25 typical, cutting 60 to 70 hours of manual work down to 8 to 10.
Real files from a live production deployment. Individual results vary by file and lender.
A deterministic engine decides what's needed next, not a wandering AI, wrapped around narrow, individually-evaluated AI judgment steps. Predictable, auditable, and safe.
Once a lead passes filtering and gets the app link, it collects the lightweight gate, photo ID and credit authorization, so real document collection can begin.
Collects the exact income and asset documents for the borrower’s income type, whether W-2, self-employed, SSI, VA, or pension, verifies them, and follows up daily until they’re all in.
Works the full document checklist, verifies each item, and assembles the complete package for the underwriter.
Parses the condition list, determines which need borrower documents, collects and verifies them, and tracks the file all the way to Clear-to-Close.
Works every file 24/7, never drops a borrower, and writes back to your system of record, with a human in the loop wherever it matters.
Reads each loan’s open conditions, derives the right documents by income type, and sends one clean consolidated request, then follows up until everything’s in.
Texts and emails borrowers in a warm, plain-English voice. Smart channel choice, TCPA / quiet-hours aware, opt-out honored automatically.
Checks every submission like a person would: readable, correct type, correct year, correct person, with multi-document sets tracked until complete.
Drafts attestation-grade Letters of Explanation, routes them for legally-binding e-signature, and files the signed copy back automatically.
A chat where the loan officer directs the processor in plain language, “draft the LOX,” “request the remaining docs,” and it executes, every action logged.
Anything the AI can’t safely resolve is escalated to a human, per loan and across the pipeline, with a one-click resolve that does the real work.
Files documents into the Encompass eFolder and recommends conditions satisfied. The AI recommends; a human clears, never silently.
A live, step-by-step view of exactly what the processor did on every file, for trust and auditability.
See it work
Clara collects the documents, a two-model check verifies them, and the condition goes back to the underwriter fulfilled. Click any file. Tell Clara what to do.
Interactive demo. Use the suggested actions or type an instruction to the Clara agent, then approve the item it escalates to you.
4 are a borrower ask. The rest are internal or vendor work Clara tracks without touching the borrower.
A new underwriter condition: pay stubs, most recent 30 days.
auto tourWant it running on your files? Book a walkthrough.
A Model Context Protocol server per system, so the AI reads and writes your system of record over a clean, governed interface, not brittle scripting. Each new LOS / POS is an integration layer, not a rebuild.
Live today
Coming soon
It replaces the expensive, manual, borrower-facing chase, not a back-office task that ships overseas.
A deterministic spine runs the sequence; AI only handles narrow, evaluated judgments. Sends are gated, and anything uncertain goes to a human.
Lives inside the LOS over a real, MCP-governed interface, so the system of record stays the single source of truth.
Documents are filed and conditions recommended satisfied, but a human clears every one. It never silently marks a condition done.
Two AI services, one platform, living inside the stack you already run. Here's where this one sits.
Also from Loandock: the MLO Engine for accurate loan pricing, a separate self-serve product on its own site.
FAQ
An AI loan processor does the borrower-facing work of loan processing: it collects and verifies documents, chases borrowers, drafts letters of explanation, and clears underwriter conditions from application to clear-to-close. Loandock does this inside your existing system of record (Encompass), with a human in the loop on every decision that matters.
Yes. Loandock is live inside Encompass (ICE) today through a Model Context Protocol server that reads and writes the eFolder and recommends conditions satisfied. Vesta and Arive are active too, with Calyx Path next. It plugs into your stack with no rip-and-replace.
Loandock parses the condition list, works out which conditions need borrower documents, collects and verifies them, and tracks the file to clear-to-close. The AI recommends a condition satisfied and files the documentation. A human clears every condition. It never silently marks a condition done.
Loandock is deterministic by design. A fixed engine runs the sequence and AI only handles narrow, individually evaluated judgment steps. Borrower messages are TCPA and quiet-hours aware with automatic opt-out, every action is logged for audit, and anything uncertain is escalated to a human.
An AI loan officer works the front of the funnel: borrower intake and initial qualification, upstream of the LOS. Loandock is an AI processor that works inside the LOS from application to clear-to-close: document collection, verification, and underwriter condition clearing in the system of record, with a human in the loop.
Loandock is priced per file with no seats and no platform fees. It comes with a pilot guarantee: give it 10 conditionally approved files, and if it does not take at least 8 to clear-to-close by their closing dates, the pilot fee is refunded in full.
On live production files, Loandock has cleared 35 underwriter conditions over a single weekend, taken a suspended file to clear-to-close in two weeks, and cleared a file with 120+ conditions and flags in 7 days versus roughly 25 typical. Results vary by file and lender.
Compare and learn more
The grind from “approved” to “clear-to-close,” automated, and escalated to a human only when it truly matters. Live with a production lender.
Live with a production lender. A human signs off on every risky call.
If it doesn't work, you don't pay.
Give us your first 10 conditionally-approved files. If we don't take at least 8 of them to clear-to-close by their scheduled closing dates, your pilot fee is refunded in full. No questions.
What Doesn't Count
If a file changes after we start, that's the loan moving, not the work. Not counted: new derogatory credit or findings that surface after conditional approval, undisclosed debt or funds such as a late secondary or personal loan for the down payment, files suspended or denied for underwriting reasons, and files where the borrower stops responding or won't provide documents. Everything inside our control, we own.