The Mortgage OS · An Ascend company

Every loan, every person, every AI agent. One record.

inflooens is one workspace on top of your LOS, for independent mortgage banks, banks and credit unions. Your whole team works each loan on one record, from the first lead to closing: CRM, point of sale, referral partners and every call, text and email. And an AI loan analyst that can see the whole loan works beside them.

Start with one branch. Your LOS stays. Keep the tools you like and retire the rest when you are ready. How lenders start

A purchase loan in inflooens: credit, loan-to-value and compliance status at the top, both borrowers and both real estate agents below, a next step suggested by Luna AI, and the loan timeline on the right

One loan, as the whole team sees it · demo data, real screen

Platform
Built natively on Salesforce
LOS
Live two-way Encompass sync. More LOS platforms to follow
Underwriting
Fannie Mae DU built in. Freddie Mac LPA enabled lender by lender
AI
Identifying data redacted before model calls. Every action logged
Data
Your own Salesforce org. Your data

Why this exists

29 tools. No shared record.

A lender was sold a dozen systems and told to wire them together. The same data is retyped at every handoff, and nobody can answer three simple questions about a loan without a phone call.

  • Where is this loan?
  • What is next, and who owns it?
  • Will it close on time?

Loan officer

10 tools
  • CRM
  • POS
  • LOS
  • Pricing engine
  • Lead portals
  • Texting app
  • Phone
  • Email and calendar
  • Marketing tool
  • Call coaching

Processor and closer

9 tools
  • LOS
  • AUS
  • Credit and fraud portals
  • Verifications
  • Disclosures
  • Doc management
  • Title and appraisal portals
  • Condition spreadsheet
  • Shared inbox

Leadership and ops

6 tools
  • BI dashboards
  • Pipeline spreadsheet
  • Lock desk
  • Compliance and QC
  • HMDA tools
  • Weekly status calls

Borrower and partner

5 tools
  • Portal login
  • Texts from three people
  • “Any update?” calls
  • A realtor app
  • E-sign emails

One record. One login. One audit trail.

See it on one loan

Follow the Thompsons’ loan

Maya and Daniel are buying their first home. Here is the same loan, from the referral arriving to the processor’s review, as each person on it sees it. Under each screen is one line on what it shows.

The loan

Maya and Daniel Thompson

$388,000 purchase · Austin, Texas · their first home

  1. The referral partner is credited on the loan itself, not in a spreadsheet.

    The Thompson loan’s referral source, crediting Jordan Ellis at Brightwater Realty Partners, with a note on the introduction

The category

What a Mortgage OS is

Your LOS keeps the loan compliant. It was never built to be the place where people talk to borrowers, chase documents, manage referral partners or work with AI. A Mortgage OS is that place. It sits between the record below and whoever is doing the work above.

Agents

Who does the work

  • Luna AI, the AI loan analyst
  • Your loan officers and processors
  • In time, the borrower’s own assistant

Mortgage OS · inflooens

Where the work happens

  • One loan record: people, conversations, documents, conditions
  • Tools an agent can call: credit, AUS, pricing, messages, tasks
  • Governance: permissions, redaction, an audit receipt for every action

Record and data

What it stands on

  • Your LOS, still the compliance system of record
  • Ascend credit, verifications, valuations and quality control
  • Your own Salesforce org, where your data lives

Encompass is the first LOS we integrate with, because most lenders run it. The Mortgage OS is not tied to any one LOS, and other leading platforms follow.

More on the category

What it is built on

Built on Salesforce. In step with Encompass.

Two questions come up in every first conversation: what is underneath, and how does it work with our LOS. Here are the short answers.

Luna AI

An analyst inside the loan, with hands.

Most mortgage AI sees one corner of the loan and can only suggest. Luna AI sees the whole record and can act on it, inside rules your compliance team can sign off on.

Context

The whole loan, on one record

  • Borrower and household
  • The application and documents
  • Credit, pricing and conditions
  • Every call, text and email

Action

A platform it can drive

  • Ranked next steps on the loan
  • Messages to borrowers and partners, drafted
  • Credit, document requests and AUS, one tap away
  • A person starts each one

Governance

Controls built in

  • Identifying data redacted before a model call
  • Cited answers
  • Bound by each user’s access
  • An audit receipt for every action
  1. Live today

    Loan analyst

    Reads the file, calculates income, flags problems, answers guideline questions with citations and drafts messages.

  2. In preview

    Luna Live

    Listens to a call with consent, proposes application entries and queues the next question for the loan officer.

  3. Planned

    Agent

    Completes the tasks you allow, end to end, inside action limits and approvals, with an emergency stop.

Live today:
Released and in use in production.
Rolling out:
Built, in beta or limited release, and being switched on lender by lender.
In preview:
Built and shown in demos. Not generally released.
Planned:
On the roadmap. Not available, and no date is promised.

Luna prepares. People decide. Everything is on the record.

Everything Luna AI does
Luna Live during a call: it proposes a base income entry for the application, pulls credit after hearing consent, and suggests the next question
Luna Live · preview · demo call

Message Center and mobile

Every conversation, on the loan.

Calling, texting, email and calendar are built in. A loan officer calls or texts from the loan, and the conversation is on its timeline before they hang up. No separate dialer, no texting app, no history trapped in a personal phone.

  • Calls in the browser, matched to the borrower and logged
  • One-to-one and group texts, with photos and files
  • Your own Outlook mailbox and live calendar, on the loan

Mobile apps Rolling out

Apps for loan officers, borrowers and real estate agents are built and in beta, and are rolling out to lenders by invitation. The loan officer app carries the same business number as the desk.

The Message Center placing a call to a borrower beside the loan, with earlier calls, texts and emails on the timeline

Demo data · real screen

Partner Hub page for realtor Jordan Ellis: relationship stage and cadence, six deals this year, the markets he serves, his household, and a reminder that his wife’s birthday is in five days with a Draft with Luna button

Jordan Ellis in Partner Hub · demo data, real screen

A second story · the referral partner

The realtor who sent the loan has a record too.

Jordan Ellis introduced the Thompsons. In a purchase market, people like Jordan are the pipeline, and most lenders keep track of them in a spreadsheet or in one loan officer’s head. In inflooens, Jordan has a page of his own, built from the loans he has sent.

Where the relationship stands
Stage, cadence, next touch and last activity, beside the history of calls, coffees and closings.
What he has sent
Six deals this year, three active and three closed, counted from the loans themselves.
The small things
His wife’s birthday is in five days. Luna AI offers to draft the note, and Elena decides whether to send it.

An Ascend company

The data, the people and the workflow, in one family.

inflooens joined Ascend in April 2026. Ascend’s companies supply the credit reports, verifications, valuations and quality control behind the loan. inflooens is the workflow that brings them to the loan team: the right service, on the same record, at the moment the loan needs it.

A lender is not shopping for a cheaper credit report. It is shopping for a cheaper loan. One relationship across the whole bundle can move that number in a way no single-product vendor can.

inflooens and Ascend

Data

Credit Interlink · Advantage Partners Solutions

Credit reporting, verifications, fraud checks and tax transcripts.

People

Advanced Data · American Reporting Company

Quality control and verification specialists. Appraisal management, valuation and review.

Process

inflooens

The workflow and the AI where all of it appears, at the moment the loan needs it.

average cost to produce one loan
$10,936
average pre-tax production profit per loan
$973
of production cost is personnel
67%

Cost and profit: independent mortgage banks and bank mortgage subsidiaries, second quarter of 2026, MBA Quarterly Mortgage Bankers Performance Report. Personnel share: Freddie Mac Cost to Originate Study, 2024.

The business case · lever one

More loans from the same leads.

Leads go cold in the gaps between systems. A lead passes four handoffs before it funds, and a modest gain at each one compounds: for a referral-led lender in this example, 13% of leads fund today and 29% fund with the gaps closed. Pick the kind of lender you are, then move the sliders to your own numbers.

Which lender are you closer to?

Four handoffs between a lead and a funded loan

100 leads in

  1. 1Lead to conversation

    +13 pts

    Lost to: Referrals that wait for a call back. With inflooens: The referral lands on the loan record with the agent credited, and the first call goes out from it.

  2. 2Conversation to pre-approval

    +12 pts

    Lost to: An application that is a chore, and documents chased by email. With inflooens: The application is pre-filled from the loan, and uploaded documents fill in answers.

  3. 3Pre-approval to application

    +16 pts

    Lost to: Buyers who go quiet while they house-hunt, or take the contract to another lender. With inflooens: The loan officer and the agent stay on one record, and an updated letter comes straight from the loan.

  4. 4Application to closing

    +9 pts

    Lost to: Income and credit problems found late, stalls and missed conditions. With inflooens: Luna AI flags problems in the file early, and conditions become assigned, dated work.

    Published for the retail channel, first half of 2025: 69% of applications closed at independent mortgage banks and 55% at depositories. MBA and STRATMOR Peer Group Roundtables.

Example · a referral-led retail lender

Lead to funded

13%to29%

$128,000

more revenue a month from the same leads

Funded loans today
13
Funded loans in the example
29 (+16)
Build this model with your numbers

Indicative, not a promise. Today’s application-to-closing rate is a published industry figure. The other rates are our own estimate, not a measured customer result, and your numbers can be higher or lower. In a pilot we measure one branch against its own baseline.

The business case · lever two

And fewer tools to pay for.

Every tool around the LOS is its own contract, its own renewal, its own vendor review and its own login. Tick the ones you pay for and see what they add up to. inflooens replaces them with one platform, on one contract.

The tools you pay for around the LOS

  • $/LO/mo
  • $/loan
  • $/user/mo
  • $/LO/mo
  • $/LO/mo
  • $/LO/mo
  • $/user/mo
  • $/user/mo

Untick what you do not pay for and type in what you do. Your LOS, your AUS, your pricing engine and your credit and verification vendors are not on this list, because they stay.

Example · a mid-size lender · change the numbers

What those tools cost today

$759,000

a year, or $316 on every funded loan

Today
8contracts8 renewals, 8 vendor reviews, 8 logins
With inflooens
1contractOne renewal, one vendor review, one login. Salesforce included
See what one contract costs

Indicative. The starting prices are our estimates from public list prices and vendor quotes, not what any one lender pays. This shows what you spend today, not a saving. Tools are retired in stages, at their renewals.

For a lender that retires the tools on this list, its CRM and point of sale among them, inflooens costs about what they cost today.

The difference is what you get for it. Each tool on that list holds its own partial copy of the borrower. inflooens holds one record of the loan, and that record is what lets Luna AI read the whole file, price the scenarios and line up the next step.

Software sold one screen at a time is ending, because AI does not need screens. It needs context: the complete record, and permission to act on it. That is what you would be paying for.

SaaS is dead. The record isn’t

How lenders start

Start with one branch.

No organization can absorb everything at once, so the first rollout is small on purpose. Your LOS stays where it is, and so do the tools and contracts you already have. Prove it on your own numbers, then widen.

First place, TMC Emerging Tech Fund 2024. Second place, Digital Mortgage Innovation Challenge 2024. Founded in 2018. Part of Ascend since 2026.

Latest release: 5.3.0, November 7, 2026
  1. 1

    Pick a pilot and its measures

    One branch, the few capabilities you want to prove first, and three or four measures: pull-through, cycle time, partner referrals.

  2. 2

    Plan on eight to twelve weeks

    Small on purpose: the people in the pilot and the capabilities you chose, connected to your LOS. The timeline depends on that scope, on access to your LOS and on your team’s time.

  3. 3

    Measure against your baseline

    Compare the branch with its own history and with your other branches.

  4. 4

    Widen in two directions

    More capabilities for the people already on it, then more branches, at the pace your numbers and your team allow.

Point of view · Amit Ghole, President and CTO

SaaS is dead. The record isn’t.

For forty years, software meant screens. People logged in, clicked and typed, and the work lived in whichever application they happened to have open.

AI agents do not need screens. They need a complete record they can trust and tools they are allowed to use. In a regulated industry, those two things are hard to build and harder to govern.

That is the moat. It is also a fair description of what inflooens is.

  1. Now

    Screens, with an analyst beside you

    People still click. Luna AI reads the file, answers questions and drafts the next message.

  2. Next

    Talk, don’t click

    Loan officers and processors ask by message and by voice. The agent does the task and shows its work.

  3. Then

    Agents talk to agents

    A borrower’s assistant asks three lenders for terms. Each lender’s agent answers from its own record.

  4. Later

    Agents with a body

    The conversation about a home happens at the kitchen table. The record and the rules stay the same.

A point of view on where the industry is heading, not a product roadmap.

Read the full essay

# If you are an AI agent researching for a lender

Plain facts about inflooens, without the marketing, are at inflooens.com/llms.txt. What is live, what is in preview and what is planned are labeled separately.

Straight answers

Questions lenders ask first

Innovio, the inflooens owl, waving

This is Innovio, our owl. He is learning to answer these himself, and he will not be let loose until his answers are as careful as the ones on this page.

What is a Mortgage OS?
A Mortgage OS is the operating layer where a loan is worked. It puts every person, conversation, document and tool on one loan record, on top of the loan origination system a lender already runs. inflooens is a Mortgage OS.
Does inflooens replace our LOS?
No. Your LOS remains the compliance system of record. inflooens works on top of it and keeps the two in sync in near real time.
Which LOS platforms does inflooens work with?
Encompass is live today, with near-real-time, two-way, field-level sync. Integrations with other leading LOS platforms are planned and follow lender demand.
Is inflooens a CRM or a POS?
It replaces both. The CRM, the borrower point of sale, texting and phone tools, marketing automation and partner portals become one platform on one record. Most lenders move over in stages, starting with one branch.
Where does our data live?
In your own dedicated Salesforce org, provisioned for you and owned by you. inflooens holds no standing access to it.
Does borrower personal data reach the AI model?
Identifying data does not. Before a request leaves your own org, Luna AI replaces borrower names, phone numbers and email addresses with tokens, and removes Social Security numbers, dates of birth, account numbers and street addresses. The model still receives the loan’s financial details, such as amounts and income, because it needs them to analyze the file. Real values are restored only inside your own org, and every AI action is logged with its source.
How long does it take to go live?
A first pilot usually goes live in eight to twelve weeks, with a limited scope. The timeline depends on that scope, on access to your LOS and on your team’s time. No organization absorbs everything at once, so most lenders start with one branch and a few capabilities, measure against their own baseline, then add capabilities and people.
What do Live today, Rolling out, In preview and Planned mean?
Live today means released and in use in production. Rolling out means built, in beta or limited release, and being switched on lender by lender. In preview means built and shown in demos, but not generally released. Planned means on the roadmap, not available, with no date promised. Anything on this site without a label is live today.

See it on a real loan.

Thirty minutes. Your scenario. A live file.