When it comes to AI, you can't mandate what you can't model.

If you were building your business today with AI underneath it, what would you build, and why aren't you?

Spend a day with twelve to fifteen senior lender peers, where you build working AI on a workflow from your own shop with an expert coach beside you. You leave with the fluency and the confidence to lead the change, not a summary of it.

Full day
12 to 15 peers
Your own workflow
Senior lending leaders at the Lab, with the rule of the day on screen behind them
The June 2026 Lab
The problem

Everyone's evaluating AI. Almost nobody's in production.

83%
Are actively evaluating AI across the business
17%
Have moved one into live production
25%
Say the blocker is trust in the output, not cost
22%
Haven't assessed fair lending risk in AI decisioning

Source: The Mortgage Collaborative, Pulse of the Network, June 2026, across TMC's member lenders.

Evaluation isn't the constraint in lending. Deployment is. The board issues the mandate, the shop buys the seats, the pilot gets built, and it never crosses into production. What stops it isn't the budget and isn't the vendor. It's that nobody senior enough has the confidence to sign off on what the thing is telling them.

It shows up in the P&L. Freddie Mac puts the spread between the best and worst quartile cost to originate at roughly $6,900 against $16,500 on the same loan, with personnel running about two thirds of production cost. Every lender in the bottom quartile can buy the same software as every lender in the top one.

Same tools, different outcomes. The difference is adoption.
From the Labs

One day. A permanent change in how you see your business.

You probably haven't heard of us, and that's fair. Here's what the last cohort said, on the day.

4.9/5
Overall value
4.9/5
Would recommend to a peer
100%
Left with a working build

Scores from the lending executives at the June 2026 Lab, collected on the day.

Watch the Lab film
90 seconds inside a Lab

It didn't take us more than 15 minutes to have a full training program that would probably have cost us $200,000 and taken six months.

Chief Retail Banking Officer

The thing I took away is that I was just overcomplicating it. It turns out you just write ‘create an agent,’ and there you go. It's really not that hard.

COO

Knowing others were in the same boat, and how easy it was to create solutions.

SVP, Underwriting

How simple it was to get a very polished product as a first draft: something that would take us weeks and months of alignment with stakeholders.

Head of Residential Lending
The setting

Seniority is the willingness to be a beginner.

It's the line on the screen when everyone sits down, and it's the rule the whole day runs on. Nobody presents. Nobody performs. Nobody is watching you fumble.

For one day, every senior person at the table is a beginner together: hands on keys, a real workflow on the table, an expert coach at your elbow. That's a setting no conference, no vendor demo, and no board deck can give you.

You can't delegate your way to judgment on AI. You learn by building with it.
A lending executive describing what the Lab changed for them
In their own words · One minute
The day

What happens in a Lab?

We run the whole thing: facilitation, the AI coaches and developers, the data, the venue. You work alongside your peers in a judgment-free environment, everyone building, everyone supporting each other. You bring one workflow and an open mind.

01

Learn

A short, plain-spoken working session on what AI agents actually do in lending today, plus a framework for finding AI use cases, managing the change, and reimagining workflows. No hype, no vendor slides.

Shown live, not pitched
02

Build

You bring a scenario from your own shop: the intake queue, the document checks, the report that eats your week. You drive. Your coach keeps you moving.

Your hands, your workflow
03

Demo

Your table shows the group a working build. Two minutes, live, no slides. The moment your peers watch your agent decision a loan file is the moment the conversation changes.

Live, or it didn't happen
04

The 30 / 60 / 90

Before you leave, you name the one thing you'll rethink in your real business, and we turn it into a one-page plan: your build, and the legacy metric it moves in 30, 60, and 90 days.

The bridge to Monday
A coach working alongside an executive at the Lab
An executive building at the keyboard
A live demo of a working build
The take-home

What do you leave with?

Not notes. Not a vendor shortlist. Fluency you built yourself, and the confidence that comes with it.

IA working AI build you made yourself, on your own workflow
IIA way of seeing you can't switch off: you'll look at any process and see the version rebuilt with AI underneath it
IIIThe judgment to greenlight, kill, or redirect an AI bet worth many times the price of a seat
IVA 30/60/90 plan naming the legacy metric your build moves
VYour prompts and build steps, written up so your team can recreate it
VITwelve to fifteen senior peers wrestling with exactly what you are
Your cohort

Who's building next to you?

Twelve to fifteen senior leaders from lending and financial services, all somewhere on the same transformation journey, all facing a version of the challenge you're facing. Nobody walks in fluent. That turns out to be the point: it's far easier to try something and have it not work when the person beside you is trying the same thing.

They're peers, not an audience. One seat per company, senior line leaders only. Which does mean a competitor could be sitting two chairs away, and that's exactly why the last rule matters.

The group splits into four tables of three or four, each with its own developer, each clustered around one shared workflow. That's why the application asks what it asks: a table only works when the people at it are wrestling with problems that teach each other something.

What's said here stays here. That's what makes the conversation worth having.
ISenior line leaders onlyPeople who own a function and a P&L. No delegates, no observers.
IIOne seat per companyYou won't be sitting next to your own team, so nobody is performing for anybody.
IIIWhat's said here stays hereNo press, and nothing you say in the room is attributed outside it. If we ever want to quote you or put you on camera, we ask first, and you can say no. The Chatham House Rule, if you know it by that name.
Your seat

How do you get in?

Twelve to fifteen seats, one per company, four tables, one developer at each. Every application is read personally, because the tables have to be built rather than filled. Three steps between you and yours.

01

Apply

Twelve questions, about ninety seconds: where AI stands in your shop, where the resistance lives, and the workflows you'd hand over first.

02

Get your read

On the spot: your lending AI stage, how you compare with the market, a fit read, and your Member Pass.

03

Take your seat

Senior line leaders can hold a seat right there. Everyone else books a fifteen-minute call and we find the right way in.

Seat  $1,795 USD
When  November 2026
Where  Orlando, Florida
Cohort  12 to 15 peers
Start my application
90 seconds · no card to start · your read on the spot