4 programs · 5 terms · 21 cash-out levels · 3 structures.
Before any conversation starts, this is the space.
Standard: 945
Insight: 945
945
remaining combinations
"I want to lower my monthly payment"
Standard
"Great! Here are your options for lowering your payment..."
700
Insight
"When you say 'lower payment,' what does that mean for you?"
567
The Pivot
567
remaining combinations
"Both. Payment is a stretch AND we want to redo the kitchen."
Standard
"Here are your rate-term options, and separately, here are your cash-out options."
400
Insight
"I want to flag something. Lower payment and cash-out pull in opposite directions."
83
83
remaining combinations
"$2,000 per month. That's the hard line."
Standard
Still at ~180 options. The user hasn't been asked for a ceiling yet.
180
Insight
Every combination above $2,000 is gone. The system asked for this boundary before generating scenarios.
11
11
to three named paths
Lower Your Monthly
$1,607/mo
$233 monthly relief. Pure rate-term refi. No cash out. Maximum savings.
Smaller Cash-Out
$1,823/mo
$30k for the kitchen. Payment still drops. The compromise path.
This path didn't exist in the standard approach. The user didn't ask for it. The agent invented it by understanding the tension.
Full Kitchen Budget
$1,936/mo
$50k cash out. Close to the ceiling. Maximum kitchen, minimum headroom.
Both approaches reach 3 options. The insight approach gets there in 5 exchanges instead of 10.
But the real difference isn't speed. It's that the user never had to discover the conflict on their own,
never saw an option that violated their constraint, and got a creative path they didn't know to ask for.