A non-bank consumer lender was losing after-hours applicants to slow callbacks. We built an AI intake agent that interviews applicants on WhatsApp, runs affordability rules instantly, and delivers qualified files to the origination system — decisions in 90 seconds, around the clock.
Typical results
24/7
Intake coverage
2.9x
Weekly application volume
12%
Intake drop-off, down from 38%
Figures reflect typical results for this kind of automation.
The challenge
More than half of loan inquiries arrived outside office hours, and applicants who waited overnight for a callback often signed with a faster competitor by morning. The intake team retyped answers from emails and calls into the origination system, pre-qualification averaged 14 hours, and 38% of applicants abandoned the process before finishing.
Before and after
Pre-qual turnaround
−99%Completed applications weekly
2.9×Intake drop-off
−68%The automation
We built a conversational AI agent on the WhatsApp Business API, connected to the lender's origination platform through Make. The agent interviews each applicant in natural language — income, employment, existing obligations, loan purpose — validates every answer as it goes, and runs the lender's affordability rules in real time. Qualified applications land in the origination system with a structured summary and full transcript; edge cases route to a loan officer with context attached.
How it works
- 01
Applicant starts a chat
Prospects message on WhatsApp or arrive from the website form; the AI agent responds instantly, at any hour.
- 02
Guided interview
The agent collects income, employment, obligations, and loan purpose in plain language, validating each answer.
- 03
Affordability rules run
Answers are scored against the lender's eligibility and affordability rules in real time.
- 04
Handoff to origination
Qualified applications push into the loan origination system with a structured summary and full transcript.
- 05
Officers close, not chase
Loan officers start each morning with documented, pre-qualified applicants instead of raw inquiries.
Starts from
- WhatsApp applicants
- Website loan forms
Delivers
- Loan origination system
- Structured applicant summary
- Loan officer alerts
The impact
Pre-qualification now takes 90 seconds instead of an average 14 hours, at midnight as reliably as at midday. Intake drop-off fell from 38% to 12%, weekly completed applications climbed from 45 to 130, and loan officers spend their days closing qualified borrowers instead of chasing cold inquiries.
Questions about this automation
What does the AI intake agent ask applicants?
The agent runs a guided interview on WhatsApp or from the website form, asking about income, employment, existing obligations and loan purpose in plain language. It validates each answer as it goes, asks again when something is unclear or inconsistent, and keeps a full transcript of the conversation for the loan officer.
Does the AI make the lending decision?
No. The agent collects data and runs the lender's own eligibility and affordability rules, which return a pre-qualification result. Final credit decisions stay with the lender's underwriting process. Edge cases, such as irregular income or answers near a threshold, are routed to a loan officer with the context attached.
How does it connect to the loan origination system?
Make pushes each completed application into the origination system through its API, with structured fields, a short summary and the transcript, and loan officers are alerted to new qualified applicants. Because it involves an AI agent, rule checks and handoffs, it is a multi-step build, usually rolled out in stages.