A US mortgage brokerage
First response time dropped from 4.4 hours to under five seconds. Conversion improved by 22 percent. Status-check call volume fell by more than 60 percent. And brokers got their working day back.
The challenge
Mortgage brokerage is a competitive, time-sensitive business where the difference between winning and losing a client is often measured in minutes rather than days. This brokerage was generating strong inbound lead volume across its website, email, and WhatsApp, but its response times were not matching the quality of its demand. The average time from enquiry to first meaningful response was four and a half hours. Its faster competitors were responding in under ten minutes.
The leads themselves were high-intent. People applying for mortgages have already made a significant decision. They are ready to talk to someone. When they send an enquiry at 7pm and receive a response at 9am the following morning, that fourteen-hour gap is not neutral. During that time, they have typically contacted two or three other brokers. The ones that responded first have started building the relationship. The ones that responded last are starting from behind, and often do not close the deal.
After-hours lead loss was the most visible problem, but the business had a second, equally damaging one running alongside it. The mortgage process from application to close typically takes sixty to ninety days, and during that time, customers become anxious. Without proactive communication, they fill the silence with phone calls and emails asking for status updates. The brokerage's team was spending a significant portion of every working day responding to those status enquiries, not because the clients were unreasonable, but because no one had told them anything recently.
Those status calls were not merely time-consuming. They were actively preventing brokers from doing the work that would move applications forward. Every hour spent reading back a status to an anxious client was an hour not spent reviewing documentation, managing lender relationships, or handling the complex casework that required genuine expertise. The business was operationally stuck: growing demand, but an increasing proportion of capacity consumed by routine communication rather than skilled work.
The brokerage had attempted to solve both problems with additional headcount. A small customer service team was added to handle inbound volume and push outbound updates. This helped at the margins but the economics of consistent 24-hour coverage were difficult to sustain, and human availability was still bounded by schedules and timezones. Some leads still went unanswered overnight. Some updates still did not go out on time. The structural problem remained.
What Emma did
- Emma was deployed as the first point of contact for all inbound enquiries across the brokerage's website, email, and WhatsApp channels simultaneously.
- Every lead receives an intelligent, personalised response within seconds of enquiry, regardless of the time of day, the day of the week, or which channel they used.
- Emma qualifies each lead against the brokerage's criteria and routes them to the appropriate broker based on loan type, property location, and application complexity.
- Discovery calls are booked directly into the right broker's calendar, with confirmation messages, preparation notes, and automated reminders sent to the lead in advance.
- At each defined milestone in the mortgage process, Emma sends a proactive status update to the customer in plain, clear language specific to their application.
- The update schedule was designed with the brokerage's operations team to match the actual process steps, ensuring customers always know where they stand without needing to ask.
- Complex situations requiring a broker's judgement are escalated immediately, with full context passed to the relevant person so the handoff is seamless and nothing needs to be repeated.
- All customer data, application information, and conversation history is held in an isolated private environment and is never shared externally or used to train any AI model.
“Leads that used to sit until morning now get a reply in seconds and a booked call. Emma genuinely works for us around the clock.”
What happened next.
The operational change was visible within the first week. The volume of follow-up messages from leads asking whether anyone had received their enquiry dropped to near zero. Every inbound message was being acknowledged and responded to the moment it arrived. The leads that had previously gone cold overnight were in the calendar by the time brokers arrived the next morning.
The reduction in status-check calls took longer to materialise fully, because existing clients needed time to learn that Emma would keep them informed without being asked. Within six weeks, the pattern had shifted clearly. Inbound call volume for status enquiries fell by more than sixty percent. Brokers who had been spending two to three hours a day on the phone with existing clients were spending that time on casework instead. Applications moved faster. The bottleneck that the team had assumed was structural turned out to be operational.
The conversion improvement was the result that caught the operations manager most off guard. She had modelled a modest uplift from faster response times and cleaner lead routing. A twenty-two percent improvement was well above any projection. The explanation was straightforward: the brokerage was no longer giving leads time to make alternative arrangements. Emma's response arrived before a competitor could. The qualification step meant the conversations brokers were having were already warm, informed, and ready to progress.
The after-hours data told the clearest story. Leads that arrived between 6pm and 9am had previously converted at a rate significantly below those arriving during business hours, because the gap before first contact was so long. After Emma was deployed, that difference disappeared almost entirely. The time an enquiry arrived became irrelevant to whether it converted.
For the compliance and legal teams, the privacy architecture was as significant as the operational results. Mortgage applications contain detailed personal financial information: income, liabilities, credit history, asset declarations. This is some of the most sensitive data a business handles, and the brokerage had clear obligations under state and federal law about how it was stored, processed, and protected. Knowing that every piece of that information lives in an isolated environment with no external party able to access it, and that no AI model is ever trained on it, gave the teams the confidence to deploy Emma at the scale the business needed. The information stays with the business. Emma helps the business handle it better.
Your business could be next.
Get started in minutes, or book a demo to see Emma work for you.