Idaho Housing and Finance Association (IHFA) is a unique financial institution that administers affordable housing resources for the residents of Idaho. As a self-supporting corporate body, they generate revenue by mortgage servicing home loans to support their non-profit mission by providing down-payment assistance to first time homebuyers, providing rental assistance for people in need, and providing affordable housing programs. IHFA has  been doing this for the people of Idaho for over 50 years. 

As a home loan mortgage servicer, IHFA provides a call center for borrowers to ask questions, make payments, and provide any assistance to keep them in their homes. Borrowers need answers that are correct, consistent, and compliant. With roughly 3,000-5000 inbound calls per day, IHFA’s call center didn’t just serve as a support function. It was the entire front door to the organization. IHFA goals is to keep the borrowers happy by having a human answer their questions and keep wait time to a minimum. 

Greg Blake, IHFA’s Chief Information Officer, has spearheaded an initiative to provide better customer experiences for the borrower by making sure every caller gets to the right agent and to answer their questions properly.

The Challenge

Before Bland, every call followed the same pattern. A borrower would listen to a recorded message and try to figure out which button to press on their phone to hopefully reach the correct department. When they reached a live agent, the agent would start with the compliance disclosure and identity verification steps before getting to the real reason for the call.

“The initial phone tree we had took about a minute of the caller's time and the verification process alone takes about another minute,” Greg explained. “then the actual call itself averages another six minutes.”

Those sixty seconds of verification were the most repetitive part of the job, but they were also one of the most important. In a mortgage servicing context, verification and authentication is not optional. It has to be done every time, the same way, with zero shortcuts.

“It’s very important that it’s validated properly every time,” Greg said.

That combination, high volume plus mandatory verification, created a cost and capacity problem.

“Our agents can only handle about 50-70 calls a day,” Greg said. “So as we keep growing, we kept having to add staff. That gets very expensive very quickly.” 

Greg put the economics plainly. “It costs us about a dollar a minute per live agent,” he said. But that first minute was not helping a borrower solve their problem. It was a gatekeeping step required for compliance and privacy. At IHFA’s scale, it became a lot of time spent on non value work, and it created pressure on both staffing and the caller experience.

It also made routing messier than it needed to be. Borrowers would land with an agent, complete authentication, then get transferred to the right department and repeat the same steps again.

“In a lot of cases, callers would get an agent to pick up, and then the agent would need to transfer it to another department” Greg said. “Then they would ask the verification questions all over again.”

So even when IHFA got the fundamentals right, callers paid the price in time, repetition, and frustration. And when call volume spikes, it translates into long holds, higher agent stress, and a worse customer experience. IHFA knew they needed to solve this problem.

The Solution

Greg had been deeply immersed in AI and voice technology for years and knew the space well. When voice models reached a point where they could sound genuinely human, he started evaluating solutions that could realistically take work off his agents’ plates.

He tested several platforms, including companies like Vapi, Retell, and ElevenLabs. But according to Greg, Bland stood out almost immediately.

“Within just a week or so, I just knew immediately Bland was the one,” he added.

“I really liked three things about your platform,” Greg said. “The first was that the price was better… Secondly, the platform was very intuitive… And thirdly, time to get up to speed on the product was really quick.”

Once they’d decided on the platform, IHFA started building pragmatically. First, they replaced their legacy push button IVR with a conversational receptionist.

“I knew I wanted to switch out from ‘push a button here, push a button there’… to have it actually be more conversational,” Greg said.

Then they created a persona named Jenna, a calm, friendly voice that represents the IHFA experience. And this is where their implementation went from “nice IVR replacement” to something much more operationally meaningful.

Instead of pushing authentication onto agents, IHFA used Bland to handle it upfront, reliably and consistently, every time.

“We built APIs that actually handled the verification questions with Bland,” Greg said. “Jenna would send variables over to the API, and the API would validate each caller.”

In practice, that meant Jenna integrated directly into all their pre-existing systems to ask challenge questions, captured the caller’s responses, and passed those values into their softwares for verification. The system then surfaced the result to live agents before they ever picked up the call.

“We are finding that most calls get fully validated on the first try,” Greg said. “People seem to consistently give all the needed information to the AI.”

They started rolling it out first with hazard insurance, the smallest department to prove it out. Then they expanded to general insurance before later expanding to the full company.

Throughout implementation, Greg leaned heavily on Bland’s team.

“The folks at Bland were very helpful,” he said. “We have weekly meetings. They helped us build and refine our agent, and even helped design custom dashboards for us.”

“In terms of the customer service, I give you an A plus,” Greg said. “You just don’t usually get that with a company.”

The Outcome

According to Greg, the impact was immediate once calls went live. Average call times dropped from roughly seven and a half minutes to around six minutes. At IHFA’s scale handling around 100k calls every month, that time savings translated directly into extreme cost reductions. “Our agents cost about a dollar a minute,” Greg said. With the ROI we get from Bland, that math is a no-brainer.”

On top of that, they saw a huge decrease in the total number of needed transfers. “Bland’s AI would transfer callers to the appropriate department after verification of the loan, which reduced the amount of transfers by an incredible amount,” Greg said.

Those savings added up fast. “We save thousands of dollars a day,” Greg said. “If I have to turn it off for any reason, I cringe because I know how much money we could be saving.”

Beyond cost, the quality of life improved for agents and callers alike. Live agents spent less time on repetitive disclosures and verification, and more time actually helping bring in more customers. Routing improved as well. Instead of agents transferring calls after the fact, Jenna sent callers directly to the correct department with a 100% success rate.

What began as an experiment to replace an outdated IVR quickly became a core part of how IHFA manages growth. By offloading repetitive, non-value work to a conversational AI, the organization built a more scalable, more human front door, without compromising compliance or service quality – and Greg saved them hundreds of thousands of dollars because of it.