AI Products
Customer-facing
Employee-facing
Who We Help
Employee-facing
Platform
Trust & Security
Copyright ©0000 Posh AI. All Rights Reserved.
How Hudson Valley Credit Union built one connected member experience across the web, phone, and knowledge, and beat its year-one goals along the way.
Hudson Valley Credit Union (HVCU, $8 billion AUM) serves 13 counties across New York State. Over the past several years, HVCU has layered Posh's Digital Assistant, Voice Assistant, and Knowledge Assistant into its operations, each one solving a different piece of the same problem: getting members answers faster, and freeing staff to focus on the conversations that need a human touch.
HVCU's first move with Posh was on the web. The credit union had watched Live Chat volume climb faster than it could hire and train new agents to its standards, and found that a large share of those chats were repetitive questions with answers already sitting in public information on its website.
HVCU chose Posh's Digital Assistant for its ease of use, default intents, credit union focus, and voice capabilities for the future, integrating it with the credit union's live chat vendor and online banking provider. The ability to add or edit its responses without an engineer was a key win for the team.
“Accessing technology in this space that doesn't require a programmer to effectively use it and serve members is huge,” Goodwine said. “If we send out a communication, five minutes later our assistant can answer questions about it. We don't need to put in an IT ticket and wait for a programmer.” Usage has grown steadily since, from both members and non-members, without HVCU needing to add resources to keep pace.
Before Posh, HVCU ran a touchtone phone menu built on pre-recorded announcements. More than 80% of calls dropped into a single bucket, because most callers just pressed zero. That made it impossible to prioritize anything: a balance inquiry and a loan application landed in the exact same queue. Updating the menu wasn't much easier: it meant tracking down the person whose voice was recorded, having IT upload the new file, and reversing all of it once the message was no longer needed.
Posh's Voice Assistant took over the overloaded queue, routing members to the right place automatically instead of leaving them to guess at a menu. HVCU named it Lee, and staff talk about it like a coworker, not a tool.
"Since we've deployed Lee, I enjoy my job so much more, because I'm only doing things that require me to think," one representative told Steve. Balance checks and simple transfers now largely happen without an agent, freeing reps for calls that use critical thinking and soft skills.
Steve sees it as a matter of symmetry. "The member's experience begins by waiting to speak to an agent. The agent's experience begins at the connection to the member. The closer those two things are starting at the exact same time, the better the experience for everybody."
Even with Lee handling more calls, agents still needed fast, accurate answers for the ones that reached them. Before Kaycee (Knowledge Assistant), that meant digging through thousands of internal documents or tracking down a coworker who might know the answer, a bottleneck that slowed down calls and left agents guessing.
HVCU deployed Kaycee, its AI-powered Knowledge Assistant, in the summer of 2023 to close that gap. Agents ask a question and get an answer immediately: no more scavenger hunts, no more relying on peers.
"Knowledge Assistant takes the pressure off employees," Goodwine said. "Members aren't left waiting for answers, and agents don't need to know everything off the top of their head. They just need to know how to ask the right question."



Faster answers turned into faster ramp-up, and faster ramp-up turned into loans.
Before Kaycee, roughly six months passed between a new hire taking calls on their own and earning the ability to originate loans. Now it takes three.
"That difference for us today is over $500,000 in loan dollars per new hire in that first year," Goodwine said.
Tenured agents saw a version of the same thing. Time they had been spending to uncover knowledge and answers went into relationship building and sales instead, and the conversations got better. The clearest signal is in the pull-through. In the 12 months before Kaycee went live, HVCU funded 39% of the loan dollars it originated. In the 12 months right after, it cleared 50%.
"That itself is a huge boom to the bottom line," Goodwine said.
Lee (Voice Assistant)
Kaycee (Knowledge Assistant)
Loan production
Digital Assistant
Overall
Splitting calls across 28 specialized skills didn't just improve routing, it let HVCU rethink how agents grow into their roles. The credit union is now building a career-progression model, so new agents ramp up on simpler call types first and take on more complexity as they're ready.
The benefits have extended beyond the call center floor, too. HVCU increased available PTO to employees, meaning more staff can take time off with fewer denied requests.
It's also changed how disputes over what happened on a call get resolved. Being able to pull a chat log and show exactly what happened ended the guessing game, Goodwine said, so agents no longer have to wonder whether a member's secondhand account of what "the virtual assistant did" is accurate.
HVCU's roadmap keeps growing. The team is expanding Knowledge Assistant beyond the contact center to every branch and operations employee, with a native Teams integration to make it even easier to access. HVCU is also piloting Posh's training simulator for sales, de-escalation, and leadership coaching, and exploring a version that could act as a guided trainer, running entire lesson plans for individuals or groups.
The business case is built on the same lever as the Kaycee results: loan production, and nothing else. Net zero cost in year one, and a five-year NPV over $500,000, with every other benefit the team expects to realize discounted to zero.
"We only see this accelerating with Simulator," Goodwine said.
"Posh really does listen to the things that we need," he added, pointing to Coach QA, a tool inspired by discussion at Posh Conversations to pair simulator-based call scoring with live phone calls.
Kaycee, the simulator, and Coach QA are three products, but at HVCU they're starting to behave like one loop around the employee lifecycle.
New hires ramp faster because Knowledge Assistant and the simulator carry part of onboarding. Once those employees mature into independence, Coach QA listens to far more calls and returns far more feedback than a supervisor ever could. That data shows where knowledge is thin and which scenarios need practice, so the team refines articles and adapts simulations to match. Then QA surfaces the next set of gaps.
"That is one place where the human is in the loop," Goodwine said, "bringing the QA data to actionable changes in the other tools."
Goodwine's advice to another credit union considering Posh: "Go for it." Every time he has been curious about a solution, Posh's answer is consistent: "Don't just take our word for it, try it."