Outcomes, Not Hype: What Banks and Credit Unions Are Seeing with Saris

When we talk to community banks and credit unions about what they expect from AI, the most common reaction isn't excitement - it's often wariness. AI is approaching the line where overhype meets underdelivery, and the institutions that need real operational help pay the price. They've pitched things that didn't work, replaced systems that promised transformation, and sat through pilots that never made it to production.

As a result, at Saris, we decided instead of leading with the technology, it’s time to lead with real customer outcomes.

Vibrant Credit Union: a 600-loan backlog, cleared in 4 days

Vibrant's HELOC team was buried. The backlog wasn't a seasonal spike but a permanent state, driven by a workflow whose volume had outgrown the team's ability to clear it. Each file took about 15-22 minutes to review, and the work compounded faster than the team could keep up.

After deploying Saris, per-loan review time dropped to between 30 seconds and 2 minutes. The same HELOC team cleared the 600-loan backlog in 4 days. Across the team, employees got back 2.5 hours per day each. Time that had been spent on document review and data validation, not on members.

With this time, they were able to make crucial investments in their members and community, and team morale improved. The technology absorbed the work that had been preventing good people from doing the work they were hired - and love - to do.

Catalyst Corporate Federal Credit Union: 4x the output, same team

Catalyst runs a regulatory filing workflow that historically required hours of manual review. After deploying Saris, the team saw approximately a 75% reduction in processing time. One representative workflow that historically took 9 hours now closes in 45 minutes.

The accuracy held up. Today, Saris is running at 99.8% field accuracy and 94.2% workflow accuracy at Catalyst, with human oversight at every meaningful decision point. Speed without accuracy is just exposure with a different timeline. A meaningful solution has to deliver both.

A regional credit union: cost per workflow, from $1.85 to $0.23

One of our regional credit union customers was carrying a lending operations cost we hadn't seen quantified before. Per-workflow processing cost was $1.85, fully loaded. That's staff time, system handoffs, and the operational overhead of moving a single loan file from intake to clean.

After Saris, that number dropped to roughly $0.23 - a 90% reduction. Annualized, the institution estimates $535K in savings on a single line item.

This is the number most institutions haven't measured. Most don't know what a single loan file actually costs them to produce. When we run a workflow analysis, surfacing that number is often the moment the conversation changes.

A $5B+ regional bank: 3x productivity, same team

This is one of the more complex deployments we've done. Seven operational teams, $4.48 million in annual labor cost, and roughly 122,490 manual document reviews per year.

Six months after Saris went live, 70% of operational tasks were automated. The team got a 3x productivity/output increase with the same amount of staff. We estimate $7.1 million in value unlocked, and $680,000/yr in direct cost savings.

The pattern

The institutions doing the most with Saris aren't the ones with the biggest AI ambitions or mandates. They're the institutions that put their members and communities first and have realized that decades of accumulated manual processes have quietly become the biggest constraint on the service they want to deliver. These are institutions not trying to replace people or entire systems, but rather have their people and processes work smarter and more efficiently.  

While statistics such as 70% of tasks automated, 99% accuracy, and 3x capacity are often repeated in pitches, what is actually harder to capture is the impact Saris AI agents are delivering inside institutions. These outcomes mean teams get to leave the office on time, loan officers who follow up with customers the same day they applied, and processors who work inside one system instead of bouncing between three.

This is what Saris is about. AI agents are the how, not the what. We are building the technology that gives teams at community banks and credit unions their time, their leverage, and their ambitions back.