Fulton Bank is a regional financial institution serving customers across Pennsylvania, Maryland, New Jersey, Virginia, and Delaware through a growing portfolio of consumer, business banking and commercial banking services. Its contact center handles over 50,000 calls each month, supporting a wide range of enquiries.

As Fulton continues to grow, the bank identified a need to strengthen oversight across customer interactions as a means for accommodating growth while maintaining its commitment to excellent customer service.

The challenge

While existing sampled customer interaction reviews provided valuable insight into individual calls, they did not easily reveal broader patterns across the full population, such as emerging service quality patterns or inconsistencies in how guidance was being applied across teams, restricting growth and insight on service effectiveness.

To address this challenge, Fulton partnered with Palqee to implement Prisma as the Oversight Intelligence platform, providing continuous visibility across all customer interactions and enabling the bank to identify growth opportunities, operational trends, and service delivery challenges that would otherwise remain hidden within sampled reviews.

The goal

The primary goal of the pilot was not only to help optimize call handle time but to answer a question that no sampling-based program can answer: what is actually happening on every call, and what does that mean for how we coach, how we grow, and how we serve our customers?

"The goal with Prisma is to have a really clear picture on the level of quality we are giving to customers on every call."                                                                              

                           — SVP Contact Center Channel, Fulton Bank

The scope covered online banking support calls, assessed for accurate information conveyed and authentication, in addition to other call types and quality metrics relevant to Fulton's broader service objectives.

The ambition behind this scope reflects a challenge that has not fundamentally changed in contact centers for decades.

Performance measurement today still operates on the same model it did in the 1980s: a manager listens to a small selection of calls, scores them against a set of criteria, and delivers feedback to the agent. The sample is too small to be statistically representative, the feedback loop is too slow to drive consistent improvement, and the vast majority of interactions, the ones that would reveal the clearest picture of where coaching is needed, are never heard.

The bank sought to demonstrate how Oversight Intelligence supports growth strategies while maintaining consistent service quality and operational control, with the following scope:

• Evaluate service delivery against Fulton’s operational guidelines and quality standards at scale;

• Identify recurring patterns and provide insights into the underlying reasons behind deviations;

• Surface potential gaps, conflicts, or outdated guidance that could contribute to inconsistent customer experiences and restrict growth;

• Assess Prisma’s ability to support quality assurance functions without requiring additional review personnel;

• Support business objectives for growth and efficiency.

Fulton recognized an opportunity to leverage customer interactions as a source of operational intelligence, enabling them to identify improvement opportunities, strengthen service delivery, and support future growth through more informed decision making.

What the pilot found

The findings fell into three categories: first, what the full population of agent performance data reveals about coaching opportunities and service quality that sampling had never made visible; second, opportunities to update policies and strengthen internal guidance; and third, how dramatically the economics of oversight change with the addition of Prisma.

Authentication

Once the full call population was validated by Prisma, failure rates related to customer authentication presented a materially different picture from what periodic sampling had suggested. In some cases, the guidance covering how agents should proceed in a specific authentication scenario could reasonably be read more than one way, which was inflating the apparent failure rate.

Collaborating with Fulton’s subject matter experts reviewing and resolving these cases, Prisma incorporated their feedback into its validation engine, establishing a more consistent interpretation of the underlying policies. Rather than forcing these calls into a simple compliant or non-compliant outcome, Prisma flagged them as ambiguous, a distinction that gave Fulton's team a clear case to review rather than a potentially incorrect result.

As a result, ambiguity rates declined substantially and the true failure rate stabilized within a predictable range. It was an early, concrete example of how full-population oversight can surface and help resolve the kind of interpretive gaps that periodic sampling may never catch.

Online banking calls

Across online banking calls, one of the clearest patterns Prisma surfaced was in how consistently agents communicated available reset channels. Agents handling password reset calls should inform customers of all available reset channels (phone, online self-service, branch), but in 10% of the interactions, agents who were already on a call with the customer consistently completed the reset over the phone without mentioning the other channels.

Agents optimizing for call efficiency were inadvertently reducing customer awareness of self-service options, contributing to repeat call volume and a customer experience gap that Fulton had no previous mechanism to quantify. It's the kind of pattern that is nearly impossible to catch consistently through sampling, but stood out clearly once every call could be reviewed.

“This insight on password reset procedure will help us to broaden out the coaching and align service delivery with business objectives.”

  — SVP Contact Center Channel, Fulton Bank

Service quality at full population scale

The pilot's initial focus was on objective, rules-based scenarios: interactions with a clear right or wrong outcome. Here, full-population coverage already changes what's possible. Under a sampling model, a manager reviewing four or five calls per agent per month is working with anecdotal evidence — improvement opportunities that exist across an agent's full call volume simply aren't visible at that coverage level. At 100% coverage, those opportunities become measurable, and coaching shifts from impression-based to evidence-based.

The next, harder frontier is subjective service quality: how well an agent listens, whether they convey empathy, whether they match the customer's pace, how effectively they de-escalate a frustrated caller, and whether they go beyond the immediate inquiry to offer relevant guidance on account status or next steps. These qualities matter as much, arguably more, to the customer experience, but they don't reduce to a clear right-or-wrong check.

Extending full-population coverage to these more nuanced, conversational aspects of service is the logical next phase of this work. One that would let Fulton Bank ultimately answer, with confidence, whether it is consistently delivering genuinely expert, empathetic service on every interaction, not just the ones it happened to review.

The efficiency gains

Across the full dataset, manual review of the two validation metrics (authentication and accurate information conveyed) would have required approximately 83 hours of performance managers’ time. In the Prisma-assisted process, Prisma took under 10 minutes to validate all calls and managers spent approximately a total of 9.5 hours providing feedback to Prisma.

Incorporating Prisma into the review process has proven to be considerably more efficient, as Prisma’s reasoning, feedback requests, and issue explanations are far more precise and explicit than the previous sampling approach could provide. Previously, agents had to listen to entire calls and assess them against multiple criteria.

To enable Prisma to quickly learn from the bank’s SMEs, the pilot was structured as ten sequential runs of customer calls, with feedback provided between each run. In a production environment, these interactions occur in real time through integrated systems.

The decrease in human review rate is notably as significant as the time spent figure.

In run one, approximately 40% of interactions required the performance manager’s input, as Prisma was encountering Fulton’s policies for the first time and navigating areas for interpretation.

By run ten, the rate had stabilized at approximately 5%. The system had, over the course of the validation runs, internalized the standards that the manager had been applying case by case, and was now applying them consistently, at scale, without requiring further input.

“Prisma created 49 internal controls on the back of manager feedback across a single metric. Those controls now tell us what we should include in the next iteration.”

    — Observed during Fulton Bank pilot engagement

ROI and value delivered

The Fulton Bank pilot produced three categories of value that may be generalizable to any regional bank.

First, and most significantly, it has the potential to transform the foundation of performance management and coaching. Prisma's validations aligned with experienced manager judgment across the policy scenarios and call types assessed in the pilot, and in doing so revealed the greatest opportunities for agent development that sampling had never exposed. The system amplified manager judgment instead of replacing it, applying the standards that the manager had established to a population of interactions that no manager could have reviewed individually.

What was previously an invisible dataset became a continuous source of intelligence about where agents are performing well, where coaching is needed, and where the institution has the clearest opportunities for growth. Contact centers have operated on the same performance measurement model for decades. This pilot demonstrated what becomes possible when that model changes.

This change was estimated to deliver improvements in call handle time, performance managers’ efficiency, first contact resolution rate, reduced customer churn, and missed revenue capture.

Second, it strengthens Fulton's ability to demonstrate consistent, well-evidenced oversight of customer interactions to regulators, reflecting the full population of interactions rather than a sample.

That advantage tends to grow with continued use, as the standards Prisma applies become more refined and consistent over time.

Third, it removed the resourcing objection to full-population oversight. The reduction in manager review time (from 83 hours manually to approximately 10 hours with Prisma across the full dataset) matters not because efficiency is the primary goal, but because it eliminates the assumption that 100% coverage requires additional headcount. It does not.

It requires redirecting the time existing reviewers spend on routine cases toward the genuine edge cases that warrant human judgment, and toward the coaching conversations that the full-population data now makes possible.