A leading US records retrieval provider for the insurance industry, collects medical and other records for the property and casualty insurance industry. The company has around 500 agents dialing out to a network of roughly 200K record custodians. At peak, that added up to more than 100K outbound calls every month.

However, the company had no visibility into what was actually happening in any of the conversations between their onshore and offshore agents. Agents worked manually off spreadsheets rather than a centralized queue. And its CSAT and NPS scores came back poor after custodians pushed back against the high volume of calls they were receiving: "Why do you call us so much?, Why do you call me 10 times a day?".

One case made the problem impossible to ignore: agents had called a single custodian 99 times without retrieving the record. It turned out the team had simply been dialing the wrong number and would have kept dialing indefinitely; with no call logging, recording, or transcription in place, nobody could have caught the problem.

"That realization was a big push. We needed to find a tool that allowed us to get true insights into all the calls we were making." — CTO at a leading US records retrieval company

No rip and replace: Why Replicant fit where others couldn't

Outbound calling ran through a homegrown application built on top of Microsoft Teams, not a standard contact center platform. Dozens of vendors tried to sell this organization a new, full contact center system to replace it. Replicant stood out by integrating directly into the existing infrastructure, so the company could add conversation logging, visibility, and automation without disrupting a setup that was already working.

The company brought in Replicant's Conversation Intelligence to analyze its calls. The first goal wasn't automation. It was simply seeing what was happening.

Visibility alone changed behavior, and the patterns showed up immediately:

• Five different agents calling the same custodian within a 20-minute window.
• An internal myth proved false: customers were expected to hang up the moment they heard "this call is being recorded," but almost nobody did.
• Scorecards revealed that a large share of calls were non-compliant.
• Significant underperformance was identified on their offshore operation, where the agents were making many unnecessary calls. 

"Just the mere act of recording and having visibility into conversations resulted in a 25% reduction in outgoing calls, while our order volume and records retrieval success rate went up." — CTO at a leading US records retrieval company

The organization has recorded and studied close to half a million calls. The team has segmented conversations into 10 high-priority, repeatable tracks that are the raw material for what to automate next.

Turning visibility into a hybrid workforce

From outbound calling, the organization expanded Conversation Intelligence into its inbound customer service line. Along the way, a capability the team wasn't originally looking for became one of the most operationalized parts of the rollout: scorecards and coaching. Managers use them to coach agents against roughly 10 markers of a "compliant call," such as proper introductions and adherence to talk track.

"As that compliance rate rises, we are getting much more productive calls out of our team." — CTO at a leading US records retrieval company

Coaching improved how agents work. The next step was reducing the manual work around each call entirely. AI agents now read call transcripts, summarize, and automatically apply the right status codes directly into its internal system, eliminating manual post-call data entry. The team is also building a "unified activity timeline" that combines Replicant's call data with email, fax, and internal system activity into a single view per case, making it possible to spot the repeatable steps that consume 15 to 20 minutes and automate them directly.

The organization found that 25% of its calls are simple status checks — "you promised us records yesterday, where are they?". This is an easy, task-based work the team can automate outright, freeing people to manage exceptions instead.

Half the calls. Half the wait. More records.

The results touched every part of the operation, from call volume and visibility to workforce costs and the speed of closing a case.

• Greater visibility into which calls needed to be made drove:
  
  • Outbound call volume down, from nearly 100K to 54K a month.
  • Record retrieval success and overall order volume rose.
  • Contacts-to-complete, the number of touchpoints needed to land a single record, was cut in half, from an average of 9 down to 4.5.
• Onshore and offshore agent performance became visible for the first time.
• That same visibility exposed underperformance in the offshore operation, letting the company cut its BPO commitment by 25% and save roughly $1 million.

The next phase: Turning insights into action

This leader in records retrieval has the foundations of intelligence and visibility to understand what is happening in its conversations and how they impact its business. The next phase is to automate the 10 mapped call tracks into AI agents.

"My vision is a hybrid team, a mix of humans and AI agents, where we automate the easy calls, a human oversees it all, and we free up the people to do more complex things." — CTO at a leading US records retrieval company