How AI-Driven Coaching Lifted Sales Conversion and Cut Handle Time at a Top US Insurance Carrier

A leading national insurance carrier — one of the largest writers of auto, home, and multi-line personal insurance in the United States — runs a roughly 1,500-agent contact operation that blends in-house teams with a BPO partner. The work spans Sales, Service, Bilingual Sales, Bilingual Service, Blended, Outbound, and Commercial Service programs. Performance is everything: every call is either a quote that converts, a service interaction that retains a policyholder, or a moment that costs the carrier on either side.

For years, the carrier's frontline managers were trapped in the same operational pattern: spend most of the week assembling performance data, then squeeze coaching into whatever time was left. The reporting tools worked. The coaching cadence was defined. What was missing was the connection between the two — and visibility into whether the coaching was actually moving the metrics it was supposed to move.

AmplifAI replaced the reporting work, surfaced the coaching opportunities, and started measuring whether each leader's coaching was converting into agent improvement. The carrier moved from a coaching culture defined by whether sessions happened to one defined by whether sessions worked.

A Blended Operation, A Single Source of Truth

The carrier's contact operation runs across multiple programs and sites, with internal employees and BPO-partner agents working side by side under a shared performance standard. Before AmplifAI, that "shared standard" was aspirational. Each program had its own scorecards, its own report stack, and its own coaching rhythm. Comparing performance apples-to-apples across in-house and BPO required a manual reconciliation that almost no leader had time to do.

AmplifAI replaced that fragmentation with a single, unified performance management platform that pulls data from telephony, QA, workforce management, and the carrier's sales quoting systems into one view. In-house team leaders and BPO supervisors now coach against the same metrics, the same definitions, and the same Next Best Coaching Actions. For the first time, an executive could ask "where is the highest-leverage coaching opportunity across the entire operation" and get an answer in seconds.

Where the Coaching Hours Were Going

A pre-deployment audit of how frontline managers actually spent their time told the story of the problem. Across a typical 25-hour coaching-related work block, managers were burning meaningful share on data gathering, data consolidation, distribution of reports, and identification of focus agents. The actual coaching conversation — the part that drives performance — was a smaller slice than anyone expected.

Three structural drags compounded the issue:

• Data overload. Frontline leaders were pulling reports from multiple systems, copying and pasting into spreadsheet scorecards, and reconciling numbers that disagreed depending on the source.
• Limited coaching visibility. Senior leadership could see that coaching was happening on calendars but had no objective view of whether the sessions were producing measurable agent improvement.
• Operational bottlenecks. Without a unified view of agent performance, coaching prioritization was a guess. Managers coached the agents who happened to be in front of them, not the agents whose performance gaps mattered most this week.

The carrier's leadership decided that fixing this was less about adding coaching headcount and more about giving the existing managers their week back.

What AmplifAI Changed

AmplifAI's deployment hit four operational levers simultaneously:

1. Automating Data Collection and Analysis

Reports that used to take a manager hours each week were replaced with real-time, role-specific dashboards. The same metrics, the same calculations, available on demand to every leader and every agent. Manual scorecard assembly disappeared. So did the variance arguments that came with it.

2. Performance Transparency for Agents

Agents no longer had to wait for a leader to send them their numbers. On-demand access to performance data and coaching actions gave agents ownership of their own metrics — a quiet but durable engagement lever, especially for the highest-performing and most competitive agents who started using the platform as a personal dashboard between shifts.

3. Next Best Coaching Actions

The platform analyzed agent performance and recommended specific coaching priorities for each leader: which agent, which behavior, which metric, this week. Recognition opportunities surfaced alongside performance gaps, so coaching time was distributed where it would actually move the needle. Managers stopped guessing what to coach and started executing against an evidence-based queue.

4. Tracking Coaching Effectiveness at the Leader Level

The Coaching Effectiveness Index measured the percentage of coaching sessions that produced measurable agent improvement, separating leader effectiveness from raw team performance. A high-performing team with a low-performing coach is a hidden risk; a low-performing team with a high-effectiveness coach is a signal to invest. The carrier could now see both, and target leader development where it would compound fastest.

See It in Motion

Outcomes

The combination of unified data, AI-driven coaching priorities, and coaching-effectiveness measurement produced durable, multi-program impact across the operation.

Sales Conversion: A 7% Lift Where It Matters Most

Sales conversion — measured as quote-to-call ratio — is the metric that most directly drives revenue. Approximately 27% of all coaching sessions were deliberately focused on improving quote-to-call ratio, and the result was a 7% lift in sales conversion across blended sales programs.

The most striking story is in Bilingual Sales. As coaching effectiveness on the Bilingual Sales program climbed from roughly 51% to over 80% inside a single quarter, agent performance moved with it: quote-to-call ratio rose from roughly 47% to over 57% — more than a ten-point swing, in a metric where one point is meaningful. The blended program saw an even larger lift over the year, climbing from the high-thirties to the mid-fifties as coaching effectiveness compounded month over month.

Average Handle Time: A 5% Reduction at Quality

Approximately 16% of coaching sessions focused on average handle time. The result was a 5% AHT reduction across the operation — efficiency gains that translated directly into capacity. Critically, AHT improvement did not come at the expense of quality: QA scores held steady at or above the 86% program goal across in-house Service and Blended teams, with most months running 92%+ on the Services program.

A 43% Lift in Coaching Activity, with Engagement to Match

Total coaching activity increased 43% as the time freed up from manual data prep flowed directly into coaching conversations. Engagement followed: 85% of completed coaching sessions were acknowledged by agents, signaling a coaching culture where feedback was being received as feedback, not delivered into a void. Coaching effectiveness improved by 12 percentage points overall — a tangible signal that the platform was helping leaders convert coaching time into agent improvement, not just adding more sessions to the calendar.

Six Hours Back, Per Manager, Per Week

The cleanest operational outcome lives at the manager level. Pre-AmplifAI, frontline managers spent the majority of their week on non-coaching activity: data gathering, consolidation, distribution, and focus-agent identification. Post-AmplifAI, that envelope shrank by 23% — equivalent to roughly 6 hours per week saved per manager, on a 1:23 supervisor-to-agent ratio.

Multiplied across the carrier's frontline-leader population, that recovered capacity is the structural ROI of the deployment. The platform did not replace coaches. It removed the work that was preventing coaches from coaching.

Why It Worked: Effectiveness, Not Just Activity

The carrier's leadership is explicit about the lesson learned. Coaching frequency is necessary but not sufficient. Without a way to measure whether sessions are producing improvement, more coaching can simply mean more wasted hours.

AmplifAI's Coaching Effectiveness Index made that distinction visible at the leader level for the first time. A team lead with a 51% effectiveness score is not failing — they are coaching, and roughly half their sessions are converting into agent improvement. A team lead with an 80% score is operating at a different level. With both numbers visible side by side, leader development became a targeted investment, not a generic training program.

The Bilingual Sales arc — coaching effectiveness rising from 51% to over 80% in a single quarter, with quote-to-call ratio rising in lockstep — is the cleanest illustration of the model in motion. Better coaching, measured and improved, produces better performance. The platform's job is not to coach for you. Its job is to make the coaching visible enough to improve.

What This Means for Other Insurance Carriers

The lessons from this deployment generalize cleanly across insurance and other regulated, sales-and-service contact operations:

• Unify the data first, coach second. Fragmented scorecards across in-house and BPO programs make consistent coaching impossible. Single source of truth is a precondition, not a nice-to-have.
• Measure coaching effectiveness, not just coaching activity. Counting sessions on a calendar is a vanity metric. The right number is the percentage of sessions that produce agent improvement.
• Use AI to prioritize, not to replace. Next Best Coaching Actions help managers spend their limited coaching time where it will move a metric. The coaching itself remains human.
• Give the managers their week back. Most contact center performance gains do not require more coaches. They require fewer hours of manual data prep stealing time from the coaches who already exist.

For insurance contact center leaders running blended in-house and BPO operations across multiple programs and languages, the structural case is simple: coaching is the highest-leverage activity in the operation, and almost nothing in your stack today is built to measure whether it is actually working.