Speedway Motors unifies distributed support to eliminate manual audit and rebuild agent trust

Every spring, call volume at Speedway Motors picks up as car builders and racers get ready for the season. Many of those calls are highly technical, with agents spending five or six minutes helping customers find the right part, check compatibility, or work through a backorder.

For Becky Ziebarth, Customer Experience Sales Chapter Lead, those conversations also represented a sales opportunity. But with agents working remotely across five states, supervisors had no consistent way to see whether a helpful conversation ended in an order.

Speedway Motors turned to Level AI to analyze every customer conversation, identify missed sales opportunities and service gaps, and provide supervisors with the evidence they needed to coach agents consistently.

Spreadsheet failures and small QA samples leave frontline teams without clear feedback

Speedway Motors handles roughly 35,000 customer conversations each month, with phone calls accounting for 90% of the volume.  Agents work remotely across Nebraska, North Carolina, Arizona, Michigan, Florida, and Texas. Their work spans Five9 for calls, Calabrio for workforce management, and a legacy order-entry system.

Previous software investments left these systems disconnected. To monitor quality, supervisors had to manually listen to Five9 recordings and score them on individual spreadsheets, evaluating just three calls per agent each month. That sample covered roughly 1 percent of interactions, leaving supervisors with little sense of what was happening across the rest of the calls.

Becky had to manually pull these scattered sheets into a master company report, which took hours and led to constant reporting errors.  In one instance, a supervisor sorted a shared spreadsheet the wrong way and disrupted months of agent coaching records in a single afternoon.

The small sample also made it difficult to see what was happening on sales calls. Agents often answered complex technical questions well but ended the conversation without offering to place the order. When an item was unavailable, supervisors could not easily see whether the agent suggested an alternative.
The same visibility gap affected other parts of the operation. Missed address checks created payment holds and unnecessary return shipping labels, while remote agents received feedback only in occasional emails, with no clear view of their overall performance.

Synchronized screen recordings and automated quality assurance bring more context-driven feedback

Speedway Motors evaluated Level AI by reviewing historical audio recordings alongside synchronized agent desktop screens.

The team quickly saw that transcripts alone did not capture the full technical context of a support interaction. A transcript showed whether an agent communicated well, while the screen recording showed whether they checked internal knowledge bases, searched alternate part numbers, or completed the order correctly in the DOS terminal.

The deployment also improved payment security through automated redaction models that protected sensitive customer information, including credit card numbers, while preserving complex automotive part numbers.

Leadership put Level AI to work across four areas:

• Documenting veteran mechanical expertise: Patrick used conversation search and transcript extraction to capture technical explanations from calls handled by retiring engine specialists. The team turned these real troubleshooting steps into internal reference guides for newer agents. Supervisors Tom and Emily also organized standout recordings in the Library feature, creating searchable folders for empathy, upselling, and part-number verification.
  
  

• Tracking the sales ask on inbound inquiries: Automated conversation tags check whether agents offer to make the sale once technical questions are answered. A separate tag tracks whether agents recommend alternative parts during inventory shortages.
  
  

• Catching fulfillment and address errors before shipment: Automated rubrics check whether agents confirm shipping addresses and repeat order totals before ending an interaction. This catches the mistakes that previously slipped through manual spot checks and reduces billing holds and return labels tied to customer return calls.
  
  

• Replacing ad-hoc emails with structured coaching files: The team replaced spreadsheet tracking with structured coaching templates. Supervisors categorize development areas using dropdown reason codes such as missed active listening, unfulfilled commitments, and incorrect transfer routing.

To build more consistent evaluation standards across the distributed team, Ziebarth also introduced regular calibration sessions. Supervisors score the same recorded calls against a moderator benchmark, reaching an initial 73% scoring alignment.

Frontline agents now have direct access to their own dashboards, including InstaScore results, manual reviews, and coaching notes.

Automated quality management cuts administrative time by 90 percent and secures agent trust

Moving QA out of spreadsheets reduced the administrative burden on Speedway Motors while giving agents and supervisors a shared view of performance.

• 90% administrative time savings: Becky reduced her quality management reporting workload by 90%, while supervisors across the organization cut routine audit work by 50%.
  
  

• 99.3% evaluation acceptance rate: Remote agents accepted 1,092 out of 1,100 automated and manual evaluations without filing a single dispute.
  
  

• Quality scores above 90%: Frontline agents maintained quality scores above 90% across both InstaScore rubrics and supervisor audits covering 35,000 monthly interactions.
  
  

• Stable handle times and sentiment during volume surges: Customer sentiment stayed between 6.7 and 6.9 out of 10, with negative sentiment capped at 7%. Average handle time remained between 315 and 344 seconds during peak seasonal volume.
  
  

• More transparent performance management: Frontline agents gained daily access to their individual performance across 30 metrics, replacing delayed feedback emails with direct access to their results.

  
  

Speedway Motors replaced manual spreadsheet work and small QA samples with full visibility across customer conversations.

Frontline agents can monitor their own performance, supervisors can coach using synchronized audio and screen recordings, and the company can preserve decades of automotive knowledge before experienced specialists retire.

As Patrick put it, “Now agents can really see their performance across all of our major metrics, which is pretty cool. It’s been a long time since they’ve been able to see that.”