Customer support · Production

Swtch reduces EV driver support costs by over 50% with Retell AI voice agent Lucas

The problem

As EV adoption accelerated, Swtch's inbound call volume grew steadily and spiked sharply during infrastructure incidents, making human-only staffing both insufficient and uneconomical. Support team costs were rising 300 to 400 percent a year, yet long wait times were unacceptable for drivers standing at a charger in real time.

First attempt

Human-only staffing could not absorb call surges during infrastructure or cloud-related incidents; even with continued hiring, wait times regularly exceeded internal standards and building for worst-case scenarios would have been inefficient.

Workflow diagram · grounded in source
1
Driver calls inbound
Trigger
EV drivers call inbound support when a charging issue must be addressed immediately.
▾ source quote
“When a driver calls, the issue must be addressed immediately.”
2
Lucas AI agent handles call
Ai action
Lucas, a driver-focused AI agent built by swtch on Retell, handles high-volume repeatable requests related to accessing charging services.
▾ source quote
“Lucas is a driver-focused support agent built by swtch to handle high-volume, repeatable requests related to accessing its charging services.”
3
Route routine vs. complex
Routing
Routine inbound calls are resolved by Lucas while complex edge cases and high-touch situations are directed to human agents.
▾ source quote
“By managing the majority of routine inbound calls, Lucas ensures drivers receive immediate assistance while allowing human agents to focus on complex edge cases and high-touch situations.”
4
Resolution delivered to driver
Output
The result is a faster, more reliable support experience for drivers and a more sustainable cost structure for the business.
▾ source quote
“The result is a faster, more reliable support experience for drivers and a more sustainable cost structure for the business.”
Reported outcome

By deploying the Lucas AI agent on Retell, Swtch reduced its support cost burden by over 50 percent, drastically improved SaaS margins, and delivered faster response times and quicker resolution — with most drivers unaware they were speaking to AI.

Reported metrics
support cost growth rate (pre-AI)300 to 400 percent a year
Support cost burden reductionover 50 percent
SaaS margin improvementdrastically improve
Driver response and resolution timefaster response times and quicker resolution
Reported stack
RetellLucas
◆ Does this fit your context?

Compare to your context

Tell us your scale, team, and constraints. We'll show what changes at your size, what fails at your scale, and whether this case is a fit, needs adaptation, or won't scale to you. Free demo, no signup.

Compare to your context →
~30 seconds · free
Source
https://www.retellai.com/case-study/how-swtch-keeps-ev-drivers-moving-with-always-on-voice-support-powered-by-retell
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

By deploying the Lucas AI agent on Retell, Swtch reduced its support cost burden by over 50 percent, drastically improved SaaS margins, and delivered faster response times and quicker resolution — with most drivers un…

What tools did this team use?

Retell, Lucas.

What results were reported?

support cost growth rate (pre-AI): 300 to 400 percent a year; Support cost burden reduction: over 50 percent; SaaS margin improvement: drastically improve; Driver response and resolution time: faster response times and quicker resolution (source-reported, not independently verified).

What failed first in this deployment?

Human-only staffing could not absorb call surges during infrastructure or cloud-related incidents; even with continued hiring, wait times regularly exceeded internal standards and building for worst-case scenarios wou…

How is this customer support AI workflow structured?

Driver calls inbound → Lucas AI agent handles call → Route routine vs. complex → Resolution delivered to driver.

Related customer support cases

WHAT TO DO WITH THIS

Now compare it to your context

This case is one data point. Whether its pattern fits you depends on your volumes, your stack, and your exception load — that comparison is the step no case study can do for you.

Browse the Customer support cases →