Driver Calls Are Your Best Reliability Data

A driver support call is a moment of truth. It shows operators how their network is really performing and gives them a chance to turn a frustrated driver into a loyal one.

Recently we met with the team at one of the most progressive charging operators in Europe. They run a large, professional network and take driver experience seriously. One thing they said has stuck with me: they treat a driver calling support as a sign that something went wrong, and they work hard to reduce call volume.

They have a point. A call usually means a session failed to start, the app froze, or the payment terminal declined a card that works everywhere else. Reducing the reasons drivers need to call is the right goal, and it’s a big part of what we work on at ChargeMate.

But reducing the reasons for calls and reducing the calls themselves are two different goals. Treat every call as a failure and you miss half of what those calls are worth. Push call volume down too hard and you can end up with quieter support lines and the same broken sessions.

Why driver calls do signal friction

When I was at Volta, our dashboards regularly showed uptime above 95%, while the share of driver visits that actually ended in a successful charge ran somewhere between 70% and 85%. Driver calls were often where we first saw that gap, because the network thought the charger was fine and the driver standing next to it knew it wasn’t.

• First-time payment, especially app sign-up or card tap on finicky hardware

• Vehicle interoperability, where a specific car and a specific charger don’t agree

• Connector and cable problems that never show up as a fault code

• Unclear pricing, idle fees, or session limits

• A charger that reports “Available” but can’t actually deliver a charge

Tracking why drivers call and fixing the root causes is worth doing.

What “fewer calls” leaves behind

Call volume is a noisy metric. Fewer calls can mean fewer problems. It can also mean drivers gave up. A driver who can’t find the support number, waits too long, and drives to another network doesn’t show up in your call stats. Neither does the one-star Google review they leave on the way out.

It is also easy to reduce calls in ways that make the experience worse: bury the phone number, add another IVR menu, or push everyone to a FAQ page. The number goes down and the driver is still stuck.

The number to push down is failed sessions. If those fall, calls fall with them.

There’s a fair objection here: calls cost money. Most operators pay a call center by the minute, so “listen to more calls” can sound like “spend more.” AI support changes that math. When an AI agent picks up first, handles the routine fixes, and passes the hard cases to a person, each conversation costs far less, and what the driver tells you is just as useful as before. For one of our customers, cost per call dropped about 68%, while support went from business hours to 24/7.

What driver calls tell you

When a driver calls, they tell you in their own words what broke, where, with which car, at which charger, and what they already tried. That is often more useful than OCPP status data, because it covers problems monitoring misses: the holster latch that won’t release, the screen nobody can read in direct sun, or the vehicle model that fails the handshake after a firmware update.

Every one of those conversations can become structured data: issue type, EVSE ID, vehicle, resolution, and time to resolve. Over thousands of calls, that tells you which sites need a technician, which hardware needs a firmware fix, and which parts of your app confuse people. If the only goal is a lower call count, most of that information never gets used.

Here’s a real example. One of our customers runs more than 900 ports across the Northeast and Mid-Atlantic, mostly at unstaffed sites. Over two weeks, drivers called three separate times about the same station. Each time, the screen told them to call customer support. The backend showed the station as occupied with no active error, so monitoring never flagged it, and a remote reset didn’t hold. One at a time, those were three routine calls. Read together, they pointed to a hardware problem, and the station went on a work order instead of collecting bad reviews.

The same customer’s calls turned up two fixes that had nothing to do with hardware. About one in fifteen callers was asking about a $50 pre-authorization hold or an idle fee, which is a pricing and messaging problem, not a staffing one. And some drivers were still on an old version of the app, which took minutes to spot on a call. Now our agent checks which app a driver has at the start of the call.

We saw a similar pattern in our pilot with Mercedes-Benz HPC. Drivers reached web-based chat support by scanning a “need help?” QR code at each charger. Those conversations surfaced more than 40 distinct driver-reported issues, which the team used to update on-screen guidance and payment flows at the pilot sites. Compared with control sites, calls fell 48% because drivers could solve most problems in a lower-cost self-serve chat, 1.9 times as many drivers got help, and the visit success rate rose 5.5 percentage points.

Calls are also a chance to retain a driver

Drivers often call at a stressful point in their trip: low on range, short on time, and not sure the charger will work. How that call goes shapes their view of your network more than a routine session does.

A quick, friendly fix builds trust. So does a follow-up message letting them know the charger has been repaired. This matters a lot for new EV drivers, who are still deciding what they think of public charging and of your network.

How we think about it at ChargeMate

AI support should have two jobs:

1. Reduce the reasons drivers need to call. Find and fix the root causes behind repeat issues, and get problems in front of operations before the next driver hits them.

2. Make every call that still happens count. Don’t make drivers hunt for help. Try to solve the problem on the first chat or call, then log what happened in a form operations can use and let the driver know when it’s fixed.

It’s also why we post a real, anonymized driver call every week in our EV Driver Issue of the Week series, so others in the industry can hear what issues drivers run into.

Better metrics than call volume

Here’s what we’d track in addition to call volume:

• Calls per successful session, by site. This adjusts for traffic and shows which locations need work.

• First-contact resolution rate. Did the driver leave with a charge?

• Repeat issues per charger. If the same problem gets reported twice, it’s a maintenance issue.

• Time from first driver report to fix. How long it takes a driver’s report to reach a technician.

• Driver return rate after a support interaction. This is hard to measure without app logins, but a returning payment card or account at any of your sites works as a stand-in.

So, is a driver calling you a failure?

No. A call can point to a failure, but the bigger problem is the driver who had a bad session and never told you about it.

The operator we met is right to work hard on root causes. I’d add that the drivers who call are one of your best guides to what to fix first, and one of your best chances to keep a customer.

Want to know what your drivers are telling you, and which ones you’re losing? Let’s look at your calls together.