Upstream on Air: After-Sales Quality Report, Zooming in on the Power of AI – Audiobook Edition

PQD Audiobook: The Deep Dive on Proactive Quality Detection

[00:00:00]
Welcome to The Deep Dive, where we cut through the noise and get straight to the core insights. Today, we’re tackling a topic that’s keeping automotive executives up at night: the rising pressure on global OEMs, such as surging warranty costs, escalating recalls, and supply chains that won’t settle.

Add in geopolitical uncertainty and, perhaps most importantly, rapidly rising consumer expectations, especially as EVs and software-defined vehicles (SDVs) take off. These cars are becoming software-centric, and that’s fundamentally redefining how we think about quality.

[00:00:34]
Exactly. The real challenge is that traditional, reactive quality strategies simply don’t work anymore. Many OEMs have relied on them for decades, but it’s like trying to fight a wildfire with a garden hose. What we need now is more like real-time satellite imagery to spot embers before they ignite.

[00:00:52]
In this deep dive, we’ll explore how integrating connected vehicle data and advanced AI isn’t just a “nice to have”: it’s becoming essential. Our goal? To uncover surprising insights and actionable strategies drawn from a major new report, tailored specifically for quality, warranty, and AI leaders.

EVs: Fewer Moving Parts, More Problems?

[00:01:09]
Let’s start with a common assumption: EVs should have fewer quality problems. They’re mechanically simpler, with fewer moving parts, so… less to break, right?

[00:01:26]
That’s the intuition. But software glitches (think of your phone or laptop) can be far more frustrating than a mechanical issue. So, does the data support the idea that EVs have fewer problems?

[00:01:47]
Surprisingly, it doesn’t. The H1 2025 Proactive Quality Detection (PQD) report reveals some eye-opening trends. According to JD Power’s 2024 research, battery electric vehicles (BEVs) have 48% more quality issues than internal combustion engine (ICE) vehicles.

[00:02:06]
48%! That completely debunks the “fewer moving parts = fewer problems” theory.

[00:02:21]
Exactly. The quality battleground has shifted from mechanical reliability to software and system integration. BEV owners and plug-in hybrid (PHEV) drivers report taking their cars in for severe issues three times more often than gas vehicle owners.

[00:02:56]
And these issues aren’t even primarily about the electric drivetrain, they’re tied to software, infotainment, connectivity, and system integration. It’s like the early days of PCs: the hardware was fine, but the software kept crashing, except now it’s happening at 70 mph.

Speed vs. Stability: Innovation’s Hidden Cost

[00:03:49]
Automakers are in a race to innovate, packing more software and features into vehicles faster than ever. But that speed comes at a cost: rising quality risks rooted in software complexity.

[00:04:08]
Traditional methods like warranty tracking, service records, recall history just can’t keep up. The systems are too complex, and the data is too slow or incomplete.

[00:04:28]
That’s where real-time connected vehicle data and AI come in. Telematics, DTCs, sensor streams, combined with AI, give you a live digital heartbeat of the vehicle. AI can sift through the noise to detect early warning signs, spot hidden patterns, and trigger action before issues explode into recalls.

The Data Doesn’t Lie

[00:05:20]
Here’s a stat that should make every executive lean in:
70% of all recalls, and nearly 90% of EV recalls, could have been detected earlier through connected vehicle data and AI.

[00:05:42]
Think of the cost savings, the avoided disruptions, the brand protection. This isn’t tweaking the process. It’s transformation.

Real-World Examples: What Could Have Been Prevented?

[00:07:13]
Let’s bring this down to earth with a couple of examples:

October 2024: Nearly 200,000 plug-in hybrids recalled due to a high-voltage battery defect and fire risk. Traditionally, this would only be caught through warranty claims or, worse, actual fires. With connected data, AI could have picked up subtle signs like irregular charging behavior or thermal anomalies much earlier.

November 2024: 150,000 EVs recalled due to ICCU failure causing sudden power loss. AI could’ve flagged early signals: specific DTCs, undercharged 12V batteries, temperature spikes, giving OEMs a chance to act fast and precisely.

The Bigger Picture: Trends from 5,000 Recalls

[00:09:17]
Zooming out, the report analyzed 5,000+ recalls and 30,000+ consumer complaints from 2020–2025. Key trends:

• EV recall share skyrocketed from 2.5% in 2020 to over 14% in 2025 YTD, far outpacing sales growth.
  High-impact recalls (affecting large volumes) grew from 9% in 2020 to 14% by 2022 and stayed elevated.
  EVs consistently showed higher proportions of large-scale recalls compared to ICE vehicles.

And the positive: Detectability keeps improving.

• 2020: 69% of recalls potentially detectable via connected data
  2025: 75% overall, over 90% for EVs

Blind Spots & Supply Chain Risks

[00:11:41]
But not all recalls show up in the data. Supplier-related recalls had a much lower detectability rate: just 58–63%. That’s a growing blind spot as vehicles become more modular and reliant on third-party components. It calls for deeper supplier integration and new forms of signal sharing.

From DTCs to Digital Co-Pilots

[00:13:03]
Diagnostic Trouble Codes (DTCs) are still the MVP, showing up in 49% of EV recalls. But EVs typically need more signals per issue (1.8 on average vs. 1.6 for ICE) due to their software complexity.

Enter Agentic AI.
These are AI systems that act like intelligent co-pilots. They don’t just raise flags, they act. They analyze, correlate, even initiate diagnostics or interventions. They learn from structured and unstructured data (like call center notes, OTA updates, complaint logs), giving teams a 360° view.

Closed-Loop Quality: The New Standard

[00:15:00]
Agentic AI can:

• Analyze warranty and service data at scale
• Spot trends across geographies and platforms
• Track OTA update effectiveness
• Forecast parts demand
• Feed insights back into design, production, and validation

This creates a closed-loop quality system—where real-world usage constantly informs development.

Rethinking Quality Itself

[00:16:52]
The takeaway is clear: the future of quality and competitiveness demands a proactive, data-driven, AI-powered approach.

Yes, it reduces costs. But more importantly, it builds trust. In a world where vehicles are computers on wheels, capable of reporting on their own health, OEMs must learn to listen.

[00:17:28]
And maybe, just maybe, the definition of “quality” itself is evolving, becoming continuous, adaptive, and rooted in data.

Thanks for joining us on The Deep Dive.