With over two decades of product leadership experience spanning FAANG companies in the Silicon Valley in California and the innovation-rich automotive sector, Alankar Agnihotri has witnessed firsthand the transformative power of artificial intelligence. Nowhere has this impact been more personal—or more profound—than in the evolution of the in-vehicle assistant.
Today, cars aren’t just about horsepower or sleek dashboards; under Alankar’s leadership, they’re becoming intelligent co-pilots—context-aware, voice-responsive systems that help drivers not just get from A to B, but do so safely, comfortably, and efficiently. Drawing on his experience at the intersection of embedded AI, real-world constraints, and human-machine interaction, Alankar has uncovered several critical insights about building next-gen automotive experiences.
Lesson 1: AI in Cars Must Be Mission-Critical—Not Just Conversational
Alankar emphasizes that today’s automotive assistants must go far beyond playing music or checking the weather. He led the development of a next-gen in-vehicle assistant that blended voice, gestures, and sensor inputs for a truly multimodal experience.
What his team discovered was crucial: AI in cars isn’t about convenience—it’s about confidence. Any lag, unclear feedback, or incorrect output instantly undermines user trust. And in a fast-moving vehicle, that trust isn’t optional—it’s vital.
Lesson 2: True Intelligence Requires Context—Not Just Commands
Another insight Alankar championed was the importance of contextual awareness. Instead of just following commands like “lower the temperature,” his team asked: why is the user making that request?
By integrating biometric sensors and external data, they built a system that could proactively adjust cabin settings. However, the real breakthrough came when the assistant began explaining why it took certain actions. Transparency, Alankar realized, was key to building user trust.
Lesson 3: Voice Alone Isn’t Enough—Multimodal Interfaces Are the Future
Unlike home assistants, in-car environments are noisy and distraction-prone. Alankar led efforts to build multimodal interaction models—layering voice commands with visual feedback and touch overrides.
These efforts weren’t just about technology—they required orchestration between engineers, HMI designers, sound experts, and even behavioral psychologists. In Alankar’s words, great cockpit assistants aren’t just engineered—they’re choreographed.
Lesson 4: New Metrics for a New Kind of Product
Traditional product metrics like MAUs and retention didn’t fully capture the assistant’s success in a vehicle. Alankar’s team pioneered new metrics like:
- Interruption rate
- Correction frequency
- Override rate
These indicators gave deeper insights into user trust, intent recognition, and how well the system supported—not distracted—the driver.
The Road Ahead: From Assistant to Co-Pilot
Alankar envisions a future where in-vehicle assistants manage not just entertainment, but also navigation, energy consumption, and even emotional state. The role of the PM, he argues, must evolve—requiring fluency in system thinking, behavioral science, and storytelling.
“We’re not just adding features,” Alankar notes. “We’re crafting relationships between humans and machines.”
Final Thoughts
Under Alankar Agnihotri’s product leadership, the automotive industry is advancing from machines that move us to machines that understand us. The AI-powered assistant, once a novelty, is now central to that transformation.
As he often says, the future isn’t just autonomous—it’s collaborative. And the products we build must be ones we’d trust our own families with.
