The new role allows me to focus entirely on AI, not just from North America, but also from a global perspective. Toyota invests heavily in AI research, and my goal is to bridge research and production to maximize these investments. I focus on scaling AI across all Toyota group companies, ensuring broader adoption and innovation. Many teams use AI to boost productivity and create new experiences, and I’m excited about driving innovations that wouldn’t exist without AI. Ultimately, my role is to shepherd AI from research to real-world applications, delivering tangible benefits to our business and customers worldwide.

Traditionally, organizations relied on relational and NoSQL databases as core data sources, primarily handling transactional data like orders and service repairs. However, the landscape has evolved dramatically. Data is no longer limited to structured sources — it now includes unstructured elements such as survey responses, social media comments, call center transcripts, repair manuals, owner’s manuals, images, audio, and even code.

GenAI has expanded the possibilities by integrating this diverse data, adding critical real-world context from internal documents and code repositories. Historically, code wasn’t viewed as a true data source, but AI-driven automation is changing that perspective.

For instance, AI tools can now generate unit tests with a simple right-click, significantly improving code quality and developer efficiency. Since developers often avoid writing unit tests, automating this process ensures better coverage, reducing errors and increasing productivity.

This shift highlights the need for a broader AI-driven approach beyond traditional data and analytics. Toyota has recognized this by establishing a dedicated enterprise AI function to analyze and leverage all forms of content — structured and unstructured. By incorporating images, voice, and code into AI models, the company aims to extract deeper insights and drive greater operational efficiency.

I can’t give an exact number, but our productivity has increased by at least 20%. A study we conducted a year ago showed this improvement, and it’s likely even higher now. I always ask for data to validate investments, and these AI tools have proven their impact. They’ve been incredible for productivity, improving code coverage, and automating repetitive tasks. This allows developers to focus on solving complex problems rather than spending time on routine coding tasks.

One major challenge is getting new employees to perform like seasoned experts. We’ve leveraged AI to ingest repair manuals, trouble tickets, and video documentation — some newly digitized — into a multimodal LLM.

This has led to knowledge accelerators that significantly reduce diagnosis time. For example, resetting calibration on line-level machinery requires understanding G-code, an antiquated and complex system with massive, non-searchable manuals. Previously, a process engineer spent seven hours troubleshooting an issue; now, with AI assistance, it takes just 15 seconds.

Our AI-driven tools are transforming how we diagnose and repair equipment, making critical information instantly accessible. By integrating these capabilities into smart assistants, we provide new employees with the knowledge of a 20-year veteran, accelerating their expertise. We’ve also developed custom AI bots tailored for manufacturing workers, delivering great success. Additionally, it can assist in creating more natural and intuitive human-machine interfaces, enhancing in-vehicle experiences.

When considering the immense computing power and costs involved in building custom models, there needs to be a strong justification. While I can’t share specifics, I can say that our research team is deeply exploring this area and evaluating it as a potential option.

There are many models available, and DeepSeek’s key value proposition is its cost-effectiveness while delivering similar results. However, we’re taking a wait-and-see approach. We’ve had significant success with other models we use, so it’s not something we’re actively pursuing right now.

AI is evolving rapidly — DeepSeek was the big trend a few months ago, and next month, something new will emerge. The landscape shifts so fast that if you take a short break, you return to a completely different world. While it’s exciting, it also requires caution.

With innovations emerging globally, it’s not only important to stay aware but also thoroughly vet new technologies. At Toyota, we focus on models that demonstrate real value. We keep an eye on new developments, study them, and decide when the time is right. At this point, we haven’t explored DeepSeek extensively, but we continue to monitor advancements in the field.

AI agents could enhance various aspects of our operations and customer interactions. In manufacturing, they can monitor equipment health in real time, predicting maintenance needs before issues arise, thus minimizing downtime.

In customer service, AI agents can provide personalized support, answering queries and offering tailored recommendations, thereby improving customer satisfaction. Additionally, within vehicles, these agents can learn driver preferences and adjust settings proactively, creating a more personalized and seamless driving experience. We see this as a game-changer but we also need to consider the right level of guardrails to ensure that there is always a way to measure the quality of the output coming from the AI.

I envision AI becoming an integral component across all facets. In vehicle design, AI will enable rapid prototyping and optimization, leading to more innovative and efficient models. Manufacturing processes will become increasingly automated and intelligent, with AI systems predicting maintenance needs and optimizing production lines for efficiency and quality.

The supply chain will benefit from AI-driven analytics, enhancing demand forecasting and inventory management. In sales, AI will provide personalized customer experiences, tailoring recommendations based on individual preferences and behaviors. Our dealers and team members will be supercharged with intelligence amplification tools that will help them all do their jobs more efficiently and allow them to focus on higher-value work. Overall, AI will drive a more connected, efficient, and customer-centric automotive ecosystem.

We consider several key factors. The potential for cost savings through process optimization and automation is crucial.

We assess how the AI solution can enhance revenue, either by improving product offerings or creating new business models. The impact on customer satisfaction and retention is also a significant consideration, as positive customer experiences can lead to increased loyalty and sales.

Additionally, scalability and the potential for the AI solution to be applied across multiple areas of the business are evaluated to maximize the overall benefit. We always keep TCO (total cost of ownership) as a way to ensure that we are seeing the ROI we are expecting and make sure that we lead with solving the business objectives. I like to steer my team away from tech for the sake of tech and always drive the focus on ROI and TCO.

I’ve been fortunate to have mentors like Zack Hicks (former Toyota Chief Digital Officer) and Vicky Colf (former Warner Bros. CTO) who have profoundly influenced my career. One of the most impactful pieces of advice I’ve received is to, “Embrace challenges as opportunities for growth.”

This mindset has encouraged me to take on complex projects and view obstacles and hard decisions as learning experiences, fostering innovation and resilience in my leadership approach. Vicky once told me “There are no good decisions or bad decisions in life. Each decision leads you to an opportunity to learn from, and it is up to you to learn what you can and apply that learning to future decision points.”

Yes! We own a 2021 Toyota Sienna and a 2024 Lexus RX 350. But if I had to pick a favorite, it would be my first car — a 1986 Toyota Camry. That was the car I drove through college, so it holds a lot of great memories.

I had the chance to drive a Waymo, and I was truly impressed. San Francisco’s busy streets can be challenging, but the vehicle handled them exceptionally well. At one point, a pedestrian suddenly walked in front of the car, and it stopped instantly, allowing them to cross before smoothly continuing on its way. It felt as if a highly skilled and attentive driver was behind the wheel. I was amazed by how intelligently the system responded to real-world situations.

I’m not the best person to answer that. My colleague is leading our work on diffusion policy-based autonomous driving and could provide much more detail. However, I can confirm that we are leveraging GenAI to accelerate our autonomous driving capabilities.