According to the recent McKinsey report on the state of AI, 78% of surveyed organizations use AI in at least one business function, up from 55% in 2023. But while most companies are still experimenting with AI implementations, some visionaries have already transformed entire industries with groundbreaking solutions that save millions of dollars and serve hundreds of thousands of users daily.
Meet Ankit Agarwal – the architect behind AI systems that process orders for McDonald’s and Starbucks customers worldwide, automate warehouse operations handling millions of Amazon packages, and have transformed how over 450,000 restaurants run during the most challenging time in hospitality history. As a founding engineer of Voice AI at DoorDash and a former senior software engineer at Amazon, Agarwal has spent more than a decade turning ambitious AI ideas into real-world solutions that generate billions in economic value.
What does it take to build AI systems that global brands trust with their core operations? How do you scale from prototype to enterprise deployment, serving millions of users? And what critical mistakes do most companies make when implementing AI that prevent them from achieving transformative results?
In this exclusive interview, Agarwal shares the insider strategies behind creating AI solutions that operate at scale, recounts hard-won lessons from leading teams of 50+ engineers on cutting-edge projects, and explains why the next wave of AI adoption will either make or break small businesses and what leaders can do today to position themselves on the winning side.
In recent years, AI and automation have been transforming industries in an unprecedented way. From your perspective, what are the most significant opportunities and challenges for businesses adopting AI today?
As AI adoption grows, its benefits become obvious both for enterprise-level companies and for smaller businesses. For instance, AI-driven innovation allows companies a wide array of opportunities for optimizing complicated processes, such as keeping warehouse inventory under control or processing multiple orders. However, for AI integrations to be beneficial, businesses need a conscious approach to adopting new technologies. Operational processes should be studied first, so the innovations do not remain superficial and become truly integrated into business operations. Consequently, investment in the employee workforce education and adaptation should go hand in hand with implementing new technological solutions.
Your career includes several projects of integrating innovative solutions into business workflows. It began at Samsung Research Institute and took you to Amazon and DoorDash, where you led transformative AI projects. How did those experiences shape your approach to Innovation?
I perceive every task I work on, regardless of its scale, as an opportunity to master new technologies and acquire new knowledge and experience that I will take further to my next projects. For instance, my early experience at Samsung offered me international exposure and taught me to work in cross-functional environments while holding up to industry standards. These skills proved useful later, at Amazon and DoorDash, where I worked on large-scale and impactful projects. Another key skill that proved useful many times throughout my career was identifying pain points and finding the most efficient way to apply novel technologies to resolve them.
Your work at DoorDash, the leading food delivery company in the US, is a clear example of such an approach. There, you led the development of GenAI-powered voice automation systems and ordering platforms that are now used by several global brands. What do these innovations tell us about AI reshaping customer experience and operations?
AI-based solutions possess the potential to resolve the problems restaurants and food deliveries are facing today, such as labor shortages, along with increasing workload. The AI-powered voice system allowed restaurants to automate voice orders and customer service, reducing the staff workload and improving service speed. The voice-ordering system integrated with POS terminals was implemented for phone-ordering in 2023 and then expanded to drive-throughs in 2025, became one of the pioneering solutions of this kind, and was later adopted by major food chains like Donatos and Chicken Express. This example demonstrates that AI-based solutions, when done right, benefit customers, businesses, and employees alike, as they reduce workload and improve customer service.
Your work for Amazon, the global retail giant that processes millions of orders per day, further demonstrates that automation is not limited to processing orders and can be integrated into different aspects of the business operation, from warehouse management to fraud detection. What lessons did the large-scale projects teach us about scaling AI projects for impact?
Compared to the previous example of implementing AI in customer-facing services at DoorDash, the projects I have developed for Amazon mainly focused on the company’s internal processes. The automation required both installing new hardware and sensors such as cameras and scales for tracking, and developing software for processing this newly acquired data, which was used for real-time warehouse management, predictive analytics, and efficient logistics. As a result, the system allowed for a reduction in labor costs and minimized errors immediately, but also provided the necessary resources for e-commerce growth that followed in 2022 and beyond.
As a staff software engineer at DoorDash, you shaped an efficient team of several dozen engineers who work together to create innovative solutions. What advice would you give aspiring tech leaders who want to drive innovations at their organizations?
The main lesson to draw from this experience is that collaboration is crucial for innovating and achieving technical excellence. For big and small teams alike, it is necessary to establish an environment where people with different backgrounds can voice their perspectives and share ideas. Moreover, stay in touch with people in your company besides your immediate team: understand what impact your innovation will have on their workflows and ensure you address the challenges that may arise from that. Avoid innovation for the sake of it: research your user pain points and focus on them. For instance, understanding the challenges small restaurants faced helped us to create a platform that is actively used by them.
As AI-driven solutions like those we have discussed proliferate in different industries, what do you think about their impact in the near future? Will they remain a prerogative of enterprise-scale companies, like Amazon, or do they have the potential to become more accessible for smaller businesses as well?
As AI technology advances, it becomes more accessible for smaller businesses, for example, through cloud-based and low-code solutions. At DoorDash, our work showed how AI-driven tools like voice automation can empower smaller restaurants to compete in a digital-first economy. Looking ahead, I want to advance AI-driven restaurant automation even farther. Combining it with logistics and autonomous delivery innovation, I plan to scale these platforms worldwide and make the service more accessible for underserved communities, driving economic and social equity
