From Autonomous Cars to Autonomous Companies

An AI-driven company model shifts the focus from merely using technology in point-solutions to creating an enterprise platform to engage and adapt in real-time with key ecosystem players.

The thrill of driving around San Francisco in a driverless taxi

Self-driving taxis have been a common sight in San Francisco ever since Waymo (a Google spin-out) introduced them in 2020. It’s quite a thrill to step in and let yourself be led, but you get used to it after a while. Many people prefer the convenience and safety of this mode of transport. Over 40 million KMs have been covered, and every ride is input to refine the algorithms that drive the car’s performance.

I’ve long been fascinated by autonomous vehicles—not only for their potential to redefine mobility but also for their engineering marvel. Their approach to integrating hardware, software, and data to solve complex problems has inspired me to think beyond transportation. We can apply similar concepts to medical devices, but that’s a topic for another blog.

Waymo’s system architecture embeds multiple modalities to collect data on the environment, the driving, and engine performance. Sensors like cameras, motion sensors, and LiDAR gather real-time data about the vehicle's surroundings, detecting traffic signs, obstacles, and road conditions. The GPS provides precise location information for navigation and route planning. Digital maps provide detailed road layouts and traffic patterns gleaned from other cars. The system fuses all the data from these different sources in real time to feed the digital twin model of the world around the vehicle.

The vehicle’s ‘control agent’ uses this real-time data to execute a simple instruction set for steering, acceleration, and braking, ensuring the car follows the planned route safely. So, multiple ‘agents’ act in concert: the route planning agent to guide, the real-time observations from the sensor agent to adapt, and the execution of instructions by the control agent.

Now, let’s take these concepts and imagine how companies could operate with the same level of autonomy.

The autonomous company concept may seem futuristic today, just like the self-driving car of a couple of years ago, but its concepts could redefine industries sooner than we think.

Imagining the Autonomous Company

A few years ago, while enjoying a sunny afternoon in our garden, my son Aki Tas (a regular Waymo user) and I engaged in a thought experiment triggered by the first autonomous cars. We explored the idea of an ‘autonomous company’. What if an organization could operate without humans, managing every step and running almost entirely through AI?

Imagine a company getting started by purchasing a self-driving vehicle. The company’s mission is to maximize its value while adhering to the constraints of safe driving, energy efficiency, and full compliance with regulations.

In this scenario, several specialized but interacting AI agents would manage the company’s business:

The Operations Agent handles ride scheduling, payments, route optimization, vehicle health checks, charging, and cleaning at the right time and place. It connects with third-party ride-sharing services to ensure the car remains active and efficient, filling idle time with additional rides.

The Administration Agent handles financial and administrative tasks like vehicle registration, license application, tax payments, and regulatory filings. This agent keeps all formalities in order, ensuring compliance and smooth operations.

The Strategic Agent scans the business environment for growth opportunities. If demand increases, it evaluates financing options and purchases another vehicle when the time is right. With multiple cars, the agent scouts optimal locations for fleet exploitation and explores the opportunity for B2B contracts to build recurring revenue streams. For example, it could seek partnerships with air and rail services to create seamless customer journeys.

We see immense potential for transformation in the $6.5 trillion professional services industry.

Bridging the Vision: AI-First Organizations Today

Let’s move from this future vision to today. Many companies have started deploying AI, often for targeted tasks like customer service. But what if AI became foundational to the entire business model?

We see immense potential for transformation in the 6.5 trillion-dollar professional services industry. It will profoundly impact management consulting, tax, legal, finance, communications, and IT services. ?

Here’s how an AI-first services organization could work: fully leveraging the technology but keeping us humans in the loop to guide and apply judgment.

The Strategy Agent explores new market segments and business models, calibrates competition, finds complementary partners, and identifies unique market propositions and positions. It supports the short-term versus longer-term trade-offs on the journey of sustainable growth while recognizing the constraints of its environment (the dynamics of interacting with customers, consumers, partners, regulatory bodies, competition, politicians, etc.).

The Sales and Marketing Agent helps with customer acquisition and engagement. This agent identifies prospects and generates presentations and proposals customized to the customer’s needs, as gleaned from the data.

The Project Agent creates project plans based on past performance data and customer needs; this agent supports resourcing and identifies the best assets (for instance, tech stacks) for each project, integrating partner services as needed.

The Delivery Agent in an IT services company might generate code, manage data integrations, write software documentation, and handle testing, maintenance, and support.

The Back Office Agent automates traditional HR, finance, tax, legal, and IT functions, reducing administrative expenses and seamlessly scaling operations.

Below, I've included a schematic of what the architecture may look like. At the bottom is a data ingestion layer, capturing relevant data from the environment and the organization's internal systems. The Data Layer caches or stores the data, making it interoperable for further processing. In the next layer, data gets contextualized, linking the data to relevant information based on associations and models, like digital twins. The orchestration layer, which understands goals, rewards, and constraints, brings together the various agents to operate in a coordinated fashion and feed different applications for execution.

Embracing the AI-First Future

An AI-driven company model shifts the focus from merely using technology in point solutions, like reporting or customer service, to creating an enterprise platform to engage and adapt in real-time with key ecosystem players. This is similar to the journey of a self-driving car to its destination: choosing the most efficient route while adapting in real time to dynamic road and traffic conditions.

In this adaptive organization, each function is represented by agents who semi-autonomously contribute both to real-time business orchestration and the longer-term strategy. Businesses can enhance efficiency, scalability, and responsiveness by embedding AI at the core and optimizing their intended outcomes based on real-time data and guiding models.

The autonomous company concept may seem futuristic today, just like the self-driving car of a couple of years ago, but its concepts could redefine industries sooner than we think.