Digital technologies have had a profound impact on telecommunications, financial services, media, transportation and other industries. Computers gave us immense calculation capabilities and software allowed us to execute sophisticated algorithms. Yet the impact of artificial intelligence (AI) on the business world over the next 10 years will be much more disruptive.

In previous eras, computer hardware and software in form of application code had to be created by people. By contrast, AI can learn directly from data and its own activity, very similar to how people learn and gain experience. Many talented scientists are working on Artificial General Intelligence (AGI) and we expect it to become available by 2030. AGI will be capable of learning and doing any human task.

One of the reasons for the ongoing explosion in AI capabilities is the availability and affordability of computing power. In the cloud, we already have access to transactional capacity equivalent to the human mind’s ability to process 1016 operations per second. Computers delivering this number of computations per second will become affordable by 2025.

Over the last five years AI software algorithms have made significant breakthroughs. Computer vision achieved human parity in 2015; computer speech recognition reached human level accuracy in 2017; and natural language understanding exceeded the human baseline in 2021. Visual processing, touch and hearing occupy approximately 40% of human brain capacity and AI now has mastered these skills. In the next 10 years, with AGI it will master the remaining 60% of human brain capabilities.

AI already went much beyond the above stated input functions (vision, hearing, language) and is showcasing advanced cognitive capabilities in comprehension and decision making. In 2020 Tesla released “Full Self-Driving" software and Waymo introduced autonomous taxis in Arizona. These cars are now on public streets, while in the US you need to be 16 years old to be legally allowed driving. In few years, the self-driving cars will prove to regulators that they are more advanced than the best adult drivers and can drastically reduce the number of accidents on our roads.

Autonomous vehicles received large investments because they represent major transformations in transportation. Tesla robo-taxis are expected to operate in few years in the US at a cost point of $0.20 per mile, which is three times cheaper than capital expense of owning a car at $0.60 per mile and ten times cheaper than today’s ride-sharing cost with Uber or Lyft at $2.00 per mile.

As we know, the exponential nature of technology evolution brought us to where we are today, and the advancements will only continue at an accelerating rate. With computing power becoming more and more affordable, much more powerful artificial intelligence will be applied with great returns in many other domains. A comprehensive study on the future of employment by Oxford University estimated that technology will automate over 50% of current professional occupations. A McKinsey report on the future of work determined that 60 percent of occupations can automate at least one-third of their activities.

Automation of knowledge work is the next major opportunity for AI. Another McKinsey report on the impact of disruptive technologies defined it as “Intelligent software systems that can perform knowledge worker tasks involving unstructured commands and subtle judgments”. McKinsey estimated potential economic impact in the range of $5.2 trillion to $6.7 trillion based on the salary spend of over 230 million impacted knowledge workers, one third of the global knowledge worker population.

For advanced robotics, the number of impacted manufacturing workers was estimated at 320 million. This is a bigger employment pool, but due to the lower group wages the estimated economic impact was at least 30% lower than for the knowledge workers. The economic impact of autonomous vehicles was estimated at least 50% lower than for the knowledge worker category. It was based on $4 trillion global automobile industry revenue.

The closest transformational technology based on the magnitude of impact was probably the Internet. Yet the low penetration of personal computers and rudimentary state of programming techniques made it expensive to build and host web applications, which didn’t allow it to capitalize on the power of mass connectivity. It took about 20 years to achieve the projected economic impact. With knowledge worker automation, though, there is no technical constraints on its utility.

Today, knowledge workers account for over 1 billion jobs, or one third of the global workforce. With the increased automation of physical labor, the knowledge worker share of the global working population will continue to grow. Over the next ten years, AI-based automation will be able to handle much more than 30% of the knowledge worker tasks. This means that the McKinsey economic impact estimate may prove to be higher in coming years.

While technically, AI is already capable of automating a good portion of today’s knowledge work, it will take time to create related products, and sell them to customers. We estimate that adoption of knowledge worker automation solutions will be very dependent on two aspects: 1) the solution provider’s ability to transfer knowledge from experts to machine; and 2) the time it takes to create market visibility and complete the enterprise sales cycle (the largest segment of the overall AI market). Typical sales cycles in the enterprise segment are about six months and buying decisions are often done within the annual budgeting cycle.

By taking into account the cost of intelligent software, it will be at least five times cheaper than the salary of knowledge workers. Applying 20% to the $5 trillion estimated by McKinsey knowledge worker salaries, we get to a $1 trillion global intelligent software market size of knowledge work automation by 2030.

The intelligent software solutions created in the next five years will be adopted in the following five. The top providers of knowledge worker automation will capture the biggest market share in the same way as other digital markets were conquered. We think a small number of leaders will capture more than 50% of the market, in the same way as in ecommerce, cloud computing and electric vehicles. The goal should be to move fast to assume leadership in several industrial and functional domains, where the solution can be built faster, and the sales cycles are shorter.

It is important to realize that never in the history of humankind have we had access to such powerful technologies as we do today. There have never been before new technology markets of the size and potential we are discussing here. This is a once in a life-time opportunity to drive innovation and and discovery. Only the winners of this next human race will be in a position to compete in the following AGI era of super-intelligence.

Microsoft says its new computer now vision system can outperform humans

IBM vs Microsoft: 'Human parity' speech recognition record changes hands again

Microsoft DeBERTa surpasses human performance on the SuperGLUE benchmark

The Vision Thing: Mainly in the Brain

Tesla Puts "Beta" Version of Full Self-Driving Capability In Hands of Select Few

Robot take the wheel: Waymo has launched a self-driving taxi service

The Future of Employment: How susceptible are jobs to computerisation ?

Jobs lost, jobs gained: What the future of work will mean for jobs, skills, and wages