Companies are looking to artificial intelligence to create business value, and as MIT Sloan Management Review’s 2018 Global Executive Study and Research Report on AI shows, Pioneer organizations are pulling ahead of their counterparts. By deepening their commitment to AI and focusing on revenue-generating applications over cost savings, these early implementers are positioning themselves to reap the benefits of AI at scale.

Even a highly skilled, knowledge-based industry like management consulting — one worth $250 billion and filled with some of the smartest people on the planet — is vulnerable to the same market forces that are disrupting services everywhere, including AI.

Technology changes faster than society can keep up, a pattern now playing out with artificial intelligence. Many CEOs are taking a wait-and-see approach to AI, while others are anxious to barrel forward. In both cases, there’s little conversation about AI’s human costs. Incremental adaption makes it more likely that AI algorithms shared across organizations and geography are spreading their shortcomings. Leaders must act to mitigate these challenges if AI is to benefit society.

Modest questions about how today’s problems could be solved more effectively lead to applications of technology with easily foreseeable gains. But when people start asking bigger, bolder questions that challenge basic assumptions about how a problem has been framed, they open up space for breakthrough innovations. That’s been the pattern in many digital realms, including cybersecurity, artificial intelligence, and the internet of things.

The robots are coming! But counter to popular belief, it’s not just low-paying jobs that are at risk of automation. According to research by Scott Latham and Beth Humberd, predicting which jobs are vulnerable requires analyzing the type of value job holders deliver and the skills they use to deliver it. Workers must understand four paths of job evolution — and factors behind each path — if they hope to adapt.

The fear of robots eclipsing human power in society, in particular the workforce, has persisted for decades despite the moderate progress of artificial intelligence systems. For professor and robotics researcher Ken Goldberg, a hybrid human-machine workforce is much more likely to take shape, and in many industries, it has already begun.

Not many companies are there yet, but there’s a developing framework for what it takes to lead an AI-driven company. Leaders at the forefront of AI have seven key attributes: They learn the technologies; establish clear business objectives; set an appropriate level of ambition; look beyond pilots and proofs of concept; prepare people for the journey; get the necessary data; and orchestrate collaborative organizations.

To deliver ROI and productivity, firms must overcome the barriers to implementing AI technology.

AI provides access to unstructured information and enables organizations to learn faster and offer new services.

AI’s largest and most enduring contributions will be in non-technology sectors, and many of them will come from China. Buoyed by the country’s latest five-year plan and enabled by centralized data, Chinese companies are investing aggressively in AI and adapting their business models to make the most of AI’s potential, but unclear business cases and bottlenecks due to lacking technical capabilities hinder adoption.

Software continues to “eat the world,” but often at the cost of user privacy and trust. In order to grow responsibly, companies must move from a “can we” to a “should we” framework when bringing products to market.

Less than 5% of companies are using AI to reinvent how they do business, but the competitive intensity surrounding the technology suggests that a wait-and-see strategy could be a costly mistake. To get a share of the global profit pool of $1 trillion that AI will produce by 2030, the McKinsey Global Institute says companies should begin adopting it at scale within the next three years.

When pundits talk about the impact that artificial intelligence will have on the labor market, the outlook is usually bleak, with the loss of many jobs to machines as the dominant theme. But that’s just part of the story — a probable outcome for companies that use AI only to increase efficiency. As it turns out, companies using AI to also drive innovation are more likely to increase headcount than reduce it.

Virtually all human achievements have been made by groups of people, not lone individuals. As we incorporate smart technologies further into traditionally human processes, an even more powerful form of collaboration is emerging.

Businesses should use AI technology as a tool to improve customer service and processes, not just to imitate humans.

AI can do much more than augment human performance in the workplace.

Using industry examples and findings from the McKinsey Global Institute’s research, presenter Jacques Bughin describes the current state of AI and offers strategies for maximizing its potential.

Companies are adopting artificial intelligence at an accelerated pace — and learning that developing and deploying AI is not like implementing a standard software program. Before diving into AI systems, companies should consider three principles that can greatly improve the chances for a successful outcome. First, they need to recognize that humans and machines are in this together. Second, they need to teach the AI systems with a lot of data. And third, they need to continually test what the systems have learned.

As companies rely more and more on AI-driven chatbots to communicate with customers, they need to consider ways to make computers’ communication and understanding better and more natural.

Digital technologies will increase the high levels of ambiguity that executives must navigate. Aspiring leaders may respond by ignoring the challenge, which isn’t sustainable. A better response is to harbor healthy skepticism of the digital technologies they champion, develop values that will lead to better decisions, and work to institutionalize those values at the organizational level.