It's already so difficult keeping up with the pace of change in artificial intelligence. So how will AI change company productivity, business models and ways of working between now and 2030? And what will best practices for getting the most out of AI look like?

To explore expectations in these areas, IBM and Oxford Economics commissioned a survey of more than 2,000 senior business leaders across 33 countries and 23 industries, along with in-depth interviews and real-world case studies. The key findings are presented in IBM's latest report, The enterprise in 2030: Engineered for perpetual innovation.

Here are five key takeaways from the report on how the most successful firms will be getting the most value from AI in five years' time.

1. Make big bets on AI and do it faster than their rivals

By 2030, 79% of executives say AI will contribute significantly to their revenue, up from just 40% today. But only 24% can clearly identify what their main sources of revenue will be in five years' time. It's very likely that businesses will need to embrace the unknown, including revenue streams and business models they can't even imagine today.

What's also clear is that enterprises won't be able to rely on steady progress. They'll very likely need the ability to adapt to changes and disrupt their markets quickly. 55% of executives expect that to gain a competitive advantage in 2030, they will need to depend more on speed of execution than on making perfect decisions.

That's because AI will level the playing field, allowing start-ups and new entrants to compete with larger established businesses while moving at a much faster pace.

To succeed, companies will need to be able to roll out big disruptive strategies - and do it fast. They will rely on AI-powered market scanning systems to identify future growth drivers. And they'll use AI to "stress-test" and refine new ideas through rapid experimentation.

There will be a need to develop "organizational intelligence" to help businesses recognize patterns, anticipate shifts and give leaders the confidence to "place bigger, smarter bets ahead of competition". In fact, AI will directly help leaders make decisions, with executives expecting that 25% of enterprise boards will have an AI advisor or co-decision maker by 2030.

2. Use today's AI gains to fund tomorrow's transformation

AI and automation technologies will help firms increase productivity, allowing them to allocate more resources to innovation and drive growth.

To underline this, 70% of executives say they're looking to use the value gained from AI to fund growth across their organization. The productivity gains from AI won't just help reduce costs; they will actively help companies drive revenue growth.

According to the survey, organizations that focus on three key areas can increase their expected productivity gains by 54% by 2030. The three areas are 1) Integrating AI into products and services, 2) Designing AI-first tasks, and 3) Using more sophisticated AI models.

The investments in innovation and growth can help firms transform business models and develop new revenue streams, which in turn can fuel further growth. In this way, the report's authors believe we'll see an "AI first flywheel effect".

Take the example of today's car manufacturers, who are using AI to optimize supply chain operations and reinvesting those gains into developing AI-integrated vehicles. They are investing their gains from AI into innovations that enable cars to learn drivers' preferences over time, predict maintenance needs and improve their overall customer experience.

These innovations create the opportunity to turn cars into platforms to deliver software-driven mobility. So, although digital and software-related revenue accounts for about 15% of total revenue today, it's expected to increase to 51% by 2035.

In this way, AI-led cost savings can become revenue engines that businesses can use to continually redefine and push the boundaries in their sector.

3. Use a mix of AI models, including smaller specialist models

If every organization has access to the same large foundational models, the main way that any enterprise can differentiate itself is by combining these different models and tailoring them with unique enterprise data.

Which is why, in 2030, the most successful enterprises will likely be finding ways to use multiple models in an optimized way. They'll be relying on the versatility and power of the large general-purpose foundational models as well as the focus and speed of smaller purpose-built models, which can take on specialist tasks and handle real-time use cases.

In fact, by 2030, 82% of executives expect their AI capabilities to be multi-model, with 72% expecting smaller models to become more widely used than larger models in their organizations.

According to the research, organizations that use AI across multiple workflows and focus on smaller AI models or a combination of custom and foundation models expect:

• 24% greater productivity gains
• 55% higher operating profit margin improvements
• twice as much reduction on project delivery times (when compared to those that rely mainly on large pre-trained models).

4. Develop human 'AI orchestrators' to manage AI agents across multiple areas

While successful companies will find ways to boost the unique strengths of both humans and machines, IBM's research suggests that AI in the future is likely to outperform workers in areas that today require a "human touch".

As expected, AI-first organizations are 79% more likely to say knowledge work, like the creation of reports, proposals, and code, will mostly be done by AI. And they're actively working to make this happen.

But AI can also learn how to perform specialist tasks that humans can take years to master. So, using humans to do some specialist jobs that they do today will be an unnecessary cost.

Instead, organizations will need humans for roles that need critical thinking, innovation and problem solving.

And rather than individual employees using AI to augment their own jobs, there will be a demand for "AI orchestrators" who can manage AI agents across multiple areas.

There is going to be a shift in focus from technical experts to all-around business strategists who can understand and maximize the value of AI outputs. As human-led, AI agent teams become the new competitive business unit, HR and IT fuctions are expected to converge.

5. Make the most of Quantum Computing, alongside AI

For a long time, experts have been saying that quantum computing will solve problems that classical computers can't. Quantum is going to push past the computational limits of today's computers, including the limits of AI.

In the future, successful enterprises will use AI and combine it with the potential of quantum computing. The problem is, as of now, most firms don't know for sure how they will apply quantum.

Hence, 59% of executives say quantum-enabled AI will transform their industry by 2030 - but only 27% are confident enough to say they expect to be using it in their own organizations.

This gap between the potential of what quantum will deliver and concrete knowledge about how organizations are going to use it creates an opportunity for those who act decisively now.

Put simply, even if they're not sure how they'll use it, they need to start preparing for quantum now. They need to be able to get the most out of it, alongside scaling AI. Their preparation can include developing quantum skills and expertise, experimenting with real-world quantum hardware, and building infrastructure to host quantum applications when they arrive.

This article was originally published on the IBM Community.