Businesses should use artificial intelligence not only to cut costs but also to create new sources of revenue, according to entrepreneur Bill Gross, who has called for companies to set growth targets alongside efficiency goals.
Gross, who says he has helped create more than 175 companies over five decades, argues that major technology shifts are often initially used to make existing processes cheaper. The organisations that gain most, he says, use them to develop entirely new products and services.
He points to electricity, which did more than reduce factories’ power bills by enabling new production methods, and the internet, which created businesses such as Amazon rather than simply making catalogue distribution less expensive.
The focus on efficiency is reflected in corporate adoption of AI. An EY-Parthenon survey published in April 2026 found that 63 per cent of growth leaders at large US companies said their organisations were using the technology primarily for efficiency and productivity, while only 7 per cent cited diversifying revenue streams.
Research by McKinsey in 2026, Gross said, found that the strongest AI performers pursued growth or innovation alongside efficiency.
AI opens up new business opportunities
The pace at which AI systems can handle increasingly complex work is also expanding. In a 2025 study of software and reasoning tasks, METR found that the length of tasks AI agents could complete with 50 per cent reliability had been doubling roughly every seven months.
That progress is making it cheaper to explore ideas at an early stage, particularly in software. Teams can use AI to research markets, write and test code, and develop an initial prototype before committing significant staff and funding.
Gross said the central challenge remained selecting worthwhile opportunities, establishing whether customers wanted them and turning them into sustainable businesses. Established companies, however, could draw on industry knowledge and capabilities that new ventures often lack.
Those assets include information about customers’ complaints, purchasing decisions and workarounds, as well as capabilities originally developed for other purposes. Companies may also hold patents, trade secrets or research projects that were shelved because they were too expensive to staff.
At Idealab, Gross said, 31 methods for finding business ideas had been developed over three decades and organised into seven broad approaches. The methods had been used with AI to create a database containing more than 1.2 million potential business ideas.
From ideas to paid trials
Four of the approaches begin with demand: identifying a problem, making an existing service substantially better, bringing a premium service to a wider group, or examining recurring questions, complaints and workarounds for signs of unmet needs.
Three others begin with what has become possible. They involve applying a new capability to an old problem, looking for opportunities created by falling costs, platform changes or regulation, and reconsidering projects that may previously have been abandoned because the timing was wrong.
Gross gave the example of a logistics company able to predict delivery disruptions. Its expertise might support a planning service for smaller shippers, although customers would still need to establish that the problem was important enough for them to pay for a solution.
He advised companies to test ideas against their expertise, customer knowledge, operational capabilities and intellectual property. AI can help identify and challenge assumptions, but demand should be tested directly through customer interviews, prototypes or paid pilots.
The strongest opportunities should have an accountable owner and clear milestones determining whether a company invests further, changes direction or stops. Gross said the next board presentation should include a revenue target alongside any target for savings, with the broader aim of improving efficiency while creating new growth.
