As organizations race toward the hype of AI, many business leaders and founders have realized that the return on investment (ROI) of their AI solutions is falling short.

MIT’s GenAI report claims that only 5% of AI pilots achieve rapid revenue acceleration, and 95% deliver little to no impact.

Similarly, a report released by the IBM Institute for Business Value claims that the average return on AI initiatives within the company is only 5.9%. However, these AI programs cost 10% of their capital investment.

The result? Leaders are now requesting ROI projections in advance to assess the returns on their AI initiatives. But is this approach truly effective?

This article aims to guide business leaders and founders in unlearning the AI hype and embracing a more intentional approach to AI development.

Breaking Down: A CEO’s POV on Demanding ROI from AI

Jensen Huang, the CEO of Nvidia, recently addressed the ROI of AI at the Cisco AI Summit 2026 in simple words: "Let a thousand flowers bloom.”

Huang advises business leaders to understand that AI is still an emerging technology, and it’s still in its experimental phase. Hence, putting rigid ROI numbers on spreadsheets can actually limit the very innovation you’re trying to fund.

Instead of demanding ROI on AI, he suggests founders ask:

• “What’s the essence of our company?”

• “What’s the most impactful work that we do at our company?”

He adds by saying, "Innovation is not in control. It’s an illusion."

To ensure innovation delivers tangible results, it is always necessary to let go of boundaries.

Most business owners today are seeking explicit, specific, demonstrable ROI. However, the reality is that calculating the value of something worth doing in the beginning is challenging.

He suggests, when your team is eager to experiment with AI and has a plan, say "Yes" and then ask, "Why?"

Huang connects his statement with an anecdote: “I want the same thing for my company that I want for my kids.”

He urged leaders to take a relaxed approach to ROI while focusing more on their basic understanding of AI.

3 Practical Tips to Define ROI of AI

To make sure your next AI project is driven by clear intent rather than a rush to implement, use these three practical tips:

Measure “Value Reinvestment,” Not Just “Time Saved”

Most leaders consider AI ROI is equal to time saved.

For example: “How many hours did automation free up?”

Well, that’s just the start—not ROI.

The real question leaders should ask is, “Where did the organization reinvest that time into?”

If AI saves a team 10 hours a week, but those hours disappear into meetings or low-impact work, the ROI of AI remains surface level.

However, when saved time is intentionally reinvested into higher-value activities—strategic planning, customer engagement, innovation, or decision-making—AI begins to generate measurable business value.

Solve “High-Friction” Problems First

AI delivers the highest ROI when it removes friction from critical workflows—not when it automates already-efficient tasks.

High-friction problems are areas where teams lose time, accuracy, or momentum every day. For example:

• Long approval cycles

• Repetitive decision-making

• Customer-facing delays or errors

These friction points create hidden costs across productivity, morale, and customer experience. When AI reduces friction in these areas, leaders see compounding returns

Factor in the “Human-in-the-Loop” Costs

One of the biggest mistakes leaders make when calculating AI ROI is assuming full automation.

In reality, most enterprise AI systems operate with a human-in-the-loop—people who train models, review outputs, correct errors, and make final decisions.

These human contributions are not failures of AI; they are essential to form trust, accuracy, and governance.