Which Technology Investments Are Delivering ROI?

Businesses are investing heavily in artificial intelligence, cloud platforms and data infrastructure. But spending on technology and getting value from it are two different things.

Some organisations report returns from their investments. Others are still working to move AI projects into production or demonstrate improvements in growth and efficiency.

Research from Deloitte and ISG highlights this gap. It also points to a more useful question for business leaders: Is the technology delivering a measurable business outcome, or is it simply being adopted?

Which technology investments are reporting returns?

Deloitte's 2025 Tech Value Survey, which included 548 business and technology decision-makers across five industries, found that respondents investing in several technology areas reported returns.

The share reporting ROI included:

• AI and generative AI: 84%
• Data management and architecture: 83%
• Data insights: 80%
• Cloud platforms and cloud-native applications: 79%

Source: Deloitte, AI and tech investment ROI

These figures suggest that organisations are seeing value across more than one part of their technology estate. However, they show the proportion of respondents reporting ROI, not the size of their returns or independently verified financial gains. Deloitte fielded the survey in May and June 2025 among firms with at least US$500 million in annual revenue. The percentages are of those investing in each area.

A positive return can also mean different things. One business may measure reduced operating costs, while another looks at faster processes, better decisions or revenue growth. The figures should not be treated as a direct ranking of which technology delivers the greatest financial benefit.

Why are some AI investments still struggling to demonstrate value?

A different picture emerges when the focus shifts from reported ROI to implementation and business outcomes.

ISG's State of Enterprise AI Adoption Report 2025, which examined 1,200 AI use cases, found that 31% had reached full production. Its findings also showed that one in four initiatives achieved expected ROI on growth, while 50% achieved expected efficiency gains. ISG's September 2025 release describes the study as a survey of 400 senior AI decision-makers.

Source: ISG, State of Enterprise AI Adoption Report 2025

The distinction matters. An AI project may be running successfully without increasing revenue. Another may improve efficiency without yet delivering the growth the business expected.

Moving from an experiment to a working business process is only one step. The next is establishing whether that process improves the outcome it was meant to change.

This is why the number of AI pilots, tools purchased or employees using a new system cannot, on its own, demonstrate business value.

Is a positive return always a quick return?

Not necessarily. The time required to demonstrate value is another important part of the comparison.

In separate 2025 research, Deloitte surveyed 1,854 executives across Europe and the Middle East. Most respondents reported achieving satisfactory ROI on a typical AI use case within two to four years. Only 6% reported payback in under a year. Deloitte contrasted this with the usual seven to 12 months.

Source: Deloitte, AI ROI: The paradox of rising investment and elusive returns

This does not mean businesses should accept long payback periods without scrutiny. It means expectations need to reflect the type of investment, the work required to implement it and the outcome being measured.

The two Deloitte studies also use different samples and measures, so their findings should be read separately rather than combined into a single performance ranking.

For business leaders, the practical question is whether progress is measurable and whether the expected value still justifies the cost and time involved.

What separates a valuable investment from an expensive experiment?

The difference often becomes clearer when a technology project is assessed against the problem it was meant to solve.

1. A defined business problem

Before choosing a technology, identify the outcome that needs to improve. Is the goal to reduce manual work, shorten response times, improve service quality or increase revenue? A clear starting point makes it easier to assess results.

2. A realistic implementation plan

Technology rarely works in isolation. Data quality, existing systems, security requirements, team readiness and process changes can all affect the time it takes to deliver value.

These requirements should be part of the investment decision, not treated as problems to solve after the tool has been purchased.

3. Evidence of adoption and impact

A system being available does not mean employees or customers are using it effectively. Businesses need to understand whether it has become part of everyday work and whether that use is improving the intended outcome.

4. A clear view of the total cost

Licensing is only one part of the investment. Integration, implementation, training, maintenance and ongoing support can also affect the final return. Measuring these costs against the benefits gives leaders a more complete picture.

How should businesses measure technology ROI?

The right measures depend on the problem being addressed. A useful evaluation should connect technology activity to a business result.

Businesses should establish a baseline before implementation, agree on the measures of success and review progress at defined intervals. They should also account for the full cost of ownership and distinguish benefits that can be measured directly from those that are still assumptions. For customer-service AI, one adoption measure is containment rate: conversations handled without a human agent, read with resolution and satisfaction.

Not every benefit will appear immediately. But if a project repeatedly misses its milestones, lacks adoption or cannot show progress against its intended outcome, the business should reassess the approach rather than relying on the promise of future returns.

Measuring customer conversations? The free CPaaS KPI Handbook covers 50+ WhatsApp, SMS and RCS metrics, with formulas.

The real measure of technology investment

The research shows that businesses are reporting returns from AI, data and cloud investments, while other AI initiatives are still working towards production or expected business outcomes.

Neither finding tells the whole story on its own. The value of an investment depends on what it is designed to achieve, how well it is implemented and whether its results can be measured.

The goal is not to invest in the technology attracting the most attention. It is to invest in a business outcome, then gather the evidence to show whether the investment is delivering it.

Want to measure AI in customer conversations? Talk to the Helo.ai team about Helo Convo, a no-code AI agent.

Frequently asked questions

What is technology ROI?

Technology ROI compares the business benefit of a technology investment with its total cost. The benefit can be financial, such as savings or revenue, or operational, such as faster processes.

Why is AI ROI hard to measure?

Deloitte's 2025 European and Middle Eastern survey names five reasons. Benefits are often intangible and platforms and data are siloed. Metrics lag the technology, adoption depends on people, and AI is rolled out alongside wider change.

How long before an AI investment pays back?

In Deloitte's 2025 survey of European and Middle Eastern executives, most reported satisfactory ROI on a typical AI use case within two to four years. Only 6% saw payback in under a year.