Data from Deloitte’s report, AI ROI: The paradox of rising investment and elusive returns, show that 85% of organisations increased their AI investment in the past 12 months and a further 91% plan to increase investment again in the next 12 months.
However, the survey of 1,854 senior executives in 14 countries around Europe and the Middle East also reveals that returns on investment (RIO) are proving slow to achieve.
A typical AI use case takes two to four years to realise a ROI
Most respondents reported that a typical AI use case takes two to four years to realise a ROI, not the seven to 12 months that many traditional tech investments yield. Only 6 % of respondents reported payback in under a year; even among the most successful projects, just 13 % saw returns within 12 months.
Slow payback
The Deloitte survey identifies several reasons why ROI takes so long to realise:
- Many of the benefits of AI remain intangible or hard to monetise – for example, better vendor relationships, improved employee satisfaction or enhanced customer experience.
- Data quality, siloed platforms and infrastructure gaps hamper full deployment. Organisations often over-invest in AI tools before foundational data and architecture are ready.
- Technology evolves faster than measurement metrics – new tools change expectations mid-project.
- The human factor is real – cultural resistance, low adoption of new workflows, or insufficient change-management.
- AI initiatives are frequently embedded within broader transformation programmes, making it hard to isolate the incremental value derived purely from AI.
Organisations must treat many AI investments as part of strategic transformation
From a finance professional’s perspective, this means the traditional ROI models (eg capex now, clear savings later) may not apply cleanly. Rather than expecting a linear payback, organisations must take a longer view and treat many AI investments as part of strategic transformation.
Generative vs agentic
One of the more relevant insights in the report is the distinction between generative AI (models creating new content, code or text) and agentic AI (autonomous systems driving multi-step processes). According to the survey, among respondents already using generative AI, 15% reported they already achieved significant, measurable ROI, and 38% expect it within one year of investing.
By contrast, the returns are slower for agentic AI: only 10% currently see significant ROI, and most expect payback within one-to-five years. This suggests a tiered investment strategy: generative AI may offer near-term gains (productivity, content creation), while agentic AI is a longer-term bet (autonomy, process redesign).
Individual productivity
Despite all the strategic rhetoric, many AI initiatives today are still focused on productivity or workflow-support, rather than full business model transformation. Deloitte concludes that if investment is broad but still targeted at incremental productivity gains, the ROI horizon is likely longer, and the value may be harder to quantify.
Only a third of CFOs report an excellent understanding of AI
A separate survey has shown that three out of four CFOs in the US, UK and Australia now lead their organisation’s AI strategy. However, according to OneStream, while two thirds of CFOs believe their AI strategy is ahead of the curve, only a third (35%) report an excellent understanding of AI.
And while 56% report real productivity gains, 32% of CFOs express concerns about ROI uncertainty.
Implications for CFOs and the board
- Capital allocation: Boards and CFOs need to calibrate AI budgets with realistic payback horizons — two to four years (or more) for typical return, not months.
- Risk-assessment: Given the intangible elements and measurement complexity, many AI projects carry higher execution risk and delayed value realisation.
- Valuation modelling: For listed companies or private equity investors, the ROI leaders’ practices suggest that investment in AI should be treated as strategic transformation capital rather than short-term cost savings.
- Metrics and KPIs: Finance teams should complement traditional ROI metrics (cost savings, incremental revenue) with broader value indicators – eg improved customer retention, increased innovation throughput, workflow agility. As Deloitte notes, 65 % of organisations now say AI is part of corporate strategy.
- Human/infrastructure investment: The report strongly emphasises that technology alone does not deliver value – investment in workforce, governance, data foundations and process redesign is critical. For finance leaders this means budgeting not just for AI tools, but for change-programmes, training, data platforms and organisational redesign.
- Monitoring and benchmarking: Finance functions will benefit from tracking AI ROI using a multi-dimensional index (as in Deloitte’s Performance Index) rather than a single ‘savings’ figure. It may also mean segmenting generative vs agentic AI investments and modelling different payback horizons accordingly.
