AI investment is now part of day-to-day operations for many organisations. What hasn’t kept pace is confidence in the return. Leaders face growing pressure to show clear outcomes, yet many struggle to quantify value beyond early pilots. Even when teams see progress, the link between AI activity and measurable business impact is often unclear.
The challenge comes from applying traditional ROI models to systems that behave differently. AI is not predictable in the same way as conventional software. Outputs can vary, performance can shift, and user perception does not always match measurable results. This makes it harder to define benefits upfront and track them consistently over time.
A more practical approach needs to reflect how AI works, how value evolves across implementation stages, and how risk and trust influence outcomes.
KPMG has collaborated with Gaby Carney (Senior Fellow, Strategic AI) and Nicholas Davis (Industry Professor, Emerging Technology) from the Human Technology Institute at UTS to develop a structured framework that brings discipline to AI investment decisions.
What is the best framework for measuring AI ROI?
It starts by defining value clearly, then tests whether that value is realised as AI moves from pilot to scale.
The approach is not a single calculation. It is a set of steps that guide how organisations scope, measure and govern AI over time.
The most effective frameworks focus on these core areas:
- defining a specific use case
- aligning expectations to implementation stage
- capturing both direct and indirect benefits
- testing realisation assumptions
- modelling full costs.
Together, these create a common language between delivery teams and senior leaders, helping organisations make better decisions about where to invest, where to adapt and where to stop.
Six practical principles for credible AI ROI measurement
Focus on a specific business problem with a clear, measurable outcome. Avoid broad AI rollouts that dilute ownership and make ROI difficult to prove. Define success with a practical metric and baseline.
ROI changes across experimentation, integration and scaling. Early stages validate feasibility and risk, while scale drives financial outcomes. Use stage gates to guide investment decisions.
Look beyond direct gains like time savings or cost reduction. Include benefits such as stronger data, improved AI literacy, better risk capability, and improved employee and customer experience.
Check whether expected benefits will hold in practice. Measure adoption, assess how individual improvements translate to team value, and ensure outputs are reliable and accurate.
Include integration, training, workflow changes and governance costs. Indirect and structural costs often have the biggest impact on realised ROI.
Sustained value depends on trust. Strong governance, controls and monitoring increase confidence and support safe scaling.
How do you measure AI ROI at scale across the organisation?
Scaling AI requires a shift from individual use cases to a portfolio view.
Each successful deployment should make the next one easier through shared infrastructure, improved data and stronger workforce capability. This is where indirect benefits begin to compound and where ROI can accelerate.
Measurement at scale also needs consistency. Organisations should track both financial and operational indicators. Financial metrics focus on the business outcome defined for each use case. Operational metrics track adoption, workflow fit, and performance over time. Together, these provide a more complete view of whether value is being realised and sustained.
A common failure point at scale is integration. Even when the technology works, value can drop if workflows are not redesigned, adoption is uneven, or risk controls are introduced too late. This highlights the need to treat implementation discipline as part of ROI, not separate from it.
A minimum viable approach to AI ROI for busy leaders
Not every organisation can apply a full AI ROI framework immediately. A minimum viable approach can still deliver strong results if applied consistently.
How KPMG can help
KPMG helps organisations define, prioritise and measure high-value AI use cases, with clear outcomes and stage gates to guide investment decisions.
We support implementation by embedding AI into workflows, strengthening adoption, and improving change management so solutions deliver real business value beyond pilots.
We also help build trust through governance, monitoring and assurance, while strengthening data, risk and capability foundations that enable AI to scale with confidence.
Get in touch
- Item 1
- Item 3
Frequently asked questions
Define a specific use case and measurable outcome, estimate benefits, adjust for adoption and real-world factors, and include all direct and indirect costs. Review regularly as the use case evolves
AI systems are less predictable, benefits can be indirect, and outcomes change across implementation stages. This makes traditional ROI models less reliable.
Use a combination of business outcome metrics, adoption rates, workflow integration indicators, and performance measures such as accuracy and reliability.
Indirect benefits include improved data quality, stronger AI literacy, better risk management capability, and improved employee and customer experience.
By defining clear use cases, testing real-world assumptions, tracking adoption and performance, and embedding governance and trust practices into every stage of implementation.