AI ROI: A practical framework for assessing the return on AI investments

Boards and executives are under growing pressure to demonstrate returns from their AI investments – particularly as the costs of some of those investments are climbing. 

The reality is that AI ROI is harder to assess than many organisations expect. 

Benefits are often diffuse. Adoption varies. Integration costs emerge late. And promising pilots don't always translate into sustained organisational value. 

Today, the UTS Human Technology Institute is releasing a new practical framework to help boards and executives assess AI ROI more credibly and consistently. 

The framework focuses on four practical steps: 

1. Assess ROI for a specific use case, not a general AI capability 
2. Align expectations with the implementation stage (ie, Is this a pilot? Are you integrating the use case with existing systems? Are you scaling it more broadly?) 
3. Capture both direct and indirect benefits (including improvements in data foundations, AI literacy and risk management capability) and test your assumptions as you go, and  
4. Ensure that you are modelling the full costs of adoption, change and governance. 
    

And then, bring it all together – and be prepared to show discipline if the AI use case is not delivering value within expected timeframes. This may involve revising assumptions, refining the use case, or it may involve dropping it altogether.  

Finally, the organisations achieving strongest returns are those that focus on their decision quality, and are prepared to leverage the benefits and costs across their portfolio of AI investments. 

HTI thank our authors, Nicholas Davis, Gaby Carney, Andrea Jachs and Brad Daffy. 

Download the framework (PDF, 1MB)