CSIRO releases guide for businesses to plan and invest in the right AI projects

Organisations are under pressure to do “something with AI”, with global investment in the technology projected to reach $980 billion by 2028, said Australia’s national science agency, CSIRO.

To help businesses make better decisions around AI projects, the CSIRO has brought out a guide to help organisations plan and invest in the right AI projects.

The CSIRO’s Evaluating and Prioritising Artificial Intelligence Projects: A Guide for Better Decision Making and Investment Outcomes, was a practical resource to help decision-makers identify AI projects that are more likely to succeed and deliver value.

The guide includes a practical checklist that blends financial analysis with ethical, strategic, and systems-level thinking, said Dr Stefan Hajkowicz CSIRO’s Data61 chief research consultant and lead author of the guide.

“Many organisations are investing in the wrong projects, with up to 80 per cent of AI projects failing,” he said. “This guide helps organisations evaluate options with a clearer view of the risks, strategic alignment, data requirements, and implementation challenges.”

It prompts leaders to weigh both financial and non-financial factors. This includes ethics, integration, and long-term value, with the assistance of globally recognised tools such as return on investment, net present value, and real options analysis, alongside qualitative insights.

The guide also encourages a portfolio mindset, treating AI projects as part of a broader strategy rather than isolated experiments, said CSIRO’s Digital, National Facilities and Collections Executive Director Professor Elanor Huntington.

“With the right evaluation frameworks in place, we can reduce the number of failed projects and direct investment where it will have the most impact,” she said.

The guide draws on real-world examples of AI projects that fell short of expectations.

One involved a predictive maintenance system that failed because it hadn’t been tested on the specific vehicles it was meant to monitor.

In another, a custom-built AI tool was soon outpaced by faster, cheaper commercial AI alternatives.

The guide helps public and private sector organisations at every stage of AI maturity plan and deliver smarter AI investments.