'AI is no longer a future concept, but an operational reality': New KPMG report claims firms are "racing" to deploy AI - but need to ensure they have the right security protections

• Nearly all companies plan to deploy or scale AI in finance over the next year and a half
• Many are already exceeding ROI expectations, but use cases need to be defined
• Data security and privacy is still a big concern, cited by three in five

New research from KPMG has claimed while AI is now at the full-scale deployment stage, companies are now being urged to focus on security over AI implementation to ensure the best success going forward.

The findings come as nearly all (93%) US companies expect to deploy or scale AI in finance over the next 18 months, with one in every two planning more complex multi-agent AI systems.

As for returns, KPMG found it's been a case of 'so far, so good'. Around three-quarters (74%) have either met or exceeded their ROI expectations, and those experiencing slower-than-anticipated returns aren't actually having problems with the capabilities of AI itself.

Latest Videos From

AI strategies need to move into a next stage

The report explained that slow organizational adoption and change management are the biggest challenges, rather than the technology itself. A lack of clear, role-specific AI use cases (64%) and a lack of hands-on practice environments (61%) were cited as two key hurdles by the more than 1,000 senior finance leaders surveyed.

Looking forward, KPMG calls for more practical training opportunities and targeted upskilling tailored to roles and AI use cases. And as for those use cases, predictive decision-making insights (45%) is a major one for finance leaders as well as vibe coding, which is being piloted or used by 47% and evaluated by a further 26%.

"Leaders are harnessing sophisticated AI to create a finance function that is predictive instead of reactive," US and Global Audit Chief Digital Officer Thomas Mackenzie explained, discussing a "human-led, agent-operated" future.

That's not to say that AI doesn't come without its challenges, and many are still worried about data security and privacy (60%), AI model performance and reliability (56%) and regulatory and compliance issues (53%).

The solution for continued AI success and ROI is to raise the status of security in AI, in turn boosting confidence and reducing the risk of threats.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.