1 in 3 firms investing in AI with little oversight

IDC surveyed 800 technology leaders across APAC, Europe and the US.

More than one in three organisations in Singapore are investing aggressively in artificial intelligence (AI) with little evaluation.

The report, based on a survey of 800 technology leaders across Asia-Pacific (APAC), Europe and the US, found that enterprise AI spending is rising quickly, but many organisations are still not measuring returns closely, according to a new IDC InfoBrief commissioned by Expereo.

Globally, around 70% of organisations are investing in AI. One in five, or 20%, said they are investing aggressively with little evaluation because they fear falling behind competitors.

The pressure is stronger in APAC, where 37% of organisations said they are investing aggressively with little evaluation. This is nearly double the global average and higher than the US at 10% and Europe at 13%.

In Singapore, more than one in three organisations admitted they are taking the same approach.

The report found that AI is now one of the top technology investment areas. Globally, 51% of organisations plan to prioritise AI or machine learning investment over the next 12 months. In APAC, the figure rises to 61%.

However, returns have not kept pace for many companies. Only 19% of organisations globally said their AI projects had exceeded expectations, whilst just 5% said they had significantly exceeded expectations.

APAC reported stronger results than the global average, with 40% saying their AI projects had exceeded or significantly exceeded expectations. 

Even so, most organisations in the region said results had not gone beyond expectations.

The main reasons for weaker AI performance globally were poor-quality or inadequate training data, cited by 51% of organisations; higher-than-expected costs or failure to achieve return on investment, cited by 47%; and AI not performing as expected, cited by 46%.

In APAC, 49% cited poor-quality training data, 54% cited cost overruns or return on investment not being achieved, and 46% said AI had not performed as expected.

The report also pointed to network and infrastructure readiness as a key issue. 
Globally, 26% of organisations whose AI projects fell short of expectations said inadequate network or connectivity performance was one reason.

In APAC, only 9% of organisations said their network infrastructure was fully ready to support new AI, cloud and digital initiatives. Another 37% said their infrastructure would need upgrading or replacing soon.

Singapore stood out as one of the markets where network needs are most pressing. The report found that 58% of organisations in Singapore said they need more flexible and scalable networks, above the APAC average. Thailand was higher at 74%.

Expereo said companies need stronger networks if they want AI projects to deliver results at scale.

Ben Elms, CEO of Expereo, said the data showed a gap between AI ambition and AI outcomes.

“More often than not, that gap comes down to the network underneath. AI only delivers on its promise when the infrastructure carrying it is built to support it,” he said.

He said resilient, scalable and cloud-optimised networks are now important to whether AI projects can deliver return on investment.

APAC is also ahead in AI adoption. The report found that 35% of organisations in the region are using AI extensively across the business, compared with a global average of 25%.

Eric Wong, President for APAC at Expereo, said companies in the region are moving quickly on AI, but need to look beyond applications and models.

“The underlying network, cloud connectivity, and operational readiness matter just as much,” he said.