Artificial intelligence is vital to business but organisations that use it don’t trust AI and still rely on manual data processes.
Artificial intelligence capabilities lag far behind business ambitions, a failure of management that leaves money on the table according to a survey which shows that companies don’t trust AI enough to forego human decision-making.
While 87 per cent of organizations consider artificial intelligence (AI) vital to their business survival, 86 per cent struggle to fully trust AI to make all business decisions without human intervention. Ninety per cent of respondents said their organizations continue to rely on manual data processes.
Only 14 per cent of organizations consider their AI maturity ‘advanced’ meaning that they use general purpose AI to automatically make predictions and business decisions, according to the survey by data integration specialist Fivetran.
More than two in five respondents (41 per cent) conceded there was vast room for improvement in how their organization used AI. That number spiked to 64 per cent when looking at U.S. respondents only.
“This study highlights significant gaps in efficient data movement and access across organizations. A successful AI program depends on a solid data foundation, starting with a cloud data warehouse or lake as its base,” said George Fraser, CEO of Fivetran. “Analytic teams that utilize a modern data stack can more readily extend the value of their data and maximize their investments in AI and data science.”
Inefficient data processes curtail revenue gains
Organizations appear to be laying the foundation for more sophisticated AI projects and plan to invest 13 per cent of their global annual revenue into them within the next three to five years – compared to the eight percent being invested today. Almost all of the organizations surveyed already collect and use data from operational systems, but their ability to use this data for AI models is hampered by deep-running data challenges:
71 per cent struggle to access all the data needed to run AI programs, workloads and models
At least 73 per cent find each of the stages of extracting, loading and transforming the data, to translating it into practical advice for decision-makers a challenge
Such inefficient data processes force companies to rely on human-led decision-making 71 per cent of the time. Underperforming AI programs are also hitting organizations financially, with respondents estimating they are losing out on an average of five percent of global annual revenues because of models built using inaccurate or low-quality data.
The prevalence of low-quality, siloed and stale data means that data scientists, employed by all large organizations surveyed, dedicate less than a third of their time to building AI models, spending the rest of it on tasks outside of their job role.
As a result, 87 per cent agree that data scientists within their organization are not being utilized to their full potential. Yet, recruitment is cited (by 39 per cent) as the biggest barrier to AI adoption, highlighting the responsibility of organizations to urgently empower the talent they already have. The survey was conducted by Vanson Bourne among 550 senior IT and data science professionals across the U.S., U.K., Ireland, France and Germany.
