In what is claimed to be the largest ongoing global study into generative artificial intelligence (AI), a Lucidworks’ survey of businesses pursuing generative AI initiatives found that only one in four planned projects fully were implemented in the past 12 months.
The results show that despite hype into generative AI, slow deployment and low success rates are commonplace.
According to the survey, which included respondents from North America, EMEA and the APC regions, found that the slow deployment resulted in 42% of companies yet to realize return on investment or significant benefits from generative AI projects.
Among these companies, tech and retail sectors stand out with higher deployment and realized gains, however, most are slow to move beyond a pilot phase, Lucidworks found.
"The initial wave of enthusiasm for generative AI is being met with a more strategic approach,” said Mike Sinoway, CEO, Lucidworks. “Businesses are recognizing the potential of this technology, but they're also cautious about the risks and costs. This is reflected in the flattened spending, which suggests a shift toward more thoughtful planning. This planning ensures AI adoption delivers real value, balancing the need to stay competitive with managing costs and potential risks."
Generative AI cooling?
Lucidworks’ survey found that global generative AI spending plans are down sharply with 63% planning to increase spending of those that responded. In 2023, 93% of vendors said they were increasing AI spending. In the U.S., organizations remain above average with 69% planning to increase AI spending.
Along with managing risks and costs of AI projects, security remains a top concern for companies. Additionally, concerns around response accuracy have risen five times among companies with most responding to a better need for careful large language model selection to balance cost and ensure accurate results.
First to deploy
So what projects are successful?
The Lucidworks survey found that qualitative applications like generating FAQs are successful but more complex quantitative applications like fraud detection in financial services or predictive maintenance in manufacturing are lagging.
Analyzing unstructured data and generating actionable insights holds the best potential for generative AI’s future value, the survey found.
Other findings in the report include:
- 36% of vendors plan to keep spending flat compared to only 6% last year.
- Nearly half of Chinese vendors plan to increase AI spending this year, a drop from 100% in 2023.
- 70% of financial services companies plan to increase spending in the next 12 months,
- Nearly eight in 10 companies use commercial large language models (LLMs) and 21% have opted for open source only.
- Retailers are most concerned about the cost of AI implementations.
The full survey can be found in Lucidworks’ Generative AI Global Benchmark Study.