Nearly half of business AI projects abandoned midway, study finds  

Nearly half of new business artificial intelligence projects are abandoned midway, a study has found.

A recent study conducted by the international law firm DLA Piper, surveying 600 key executives and decision-makers from global corporations, sheds light on the significant challenges businesses face in integrating AI technologies. Despite the promising potential of AI to revolutionize various sectors, the journey toward successful implementation is fraught with obstacles. This article delves into these challenges and provides expert commentary on navigating the complex landscape of AI integration.

The study revealed that although over 40% of organizations fear that their core business models will become obsolete unless they adopt AI technologies, nearly half (48%) of the companies that embarked on AI projects have been forced to pause or rollback them. The primary reasons for these setbacks include concerns over data privacy (48%), issues related to data ownership and inadequate regulatory frameworks (37%), customer apprehensions (35%), the emergence of new technologies (33%), and employee concerns (29%).

DLA Piper has been active in Israel since 2009 and is led by Adv. Jeremy Lustman and Adv. Naomi Maryles. According to Jeanne Dauzier, Partner, Global Co-Chair, AI Practice Group at DLA Piper, “Two clear messages ring out from this research. First, there is an urgency to adopt AI – this is not an area where businesses feel able to wait and see. Second, the need to ensure opportunities in productivity and efficiency do not come at an ethical cost to the business and community. There is a real imperative for values-driven value creation with AI.”

According to Arik Faingold, President & Chairman of Commit, a technology company that advises large enterprises on the implementation of AI-powered tools and provides them with AI-based technological solutions, "the findings of the survey are unsurprising, if not conservative. Based on our industry knowledge, the number of companies that have begun exploring the implementation of AI tools and ultimately decided to hold implementation for the time being are likely higher than 50%."

However, in contrast to the survey's findings, Faingold believes that the reasons why organizations abandon AI projects are fundamentally different.

"One of the primary reasons is the gap between the capabilities of AI-powered tools in their current state of development and the processes that these organizations seek to streamline, some of which cannot yet be adequately addressed by the tools available on the market,” he said. “This gap is relatively easy to identify even in the early stages of the process.” 

“Another reason why transitions to AI-powered tools are halted is the difficulty of integrating multiple disciplines, such as data, cybersecurity, and user interface,” he added. “This is something that we at Commit have been doing regularly for many years in other contexts as well, but those who are not experienced in this area may encounter difficulties."

Faingold explained that customer service and support are the areas where AI is currently making the most satisfactory improvements and efficiencies. 

"Many organizations are already implementing chatbots and providing efficient and rapid customer responses using AI,” he stated. “However, the situation is not yet encouraging when it comes to software development tools, and as a result, their implementation is also lacking. This is a gap that I anticipate will narrow significantly in the coming months and years."

Orna Kleinmann, Managing Director of SAP's R&D Center in Israel, emphasized the paramount importance of responsible, relevant and reliable data in business AI, where the stakes are significantly elevated. 

“The consequences of bias, errors, or 'hallucinations' within a business AI model could be catastrophic for a company, resulting in revenue loss, damage to reputation or even affect society itself," Kleinmann warned. "For businesses to trust generative AI, they need to be certain their data is handled responsibly and safely and that it is the relevant data that is taken into consideration.  Generative AI tools must respect and observe data privacy, data ownership, and data access restrictions by design, and operate only in areas where explicit consent has been given.” 

The three “R”s—relevance, reliability, and responsibility

Kleinmann underscored that the three “R”s—relevance, reliability, and responsibility—are the cornerstones of trustworthy AI for the business world."

Based on the study's findings and expert insights, Kleinmann pointed to several strategies that emerge for businesses aiming to successfully integrate AI. 

“Developing a clear AI strategy that outlines the vision, objectives, and specific use cases with clear KPIs is crucial,” she said. “This strategy should be integrated into the broader business plan to ensure alignment and coherence. Investing in data governance is equally important; establishing robust data governance frameworks addresses privacy and ownership concerns, including implementing clear policies for data collection, storage, and usage, and ensuring compliance with relevant regulations.”

Kleinmann and Faingold emphasized the importance of fostering collaboration between different departments within an organization. According to them, Cross-functional teams can provide diverse perspectives and expertise, leading to more innovative and effective solutions.

“In addition to strategic alignment and data governance, selecting the right AI vendor is critical. Businesses must navigate a complex landscape of technology providers, ensuring they choose a partner capable of meeting their specific needs,” Faingold added.

According to Faingold, "Google, AWS, and Microsoft platforms are well-equipped to handle data privacy issues.” 

“Businesses should recognize that privacy concerns in the context of AI should not be so intimidating and should not prevent them from exploring this technology.,” he added. “Cloud providers are skilled in managing these concerns. This is also true of regulation, which, while evolving and changing, still leaves enough room for businesses to operate without unnecessary risk."

In navigating the intricate landscape of AI integration, businesses face a multitude of challenges and decisions. From aligning AI initiatives with strategic objectives to fostering a culture of innovation, the journey toward successful implementation requires careful planning and collaboration. 

As highlighted by Kleinmann and Faingold, the stakes are particularly high in the realm of business AI, where the consequences of missteps could be harmful for revenue, reputation, and even society at large. As businesses continue to grapple with these complexities, one thing remains clear: the path to AI adoption must be guided by transparency, responsibility, and a commitment to ethical practices.