Aptean Research Finds General-Purpose AI Falls Short of Expectations
A year after an MIT study reported that 95% of generative AI pilots were failing to deliver measurable value, new international research from Aptean points to a core reason: general-purpose AI lacks the relevance and accuracy that businesses actually demand. The study also surfaces the rise of "Shadow AI," a growing call for governance, and enduring optimism about AI's long-term payoff.
The findings come from a survey of 1,535 decision-makers at companies with $10 million or more in annual revenue, conducted by independent research firm Vanson Bourne across six markets and five industries. Participants ranged from the C-suite to management, all with direct influence over their organization's AI, ERP, and supply chain strategy.
"Seventy-seven percent of leaders told us general-purpose AI simply can't handle the complexity of their operations. That's not a knock on the technology; it's a mismatch of design."
— TVN Reddy, CEO, ApteanThe Case for Industry-Specific AI
Over the past year, companies using vertical, industry-specific AI outperformed those relying solely on general-purpose tools on seven of eight operational KPIs measured. Respondents made their preference plain: 88% said purpose-built, industry-specific AI is critical or very important to their business. The top reasons cited were easier integration with existing systems, more relevant and accurate outputs, and better strategic guidance from vendors who understand their sector.
"A generic model doesn't know your business, your compliance rules, or what success looks like in your industry. Purpose-built AI does."
— TVN Reddy, CEO, ApteanIntegration remains the sticking point. 82% of respondents said connecting AI to core systems is harder than the AI technology itself, and most lack that capability in-house. As a result, 92% agreed their organization would benefit from outside expertise, and integration ranked as the single biggest factor in choosing a vendor.
A Call for Governance and Guardrails
When early pilots stalled, many workers simply routed around approved tools. The research found 40% of employees using unsanctioned "Shadow AI" at work, driven by a gap between how fast individuals adopt AI and how ready their organizations are, a shortage of sanctioned options, and personal priorities winning out over governance. Even so, 96% said a formal AI governance framework with clear accountability and ethical standards is necessary, and 75% called the absence of one a major barrier to success.
Optimism Endures
Despite the early failures and governance gaps, sentiment stays overwhelmingly positive. 83% said opportunity is their primary motivator, and 82% agreed AI offers clear value that hasn't been fully realized yet. Rather than reading today's failure rates as a reason to pause, respondents treat them as a reason to keep investing and pull ahead of competitors.
Vertical AI wins. Companies using industry-specific AI beat general-purpose-only users on seven of eight operational KPIs.
A design mismatch, not a tech flaw. 77% say general-purpose AI can't handle operational complexity, and 88% call purpose-built AI critical to their business.
Integration is the real barrier. 82% find connecting AI to core systems harder than the AI itself, and 92% want outside expertise to unlock its value.
Shadow AI is spreading. 40% of employees use unsanctioned tools, yet 96% agree a formal governance framework is necessary.
Optimism holds firm. 83% cite opportunity as their main motivator and 82% see value that AI has yet to fully deliver.