Enterprises are investing heavily in artificial intelligence, but most have yet to realize the returns they expected. New research from HFS Research, conducted in collaboration with Rackspace, finds that only 33% of enterprises have met or exceeded the business case for their most significant AI program, while 96% report significant governance gaps. The study of 502 senior AI decision makers points to a central obstacle: fragmented accountability for how AI performs, what it costs, and how it is controlled once deployed.
The report, “Run Your AI as One, or Keep Paying for Parts That Never Add Up,” finds that the challenge intensifies as organizations expand AI across multiple business functions. Companies have invested in models, platforms, infrastructure, and services, but too often no one is responsible for running the pieces as one system.
Key findings from the study include:
33% have met or exceeded their AI business case. The share falls to 28% among enterprises deploying AI across several functions.
96% report significant governance gaps, while 78% view governance failures as a bigger threat to scaling AI than technology failures.
44% have no single, consistent approach to running AI in production.
61% prefer one accountable partner across AI models, platforms, and infrastructure, rising to 74% among production-stage firms whose programs have not delivered.
73% would pay a premium over raw infrastructure for a fully managed AI service with governance accountability.
“Enterprise AI has an accountability problem that nobody in the industry wants to admit,” said Phil Fersht, Founder and CEO of HFS Research. “We’ve built an enormous ecosystem of companies selling models, platforms, infrastructure and consulting, yet when the promised results don’t materialize, everyone points at someone else and the enterprise is left trying to make the pieces work together. Only a third of enterprises have met the business case for their most significant AI program, which should force us to confront how addicted we’ve become to selling AI rather than taking responsibility for making it work. We have no shortage of technology or talent, but we are missing people and partners willing to take ownership of how the whole system operates once it goes live and whether it actually delivers what the business needs. Until that accountability exists, enterprises will keep paying for parts that never add up to the outcomes they were promised.”
AI business-case achievement varies by deployment stage
The study finds that the share of enterprises meeting or exceeding their AI business case varies by deployment stage. Among enterprises deploying AI in one function, 37% have met or exceeded their business case, compared with 28% among those deploying across several functions. Among organizations with AI embedded in standard operations, 56% have met or exceeded their business case. These figures describe different groups and do not establish that operating maturity causes stronger returns.
Share of enterprises that met or exceeded their AI business case
Most reported gains remain concentrated in efficiency: 58% cite workforce productivity improvements and 42% measurable cost reductions, compared with 28% reporting better customer experience and 21% measurable revenue growth.
HFS calls for an accountable AI operator
To close the gap, HFS calls for a Services-as-Software™ (SaS) Operator: an internal team, group of existing providers under one remit, or managed services partner accountable for the day-to-day performance, economics, and governance of AI once it is live. The operator coordinates the technology and partners already in place while the enterprise retains its business context, policy, decision rights, and legal accountability.
“Turning AI investment into measurable business value requires clear operational accountability,” said Gajen Kandiah, CEO, Rackspace. “As enterprises move from pilots into production, infrastructure readiness, security, governance and cost have to be managed together. Customers need an operator accountable for integrating, running and continuously improving the environment across infrastructure, applications, data and AI. They should retain control of their data and technology choices, with clear measures of performance and cost, and deployments paced to their business needs.”
The research also points to demand for a new operating model: 63% of respondents recognize the proposed operator role as meaningfully different from what they buy today, while 73% say they would pay a premium over raw infrastructure for a fully managed AI service with governance accountability. However, providers must demonstrate clear operational responsibilities and measurable outcomes to earn that premium.


