Software
US firms lose 2.4% of revenue on failed AI projects
According to a report released by Emergn on Wednesday, U.S. companies lose an average of 2.4% of annual revenue due to AI projects failing to deliver expected value. The technology and consulting firm surveyed 700 senior business leaders. Only 30% of organizations consider shutting down underperforming AI projects a normal practice, and nearly half say projects are typically halted only after significant time and money have been invested.
HP prepares to roll out OpenAI’s agentic platform
HP and OpenAI announced an agreement on Sunday to deploy OpenAI's enterprise AI platform Frontier across HP's global operations, making HP one of the first enterprises to adopt the platform at scale. The two parties plan to jointly develop AI agents for internal operations such as customer support, employee productivity, and software development. HP initiated an exploration phase when the platform was launched in February, assessing the alignment of technical capabilities with business priorities.
AI programming costs expected to exceed human developer salaries by 2028
According to the latest Gartner report, by 2028, the token costs of AI programming will exceed the average salary of ordinary developers. The shift from subscription models to pay-as-you-go pricing intensifies fluctuations in AI spending, making it difficult for enterprises to predict and control budgets. Experts advise CIOs to strengthen usage monitoring and cost governance.
CEO: Existing IT operating models are difficult to adapt to the AI era
Gartner's latest report indicates that more than 70% of CEOs believe their organization's IT operating model cannot adapt to the AI era, with traditional centralized frameworks struggling to cope with rapid AI disruption, geopolitical volatility, and the proliferation of AI within enterprises. Only 24% of CIOs say their existing models can effectively adapt to business changes, leaving CIOs facing an imbalance between responsibility and control. Gartner recommends shifting to an orchestration-centric 'technology operating model' with embedded decision rights and guardrails.
Technical debt and process shortcomings trap enterprises in AI 'pilot purgatory,' research reveals root causes
Latest research indicates that the world's 2000 largest listed companies are unable to unlock nearly $18 trillion in AI investment value due to internal operational shortcomings—including poor data quality, inefficient processes, outdated technology, and talent gaps. The report states that most enterprises have not yet addressed the fundamental issues constraining AI returns, with only 6% of surveyed organizations classified as 'mature debt resolvers.'
IBM and Google Cloud Reach Strategic AI Platform Partnership to Expand Enterprise AI Services Market
IBM and Google Cloud announced a deepened strategic partnership, integrating IBM Consulting Advantage platform with Google Gemini enterprise agent platform and forming a Google Cloud business unit staffed with thousands of certified consultants. This initiative aims to help clients modernize legacy systems, manage hybrid cloud infrastructure, and drive AI application adoption.
Salesforce bets on Agentforce growth while other business segments lag
Salesforce achieved revenue of $11.13 billion in the first quarter of fiscal year 2027, up 13% year-over-year, but analysts noted slowing growth in products such as Tableau and Agentforce Commerce, with the company relying on the Agentforce platform to sustain growth momentum.
Google Cloud and Workday Deepen Partnership, Launch HR and Finance AI Agents
Google Cloud and Workday announced an expanded partnership on Thursday, integrating Workday's Sana self-service agent into Gemini Enterprise, enabling employees to query leave balances, payroll, tax information, and expense policies in a single interface, and supporting managers in approving timesheets and initiating performance reviews. The two companies aim to automate HR and finance processes through AI agents, responding to transformations in the SaaS industry.
Uniform governance policies may cause enterprise agent projects to fail; Gartner recommends tiered control
Gartner's latest report shows that adopting a uniform governance strategy for enterprise AI agents will significantly increase project failure rates. The firm predicts that by 2027, 40% of enterprises will discontinue some agents due to technical teams failing to distinguish between agents' action capabilities and authorization scope. Analysts recommend a four-level proportional governance model—observe, advise, act after approval, and act autonomously—to balance efficiency and risk.
Liberty Mutual, Travelers Discuss When Not to Modernize Systems
In a panel discussion at the 2026 MIT Sloan CIO Symposium, Liberty Mutual Insurance Global CIO Monica Caldas and Travelers Insurance Chief Technology and Operations Officer Mojgan Lefebvre explored the trade-offs of enterprise system modernization. Both executives noted that in the new AI-driven era, not all legacy systems require immediate upgrades, and knowing what not to do is equally important. Caldas introduced Liberty Mutual's technical debt radar framework to distinguish systems that need immediate attention, monitoring, or stability maintenance; Lefebvre shared Travelers' strategy of isolating subsystems behind APIs and investing around customer value. Thanks to a cloud migration started eight years ago, Travelers has placed 70% of its computing workloads in the cloud, laying a solid foundation for AI applications.