IDC: Predicts 60% of Businesses Will Prioritize AI for Most Business Processes by 2029
The Zhitong Finance App learned that the International Data Corporation (IDC) predicts that by 2029, 60% of the world's top 2000 (G2000) companies will allow most business processes to operate in an AI-first manner. The premise of this prediction is that operations and decisions are left to AI agents. IDC also predicts that by 2028, 70% of the world's top 2000 companies (G2000) will have the ability to continuously verify decisions made by AI agents; currently, this ratio is only 14%.
To help business leaders grasp the pace of change and clarify action priorities, IDC divides them into three major categories based on the time windows corresponding to different forces of change:
Closing Window (Closing Window): The pioneers have established a leading edge, and this advantage continues to expand, and the gap between latecomers and leaders is getting wider.
An open window (Open Window): Business leaders still have an opportunity to shape the competitive landscape and influence future trends, but the time to act is rapidly shrinking.
Emerging Window (Emerging Window): New markets and value creation opportunities are emerging, and companies that take the lead are expected to gain a competitive advantage far beyond their latecomers.
The first type of window is closing - pioneering advantage is accumulating at an accelerated pace
1. The rise of open models is impacting the difficult problem of AI economics
By 2028, enterprises will shift 60% of AI workloads from proprietary large-scale language models to open source models, and use model routing strategies to reduce costs and hedge the risk of vendor lock-up. Companies that move faster will take the lead in solving the problem of overspending on intelligent inference — a problem that is limiting the large-scale expansion of AI applications.
2. The growth in the scale of data is surpassing the ability of enterprises to create value in data
IDC predicts that by 2028, 60% of companies will be unable to effectively transform their data into measurable business value. The root cause is that the scale of enterprise data is growing faster than the company's ability to define, manage, and govern the value of data. The challenge facing enterprises is no longer just how to obtain and store more data, but how to truly understand the value of data, establish effective data governance mechanisms, and transform data into business results.
The second type of window that is still open — the time to influence the future competitive landscape is rapidly shrinking
3. The speed at which agents are being deployed is surpassing corporate governance capabilities
A team of IDC analysts from 10 different research fields discovered the same key problem in independent research: the deployment of AI agents is accelerating, yet corporate supervision and governance mechanisms are failing to keep up.
IDC predicts that by 2028, the top 2000 global companies that fail to configure AI agents in production environments with an “emergency stop switch” (an emergency stop switch, a control mechanism that can shut down agents with abnormal behavior in a timely manner) will face two obvious risks and cost increases: liability insurance premiums will increase by 25%; and the number of issues found in major audits will triple.
Meanwhile, the deployment speed of AI agents continues to accelerate. According to IDC's “Future Enterprise Resiliency and Spending Survey” (Future Enterprise Resiliency and Spending Survey), as of July 2026, 95.4% of companies had deployed agents into an average of 11 production workflows, with a monthly estimated budget of close to US$118,000.
This set of data reveals an increasingly prominent contradiction: intelligent entities are entering the core business of enterprises at an unprecedented speed, and traditional governance systems that rely on manual approval, regular audits, and subsequent correction are becoming increasingly difficult to meet the requirements of real-time supervision.
4. Software contracts are redefining the boundaries of responsibility around AI agents
IDC predicts that by 2028, 75% of the world's top 2000 companies signed contracts with technology suppliers will include liability clauses for AI agents to initiate acts independently. At that time, if there is no such liability agreement in the supplier contract, it will be an important risk warning signal for companies when evaluating and selecting technology suppliers.
At the same time, enterprises will also face new challenges in terms of cost transparency for AI agents. IDC predicts that among global companies, although suppliers have not publicly adjusted the prices of products or services, the proportion of smart-related expenses actually borne by companies will increase from 20% now to 55% by the end of 2027.
This means that enterprise software procurement and cost management are undergoing profound changes. In the past, enterprises mainly focused on software licenses, subscription prices, and number of users; in the smart economy, enterprises must also pay attention to the dynamic costs generated by intelligent agents performing tasks independently, as well as corresponding behavioral responsibilities, risk sharing, and contractual constraints.
5. The globally unified technology architecture is becoming fragmented
A team of IDC analysts from nine research fields pointed out that increasingly fragmented regulatory requirements in different countries and regions are making it difficult for companies to continue to rely on a single global unified technology architecture. IDC predicts that by 2029, 35% of the world's top 2000 companies will establish AI model registration management systems and corresponding control mechanisms for different jurisdictions.
As countries develop different regulatory systems for data security, privacy protection, AI governance, and compliance requirements, multinational enterprises must reconsider the design principles of their global AI architecture. In the future, the consistency of global technical capabilities and the differentiation of local regulatory compliance will become two core requirements that must be taken into account simultaneously in enterprise AI architecture design.
A third type of window is opening - emerging market opportunities will be prioritized as pioneers
6. Supervisory and orchestration agents are spawning new job categories
IDC predicts that by 2029, the world's top 2000 companies will promote 30% of junior and intermediate executive positions to transition to AI supervisory and orchestration related jobs. This trend means that the way companies work is undergoing structural changes.
More and more employees will no longer directly perform every specific task, but will be responsible for setting goals, coordinating agents, overseeing the execution process, and judging and taking responsibility for the final results. It is worth noting that AI does not necessarily mean an overall decline in the number of junior jobs in enterprises. According to IDC's “Future of Work Survey” (Future of Work Survey), among mid-market companies with 1,000 to 4,999 employees:
49% of companies expect the number of entry-level employees to increase in the next 18 months.
17% of companies expect a reduction in the number of entry-level employees.
This shows that the impact of AI on the job market is not simply job replacement, but is accompanied by a reconfiguration of job content, skill requirements, and organizational division of labor.
7. AI is stepping out of the screen and into the real physical world
IDC analyst teams from eight research fields, including energy, manufacturing, retail, and connectivity technology, all observed an important development direction: AI is moving from digital space to physical operation scenarios, and is beginning to directly participate in real-world production, control, and business execution.
For example, IDC predicts that by 2030, 40% of power utilities will adopt a hierarchical authorization framework that allows AI agents to independently perform selected and authorized routine grid control operations, thereby reducing the need for intervention by manual operators by 25%.
This trend means that AI applications are breaking through traditional scenarios such as information processing, knowledge questioning, and decision support, and going deeper into fields such as production and operation, energy management, and physical infrastructure control. As AI agents gradually gain authority to perform real-world operations, enterprises must also establish authorization, supervision, security, and responsibility management mechanisms to match them.
8. Mid-market companies are already leading the big players in deploying smart devices
IDC's “Future Work Survey” of 1,146 business leaders found that mid-market companies showed greater enthusiasm than large enterprises in terms of cross-departmental deployment of AI agents. More specifically:
Of mid-market companies with 1,000 to 4,999 employees, 46% have deployed proxy AI across multiple departments.
In large companies with 5,000 or more employees, this percentage is 40%.
However, when it comes to building an AI governance system, the opposite is true. Only 31% of mid-market companies have established fully operational AI Centers of Excellence (AI Centers of Excellence) to coordinate and oversee AI governance, compared to 43% of large enterprises.
This discovery reveals a new phenomenon worth watching in the smart economy era: mid-market enterprises may gain a first-mover advantage in smart device deployment with more flexible organizational structures and faster decision-making speed; however, large enterprises still have a comparative advantage in establishing systematic governance mechanisms, improving organizational capabilities, and managing risk.
Future competition depends not only on the speed at which enterprises deploy AI agents, but also on whether they can establish governance and operation capabilities that match them while applying them on a large scale.