Businesses
Why Some Companies Succeed With Agentic AI While Others Remain Stuck in Pilot Projects
05 Nov 2026
New IDC research commissioned by Solace finds that companies with mature real-time data systems are more than three times as likely to achieve measurable results across most of their AI projects. Solace CTO Josh Carroll explains how data access, governance and enterprise architecture influence the success of autonomous AI deployments.
Companies are investing heavily in autonomous AI agents, but getting these systems to work reliably across business operations remains difficult. New research suggests that the difference between successful deployments and stalled pilot projects often comes down to how businesses manage their data.
A 2026 study conducted by IDC and commissioned by Solace surveyed 623 senior technology decision-makers at companies generating more than $1 billion in annual revenue. The findings show that organisations with mature real-time data infrastructure are more likely to deploy AI agents successfully and report measurable business results.
The research, titled 2026 State of Real-Time Data: Agentic Enterprises Are Running on Real-Time Data, examines how enterprise data architecture affects the adoption of autonomous AI applications.
Joshua Carroll, Chief Technology Officer at Solace, argues that businesses need to reconsider how information moves between their systems as AI agents take on more operational responsibilities.
Real-Time Data Separates Successful AI Deployments From Early Experiments
Many enterprises have already introduced AI agents into their operations, but the research identifies substantial differences in how extensively these systems are used.
According to IDC, 80% of surveyed organisations are investing in or already operating AI agents, while 90% have increased their attention to real-time data to support their plans.
The findings suggest that interest in agentic AI is widespread among large businesses, although operational maturity varies considerably.
IDC divided respondents into four groups based on their real-time data capabilities, AI adoption and reported business outcomes: Emerging, Developing, Advanced and Leaders.
Among organisations classified as Leaders, 76% use real-time data across most or all of their workflows, compared with only 7% of Emerging organisations.
The more mature group also reported better results from AI deployment, with 67% of AI projects delivering measurable outcomes, compared with 35% among Emerging organisations.
These findings demonstrate a strong relationship between real-time data maturity and reported AI performance, although the survey does not establish that better data infrastructure alone causes those results.
Why AI Agents Struggle With Outdated Enterprise Information
AI agents need access to business information to understand situations and decide which actions to take.
A customer service agent may need recent account updates, while an inventory management system must know which products are available before processing orders.
Problems arise when information is stored across disconnected applications or updated through scheduled batch processes.
An AI agent working with yesterday's inventory figures could recommend products that are no longer available. A system handling customer enquiries might also provide inaccurate information when recent transactions have not reached its database.
IDC found that 40% of surveyed organisations identified connecting AI agents to reliable enterprise data in real time as one of the main barriers to production deployment.
Data quality and consistency were also identified as the most frequently cited technical difficulty, affecting 46% of respondents.
For businesses, these results suggest that improving AI models may have limited value when the underlying systems cannot supply accurate information when it is needed.
Companies With Mature Data Systems Report Better AI Results
The research found that organisations with stronger real-time data capabilities were more advanced in moving autonomous agents into production.
Among Emerging organisations, 20% reported having several AI agents operating in production, compared with 59% among Leaders.
The survey also found that companies in the leading maturity group reported average annual improvements of approximately 23% across selected business outcomes, including speed to market, risk reduction and responsiveness to changing conditions.
These improvements were reported by the organisations surveyed and should not be treated as independently audited financial returns.
Interestingly, the study found that company size had little relationship with real-time data maturity, while differences between industries were relatively modest.
This suggests that having a large technology budget does not necessarily translate into a more advanced approach to enterprise AI.
Instead, organisational practices and the way technology teams manage data appear to influence how effectively businesses move from experiments to operational systems.
Five Approaches That Can Help Businesses Scale Agentic AI
Drawing on the IDC findings, Carroll identifies five areas companies should consider when developing enterprise AI strategies.
The first is prioritising real-time data movement. Businesses need systems that can provide current information to AI applications without depending entirely on historical databases or delayed updates.
This may require event-driven architectures, where changes in business systems can be communicated to other applications as they occur.
The second involves reducing tool fragmentation. Large companies often maintain separate platforms for application integration, messaging and data streaming, which can create duplicated work and complicate maintenance.
IDC found that mature organisations were more likely to standardise how information moves across their technology environments.
Governance forms the third area. As AI agents gain permission to interact with business applications, companies need clear policies covering data access, system permissions and responsibility for automated actions.
The fourth involves supplying agents with reliable business context. Current data is useful, but AI applications also need to understand the rules and operational conditions surrounding that information.
Finally, companies need to continue investing as AI deployments expand, ensuring that the systems supporting autonomous applications can handle growing demand.
Governance Becomes More Important as AI Agents Gain Autonomy
Moving AI agents into production introduces questions about how businesses control their actions.
An agent connected to customer records, financial systems or internal applications may have access to information that requires strict handling.
Organisations therefore need to define what each agent can access and which actions it is authorised to perform.
Governance arrangements may also include monitoring automated decisions, reviewing unexpected behaviour and maintaining records of actions taken through connected applications.
These measures can help businesses identify problems before they spread across multiple systems.
For companies operating in regulated industries, the ability to demonstrate how AI agents reach decisions may also affect whether particular applications are suitable for deployment.
A well-designed data infrastructure can support these controls by making information more accessible while maintaining defined permissions.
AI Investment Is Expected to Continue Growing
Despite the difficulties associated with deploying autonomous systems, businesses surveyed by IDC remain committed to increasing investment.
Between 94% and 97% of organisations across the four maturity groups expect to increase spending on real-time data and agentic AI over the following year.
The research also found that approximately 90% of respondents said their real-time data investments had met or exceeded expectations.
These results suggest that companies see value in improving their data systems, even when their AI programmes have not yet reached full operational maturity.
However, continued investment will also require businesses to assess whether the financial benefits justify the costs of infrastructure, integration and ongoing management.
AI deployment can create expenses beyond purchasing models or software, including data preparation, application maintenance and employee training.
Measuring those costs alongside operational improvements can help organisations determine where additional investment makes sense.
What the Research Means for Enterprise AI Strategies
The IDC findings suggest that businesses planning larger AI deployments need to examine the condition of their existing data infrastructure before introducing more autonomous applications.
Disconnected systems and inconsistent information can restrict what an AI agent is able to accomplish, regardless of the model's technical capabilities.
Companies with stronger real-time data arrangements appear better positioned to move agents into daily operations and measure their results.
For business leaders, this creates an opportunity to connect AI investment decisions with broader improvements in technology architecture.
Rather than funding isolated projects without considering their dependencies, organisations can examine how shared data services and integration platforms might support several applications.
The research also indicates that successful AI adoption depends on organisational decisions about governance and how technology teams collaborate.
As businesses continue developing autonomous applications, their ability to provide accurate information and maintain control over automated processes will influence how successfully these systems move beyond experimentation.
About Solace
Solace is a Canadian enterprise technology company headquartered in Ottawa, specialising in real-time data distribution and event-driven application integration.
Its platform helps organisations connect applications, cloud services and operational systems so information can move between them as business events occur.
Solace works with enterprise customers across industries including financial services, manufacturing, retail and transportation.
The company commissioned IDC to examine how real-time data capabilities influence agentic AI adoption and business outcomes.
Official website: https://solace.com/
About Joshua Carroll
Joshua Carroll is Chief Technology Officer at Solace, where he leads the company's technology strategy and works with enterprise customers on real-time data systems and AI deployment.
He has more than 20 years of experience in financial services and enterprise technology, including previous leadership roles at GitLab, Moody's Corporation and RBC Capital Markets.
Carroll's work focuses on enterprise architecture, secure software development and the systems required to support AI applications in production.
Share
Sara Srifi
Sara is a Software Engineering and Business student with a passion for astronomy, cultural studies, and human-centered storytelling. She explores the quiet intersections between science, identity, and imagination, reflecting on how space, art, and society shape the way we understand ourselves and the world around us. Her writing draws on curiosity and lived experience to bridge disciplines and spark dialogue across cultures.
