Gartner predicts AI agent surge is reshaping expectations for enterprise automation and customer experience. Recent Gartner reports and forecasts from 2025 through 2026 outline rapid adoption scenarios alongside stark warnings about cost, governance and vendor quality.
The forecasts are vivid: large shares of enterprise applications and daily decisions will be influenced or driven by agentic AI by the end of the decade, even as Gartner cautions that many projects will be canceled unless organizations sharpen ROI, controls and integration plans.
The line numbers
Gartner's June 25, 2025 forecast grabbed attention by predicting that over 40% of agentic AI projects will be canceled by the end of 2027, citing escalating costs, unclear business value and inadequate risk controls. The same release warned of pervasive 'agent washing' and estimated that only about 130 vendors are truly agentic amid thousands marketing themselves as such.
On the adoption side, Gartner projected that at least 15% of day-to-day work decisions will be made autonomously through agentic AI by 2028, up from roughly 0% in 2024. It also forecast that 33% of enterprise software applications will include agentic AI by 2028, compared with less than 1% in 2024.
Additional Gartner forecasts extend the timeline: by the end of 2026, 40% of enterprise applications could be integrated with task-specific AI agents, and in an upside scenario agentic AI might capture roughly 30% of enterprise application software revenue by 2035, an estimated $450 billion opportunity.
Why cancellation risk is so high
Gartner's analysis points to three root causes for the projected cancellations: rising and unpredictable costs, unclear or unproven business value, and weak risk management or governance frameworks. Many early projects were launched as experiments and proofs of concept driven by hype rather than clear ROI, a point Gartner analysts such as Anushree Verma have emphasized.
'Agent washing' compounds the risk: vendors relabel existing chatbots, assistants or RPA as agentic without delivering true autonomous capabilities, creating market confusion and false expectations. Gartner recommends buyers be skeptical of vendor claims and insist on demonstrable task autonomy and measurable outcomes.
Technical complexity is another contributor: integrating agentic systems with legacy stacks, data stores and business processes can be costly and time consuming, and without rethought workflows the expected efficiency gains often fail to materialize.
Adoption timeline and evolving market structure
Gartner maps a progression from embedded assistants today to task-specific agents in 2026, collaborative multi-agent applications in 2027, and cross-application agent ecosystems by 2028, 2029. This staged view underscores that organizations have narrow windows to define strategy and integration approaches.
The IT Symposium coverage and forecasts from late 2025 raise the stakes: organizations using multi-agent AI for the majority of customer-facing processes are expected to dominate, and Gartner even projects that by 2028 up to 90% of B2B buying could be intermediary-mediated by AI agents, routing more than $15 trillion in B2B spend through agent exchanges.
Despite the hype, Gartner's market sizing and vendor-count estimates suggest a consolidation and certification phase is likely as buyers separate genuinely agentic platforms from rebranded tooling, and as interoperability, orchestration and ecosystems become deciding purchase criteria.
Customer service and workflow transformation
Gartner's March 5, 2025 prediction is transformational: by 2029, agentic AI will autonomously resolve 80% of common customer-service issues without human intervention, producing an estimated 30% reduction in operational costs. Daniel O'Sullivan of Gartner described agentic AI as a potential game-changer for customer service.
Coupled with a Gartner Q&A prediction that by 2030 half of all service requests will be initiated by machine customers powered by agentic AI, enterprises should expect both volume and complexity shifts in service routing. This implies redesigning escalation paths, SLAs and monitoring to accommodate AI-to-AI interactions.
For workflow owners, the implication is clear: don't simply wrap legacy processes in an agentic veneer. Gartner advises rethinking workflows to exploit agent autonomy, measuring impact on end-to-end outcomes and retraining staff into new oversight and exception-management roles.
Security, privacy and data governance challenges
Gartner's April 2025 security forecast warns that agentic tools will accelerate attackers too, predicting that AI agents could reduce the time to exploit account exposures by 50% by 2027. That projected acceleration means incident response, detection and authentication systems will need agent-aware capabilities across web, app, API and voice channels.
Data collection and consent are also in focus: Gartner's Nicole Greene warned that by 2027 up to 85% of customer data could be collected from automated interactions or those led by AI agents, raising privacy, governance and consent challenges for marketers and CX teams. Organizations must update consent models and data inventories to account for machine-initiated events.
Operationally, firms should invest in agent-aware security controls, provenance and audit trails, and privacy-by-design for agent interactions, because the combined effect of more automated interactions and faster exploitation timelines increases systemic risk.
Vendor landscape, agent washing and ecosystem dynamics
Gartner explicitly calls out 'agent washing' as a major market problem: many vendors reposition legacy products as agentic without delivering genuine autonomy or the necessary orchestration features. The June 2025 Gartner research estimated only about 130 truly agentic vendors among thousands, signaling a high bar for credible capability.
Media and analyst coverage from outlets such as Reuters and trade publications echoed Gartner's core claims, amplifying concern that buyer due diligence and reference checks will be critical. For buyers, the immediate challenge is separating marketing from substance and demanding real-world performance metrics.
At the same time, Gartner's scenario planning points to the emergence of agent ecosystems: interoperability and agent orchestration will become strategic levers. Vendors that enable safe, observable and composable agent networks are likely to capture premium positioning in the years a.
Practical guidance for CIOs and business leaders
Gartner's tactical recommendations repeat across its 2025 releases: prioritize agentic use cases that deliver clear business value and measurable ROI, prefer workflow redesigns over superficial embedding, and enforce governance and risk controls from day one. These steps reduce the chance of wasteful projects and cancellations.
Leaders should also invest in orchestration, interoperability and monitoring: treating agents as first-class components in architecture and operations helps manage complexity and unlocks multi-agent collaboration. Gartner advises defining agent strategy within a 3, 6 month window to avoid being outpaced.
Finally, build cross-functional teams combining business owners, security, privacy, legal and platform engineers to set performance targets, compliance guardrails and escalation rules. That cross-functional approach is precisely what will separate successful agentic deployments from canceled experiments.
Gartner predicts AI agent surge brings both immense opportunity and significant risk. The forecasts are not a green light to rush: they are a call for disciplined, value-focused adoption backed by governance, security and realistic integration plans.
Organizations that marry clear ROI use cases with solid controls and a willingness to rethink workflows can capture major benefits, while those that chase hype without preparation risk wasting investment or harming customer trust. The next three years will separate leaders who thoughtfully adopt agents from those who merely market them.
