Gartner says AI ops will add tools before it cuts work

Gartner expects AI to take over a quarter of IT infrastructure and operations work by 2030, but the analyst firm says the shift will first add tools, management layers and outage risk. In its 2026 Hype Cycle for AI in IT Operations, published July 10, Gartner forecasts rapid adoption of agentic AI in production operations even as the technology makes environments harder to control in the near term.

The near-term message is a warning against the clean consolidation story many AI operations vendors are selling. Gartner says agents may eventually query multiple systems and reduce reliance on specialist tools, but for the next few years IT organizations should expect additional control points and more specialized observability, orchestration and management products. The firm says consolidation across vendors and markets could later shrink that tooling footprint, but it does not give a date for when that cleanup arrives.

Gartner’s outage forecast is blunt. By 2028, it expects 40% of infrastructure and operations organizations using agentic I&O at production scale to suffer a business-critical service disruption, compared with less than 1% in 2026. The firm attributes that risk to AI becoming part of the operational control plane, rather than remaining a recommendation layer.

Adoption is still expected to accelerate. Gartner predicts 60% of enterprises will use agentic AI in IT infrastructure operations by 2029, up from under 10% today. It also expects human approval to become less common: only 20% of AI-suggested actions will require human-in-the-loop approval in 2029, down from 80% in 2025. Gartner says that change will rely on deterministic guardrails, meaning policy-based limits that define what an AI system is permitted to do.

By 2030, Gartner expects half of infrastructure and operations teams to have been restructured after companies invest in agents for complex management tasks. For the work that remains, the firm forecasts AI will touch every human-performed task: 75% of work will be done by people using AI assistance, while 25% will be performed by AI without human execution.

Which tools Gartner sees maturing first

Gartner identifies several categories as likely to mature over the next couple of years. Those include vendors building generative AI-first IT operations products, virtual assistants for self-service troubleshooting, and generative AI-augmented CloudOps tools that analyze logs, metrics, traces, configuration data and change events to create scripts, infrastructure-as-code templates, runbooks and post-incident reports.

The firm also points to autonomous endpoint management, which can configure machines to match user profiles and handle patches, as well as network AI and automation tools that monitor networks and recommend configuration changes. Gartner says service providers are the direct beneficiaries of some network automation, with customers potentially seeing performance gains downstream.

Four categories are flagged as the most impactful over the next two to five years: agentic AI observability, agentic NetOps, augmented FinOps and multiagent systems. Agentic AI observability is aimed at monitoring AI agents themselves and detecting when they behave incorrectly, a requirement Gartner links to governance and cost control. Agentic NetOps applies automation to network management. Augmented FinOps uses AI for cloud budget planning and optimization. Multiagent systems coordinate several agents to complete a task.

The forecast gives buyers a more complicated procurement picture than vendor marketing suggests. AI operations tools may reduce manual work over time, but Gartner’s own assumptions point to a period of expanded tooling, looser human review and higher operational risk before any promised simplification shows up.

This story draws on original reporting from The Register.