What 410 Enterprise Decision-Makers Revealed About Deploying AI Agents.
Two thirds already use agents in production. Not one has given an agent full autonomy. IDC asked 410 enterprise decision-makers what they have deployed, what those agents are allowed to do on their own, and what they are budgeting for the next twelve months.
66% are already using agents in production6x growth in deployments planned by January 2027
0% have deployed a fully autonomous operation with policy guardrails37% recommend actions requiring sign-off before an agent acts at all
Agent budgets double inside twelve monthsBudget allocation for agentic AI doubles in twelve months
410Decision-makersDirector level and above
US + UKMarkets coveredDecision-makers, not practitioners
8Industry sectorsBFSI through utilities
$500M+Company revenueGlobal annual minimum
Agent Counts Are Climbing. The Need for Coordination Is Paramount.
The average organization ran 2.7 agents in January 2026 and expects 15.6 a year later, six times as many. Coordination has not moved at anything like that rate. 51% still pass work between agents using basic rule-based handoffs, which IDC calls structurally inadequate, and only 7% have reached multi-agent collaboration with monitoring.
Average AI agents per organization
2.7Jan 2026 actual
5.3Mid-2026 actual
15.6Jan 2027 planned
Agent orchestration maturity today
Basic: rule or workflow-based handoffs51%
Emerging: shared memory or event-based36%
Advanced: multi-agent collaboration7%
None: agents run independently6%
Five Findings Worth Planning Around.
The deck runs to seven sections, from deployment maturity through to how buyers pick a vendor. These five are the ones teams are acting on first.
66%
Already in Production
Two thirds of organizations are already using agents in production, and they plan six times as many deployments by January 2027. IDC notes that growth is far outpacing governance.
2×
Budgets Double in a Year
Organizations are doubling their budget allocation for agentic AI over the next twelve months, while deployments grow sixfold over the same period.
0%
Nobody Is Fully Autonomous
Not one organization has a fully autonomous agent in production. 46% execute low-risk actions autonomously, 37% recommend actions requiring approval before an agent acts, and 9% permit execution across workflows with audit controls.
4–6
Contracts Stall Between Teams, Not on Someone’s Desk
59% of contracts pass through four to six separate systems before signature, and 79% of organizations say contracts hit handoff delays often or sometimes. 59% say those delays add four to seven days. The most common source of value leakage is simply not being able to see what the contract says, cited by 43%.
68%
Generic Chatbots Fall Short Here
68% of decision-makers say today’s generic LLM-powered chatbots are only partially sufficient or worse for legal and procurement tasks. 37% point to data privacy and security limitations.
What Legal, Compliance and Risk Teams Want Agents to Fix.
Legal, compliance and risk teams know exactly what they need from agentic AI. Compliance risk insights, obligation tracking and faster cross-functional handoffs are all highly ranked priorities.
Legal, Compliance & Risk
Top 5 functional improvements expected from AI agents
Improved risk and compliance insights66%
Improved visibility into obligations and renewals42%
Faster contract drafting and review41%
Improved supplier performance visibility37%
Improved collaboration with procurement and finance34%
How the IDC Survey Was Conducted.
Methodology
Survey interviews
Sample
410 enterprise decision-makers actively deploying or evaluating AI agents across legal and compliance, procurement and sourcing, finance and accounting, and logistics and supply chain