The rapid rise of artificial intelligence (AI) agents is unlike any past technology shift. It creates a significant opportunity for managed service providers (MSPs) to expand and enhance their offerings.
AI agent adoption accelerates at unprecedented speed
A survey of 600 global data leaders across the U.S., Europe, and Asia Pacific conducted by Informatica, a unit of Salesforce, finds nearly half (47 percent) have already embraced agentic AI to some degree.
Similarly, analysis of telemetry data from 20,000 organizations published by Databricks shows multi‑agent AI workflows have surged 327 percent in just four months.
The pace of adoption is rapid, and the ripple effects across IT environments are only just beginning.
Infrastructure, skills, and data quality create major barriers
As organizations deploy AI agents, they are running squarely into multiple challenges. A global survey of 505 data and analytics leaders finds technological infrastructure (54 percent) and skills (51 percent) are the top two barriers to AI adoption.
Conducted by LeBow College of Business at Drexel University in collaboration with Precisely, the survey also reveals gaps in an organization’s ability to deploy AI at scale (30 percent), responsible AI expertise (29 percent), the ability to translate business needs into AI solutions (28 percent), and overall AI literacy (27 percent).
In addition, many organizations will soon realize their data quality is nowhere near sufficient for successful agentic AI workflows. This leads to adoption becoming further complicated.
Security risks multiply as AI agents become “new end users”
The most concerning issue, however, is the significant security risk introduced by AI agents without proper guardrails. It has already been demonstrated how easily an indirect prompt‑injection attack can instruct an AI agent to send sensitive data to a malicious actor.
This risk is escalating as end users independently adopt AI tools. Shadow AI usage is already widespread, and the problem will worsen as employees begin sharing data with self‑deployed agents or the expanding number of AI‑enabled SaaS applications.
Functionally, AI agents are becoming a new class of end user—often inheriting the permissions of whoever deploys them. Soon, for every human user, there may be ten agents performing tasks autonomously.
Most organizations already struggle to manage and secure traditional endpoints. The rapid influx of AI agents will only exacerbate these challenges.
A massive opportunity for MSPs to lead in AI governance
It’s only a matter of time before the first major breaches are traced back to compromised or misconfigured AI agents. These incidents may occur either through direct attacks or through unintentional actions resulting from poor training and weak guardrails.
When that happens, MSPs who can clearly demonstrate expertise in governing, securing, and managing AI agents will be in exceptionally high demand.
The organizations that get ahead of this trend today will be the ones shaping the next era of IT services.
Photo: Lightspring / Shutterstock
This post originally appeared on Smarter MSP.