AI Agent Sprawl Hits Corporate Frontlines

In 2023–2024, enterprises across sectors began deploying custom AI agents at unprecedented scale. Platforms such as Anthropic’s Claude Cowork let any employee build a task-oriented bot, prompting a surge that executives label “AI agent sprawl.” At Magnum Ice Cream, CIO Michael Friedlander warns of uncontrolled growth. FICO’s chief customer officer Mike Trkay reports daily creation of dozens of agents by a 3,500-person workforce. DaVita’s CIO Madhu Narasimhan cites more than 10,000 agents in use for patient-care operations.

Futurism

A Gartner survey found only 13 % of firms believe they have robust AI-agent governance. Gartner projects the average Fortune 500 company will run over 150,000 agents by 2028, up from fewer than 15 today. Magnum Ice Cream notes that each agent consumes token-based compute, creating unpredictable cost pressures. FICO’s staff generate dozens of agents daily, while DaVita reports more than 10,000 active agents.

Futurism

Friedlander says Magnum Ice Cream must “condense and centralize” agents to curb cybersecurity risk and token spend. Trkay emphasizes that agent creation occurs at every hierarchical tier, reflecting a company-wide automation push. Narasimhan stresses that scaling AI tools for patient care must be done “with safety.” Gartner’s guidance calls for formal governance frameworks to address the gap.

Futurism

Analysts view the unchecked proliferation as a self-inflicted governance gap. Without oversight, firms risk heightened cybersecurity exposure and unpredictable operating expenses.

Futurism

• “Because everybody can do it, we’re probably going to end up with a lot of people having the same types of agents,” — Michael Friedlander, CIO, Magnum Ice Cream

Futurism

• “Depending on how all of this will turn out, there’ll be tokens, and then there’s going to be cost, and then you end up with, ‘How do we manage this to make sure that it’s under a financially responsible model?'” — Michael Friedlander, CIO, Magnum Ice Cream

Futurism

• “The chief customer officer of FICO Mike Trkay bragged to the WSJ that its 3,500 employees are creating dozens of new AI agents every day “at every tier of the hierarchical structure.” — Mike Trkay, Chief Customer Officer, FICO

Futurism

• “Because we care for our patients, we have to scale with safety,” — Madhu Narasimhan, CIO, DaVita

Futurism

Gartner’s 150,000-agent forecast signals that centralized governance will become a strategic imperative. Magnum Ice Cream plans to consolidate agents, while DaVita emphasizes safety-first scaling. Industry observers expect the formal governance frameworks outlined by Gartner to guide the next phase of enterprise AI adoption.

Futurism