As organizations scale their use of AI, security teams are confronting two converging challenges: AI sprawl, the rapid proliferation of AI services and autonomous agents, and agent sprawl, the uncontrolled growth of individual AI agents that can act independently, traverse systems, access sensitive data and execute actions with far-reaching consequences. The result is critical blind spots for security teams; they can’t see what AI agents are running, what permissions they hold, what data they can access, or which applications they touch. Reco AI Agent Security was developed to address these blind spots. Reco’s first-of-its-kind capability gives security and risk teams complete visibility and control over all AI agents operating across their SaaS ecosystems. The new capability is available immediately as part of the Reco unified SaaS security platform. “Enterprises today don’t just have hundreds of connected SaaS apps — they have thousands of connected AI agents operating in the background,” said Ofer Klein, CEO and co-founder of Reco. “Unlike traditional SaaS plugins, AI agents can act autonomously and span identity, data and systems, exponentially increasing risk when misconfigured or unmanaged. Reco AI Agent Security gives security teams the visibility and control they need — all from the same platform they use to govern their SaaS estate.” Traditional SaaS security tools provide visibility into applications and user identities, but they were not built to account for autonomous AI agents — which can act without direct human interaction, possess broad permissions to sensitive systems, connect across multiple SaaS platforms and operate outside the visibility or control of IT and security teams, explained Reco executives. Reco AI Agent Security inventories every AI agent, maps its access, permissions, connections and risk posture and enables security teams to determine which agents should be sanctioned, restricted, or blocked before they introduce risk. While recent trends in AI security focus on posture management and runtime protections across cloud and AI models, those capabilities do not directly address the risk introduced by autonomous AI agents embedded in SaaS workflows — even when they have broad permissions and cross-system reach. Reco is not an AI-only security tool; Reco secures applications, identities and now AI agents across the entire SaaS environment — bringing agent security into the same framework used for SaaS security governance. Global telecom operators currently use AI most extensively for customer service and support, followed by cybersecurity, showed figures presented by the World Broadband Association. Significantly fewer telcos are applying AI to network automation and design. Analysts at WBBA, however, argued that moving forward, network enhancements carry the biggest potential for AI among telcos. With business either starting cloudnative or migrating over to cloud-based platforms at the same time as the demand for AI services has increased rapidly, the shape and speed of network growth has changed too, said WBA. Most data‑center networks were never built for the scale, performance or resiliency that AI workloads demand. In turn, “this provides a major opportunity for communications service providers to refresh their networks and to reposition for optimal AI usage,” said WBBA. Reco Tackles AI & Agent Sprawl Across SaaS Stacks How Telcos Use AI AI & AUTOMATION Source: Splunk CISO survey, 2026 How has your contact center operations head count changed over the past 12 months? Source: Execs in the Know Which risks related to AI are most pressing for your organization? Source: JumpCloud survey of IT leaders, 2026 In which areas of your organization are you using (or planning to use) AI the most? Source: WBBA; Omdia survey of global telecoms. Incident reduction Improved MTTD Improved MTTR 4% 13% 14% 80% 5% 74% 33% 38% 29% 2023 2024 2025 29% IN-HOUSE AND OUTSOURCED COMBINED 46% 25% 19% 42% 39% 43% 42% 48% 38% 39% 38% 46% 36% 42% 28% 41% 26% 33% 24% 32% 41% 39% 31% 22% 21% 17% 11% Unauthorized access to sensitive data AI-generated phishing or malware Data leakage or compliance violations Unregulated use of AI tools AI attacking AI Model inaccuracy or hallucination Shadow IT via AI apps Lack of AI explainability/ auditability Customer service and support Cybersecurity & treat detection Cloud & IT operations Sales, marketing, and customer insights Today In two years Network automation, operations, and design Product development & innovation Corporate functions Increased Kept same Decreased 10 CHANNELVISION | SPRING 2026
RkJQdWJsaXNoZXIy NTg4Njc=