Shield has added two AI agents to AmplifAI, its agentic suite for digital communications surveillance and investigations, extending the platform’s focus from detection and investigation into alert resolution and language coverage.
The new Alert Closure Agent is designed to assess flagged communications using message content, risk language and conversation context, and to close alerts where the surrounding context indicates that no compliance risk is present. The company says customer evaluations of the agent resulted in a 77.3% reduction in false positives.
The launch addresses one of the most persistent operating problems in communications surveillance: large volumes of low-risk alerts that absorb reviewer capacity before cases can be escalated for more meaningful investigation. The company cites industry estimates that firms process roughly one million Level 1 alerts each year, with fewer than 0.02% progressing beyond initial review and 93% of firms identifying false positives as a significant operational challenge. The agent is positioned as a governed workflow tool rather than a standalone decision-maker, with closure reasoning recorded in the alert detail, closed alerts capable of being reopened, and quality assurance workflow steps configurable by the firm.
The operational significance is the move from AI-assisted triage to controlled case disposition. Surveillance providers have increasingly used artificial intelligence to classify communications, prioritise alerts or surface contextual information, while regulators have observed firms exploring AI to filter false alerts. This latest release pushes further into the workflow by allowing contextually clear false positives to be closed under defined oversight controls.
That distinction matters for regulated firms. The practical challenge is not simply whether AI can reduce alert noise, but whether it can do so in a way that leaves a defensible evidence trail for compliance, audit and supervisory review. The company’s emphasis on recorded reasoning, reopenable alerts and configurable quality assurance is intended to support that governance requirement.
The second addition, the Language Expansion Agent, is aimed at multilingual surveillance gaps. Shield says the agent can identify risk across unmonitored or rare languages, bringing communications inside the compliance perimeter regardless of the languages a firm has selected for active monitoring.
Together, the two agents broaden AmplifAI across detection, investigation and governed resolution. The suite already includes a Noise Reduction Agent and Coverage Expansion Agent for detection, a Risk Reasoning Agent for triage and analysis, and Shiela, an agentic assistant for natural-language queries and investigation.
Shiran Weitzman, chief executive officer of Shield, framed the launch as part of a shift from task-specific AI toward “a coordinated system of specialized agents” across the surveillance lifecycle. He said the new agents are intended to support autonomy “where needed” while helping firms “keep human judgment at the centre.”
Tamar Sharir, chief product officer of Shield, linked the release to the long-running trade-off between scale, coverage and efficiency in compliance operations. She said the new agents are designed to give compliance programmes “coverage and capacity” across channels, languages and alerts.
The Alert Closure Agent and Language Expansion Agent are available as part of the AmplifAI suite. The company says the Alert Closure Agent is already in deployment with a Tier 1 financial institution.
Subscribe to our newsletter
