Agentic AI adoption is increasing rapidly in Australia, with new technology creating both opportunities and risks in sectors such as banking and financial services.
According to research from Gartner, 41% of Australian organisations are already utilising agentic AI in some capacity. Projections suggest that by 2029, agentic AI will autonomously resolve around 80% of customer service issues, promising significant cost reductions and heightened operational efficiency.
Yet, as businesses leverage these advancements, concerns have grown regarding the use of agentic AI by cybercriminals. Analysts note that 80% of ransomware attacks now incorporate AI, facilitating more sophisticated malware, phishing schemes, and deepfake-driven social engineering tactics such as fraudulent customer service calls. These developments have made it imperative for organisations to address both the benefits and the risks associated with AI-driven automation in customer-facing operations.
Banking sector challenges
The banking industry, long a target for financial crime, faces particular concerns due to the dual nature of AI: while it increases efficiency, it also opens new avenues for abuse. In response, Proofpoint has launched what it describes as the first agentic AI solution focused on Human Communications Intelligence, aiming to assist enterprises in detecting and mitigating conduct and compliance risks in real time.
This new approach, centred on Digital Communications Governance (DCG), shifts from post-incident compliance reviews to real-time, AI-powered risk detection and prevention. Proofpoint's system is designed to help organisations understand human intent within digital communications, transforming these interactions into actionable intelligence for compliance, legal, and security teams.
A study involving 1,600 global Chief Information Security Officers (CISOs) found that over one-third consider collaboration tools such as Slack, Teams, and Zoom to be the primary source of organisational risk - surpassing concerns related to GenAI chatbots, perimeter devices, cloud storage, and Microsoft 365. Historically, compliance teams have relied on passive tools that capture communications but offer minimal AI-based interpretation. Growing insider threats, financial misconduct, and increased regulatory scrutiny have signalled a need for intelligence solutions that can evaluate real-time conversations and assess risk based on detected intent and behaviour.
Technology and capabilities
Proofpoint's Human Communications Intelligence solution, powered by its Nuclei technology, claims to go beyond existing solutions that primarily flag keywords or communication patterns. Instead, it uses "true agentic AI" to interpret and contextualise human conversations in real time, spanning more than 80 types of communication channels. The intelligent agents assess conversations and flag emerging risks, providing transparent explanations for each action.
"Legacy connectors are merely proxies that pass content downstream, offering little intelligence. Proofpoint has reimagined capture by moving beyond simply collecting messages to interpreting and reasoning in real time. This transforms the world's largest source of behavioural data - human conversations - into actionable intelligence that empowers compliance, legal, and security teams to prevent risk before it escalates," said Harry Labana, SVP & GM of Proofpoint's DCG business.
Among the key capabilities of the Proofpoint DCG portfolio are:
- Real-time capture and reasoning across various channels including GenAI chatbots, social media, voice, and files.
- Explainable AI-powered supervision that reduces manual review workload by analysing communication tone, shorthand, emojis, and code snippets across multiple languages.
- Integration with Insider Threat Management tools to correlate risk-indicative communications with user activity, providing a comprehensive record of employee actions and communications.
The system seeks to address the challenge of false positives commonly associated with legacy supervision methods, claiming up to a 90% reduction in such occurrences, thereby enhancing accuracy and reducing compliance overhead.
Ethical and operational considerations
The increasing sophistication of agentic AI has prompted renewed discussions around ethical deployment and the need for robust safeguards. As AI becomes more capable of interpreting and acting on human intent, organisations are advised to maintain strong oversight mechanisms and continuously review the outcomes produced by these technologies.
Proofpoint has stated that the new Human Communications Intelligence agents are currently available, with further AI-driven features expected to be introduced later in 2025, and integration with Insider Threat Management scheduled for 2026.
