Technology 4 min read By Arthur Ellington
Palo Alto Networks launches AI-driven cybersecurity service using Claude and GPT models
Palo Alto Networks has unveiled a new AI-powered cybersecurity service that draws on large language models from Anthropic and OpenAI, deepening the push to automate threat detection and response for enterprise customers.
Palo Alto Networks has launched an AI-powered cybersecurity service that uses large language models from Anthropic and OpenAI, as the company moves to embed generative artificial intelligence more deeply into enterprise security operations.
The service combines Palo Alto Networks' existing security platform with models including Anthropic's Claude and OpenAI's GPT family, according to the company. The aim is to help security teams analyse threats, triage alerts and respond to incidents more quickly, reducing the manual workload that has long strained understaffed security operations centres.
The announcement places Palo Alto Networks alongside a growing group of cybersecurity vendors racing to integrate generative AI into their products. Microsoft, Google and CrowdStrike have all introduced AI assistants and copilots for security work over the past two years, betting that automation can help defenders keep pace with attackers who are themselves adopting AI tools.
For Palo Alto Networks, the move is also a commercial play. The company has been expanding its platform strategy, bundling firewalls, cloud security and threat intelligence into subscription packages that aim to lock in enterprise customers. Adding AI capabilities gives it another differentiator as it competes with Cisco, Fortinet and Check Point in the network security market, and with cloud-native vendors in the fast-growing segment for protecting workloads and data.
Using third-party models from Anthropic and OpenAI rather than relying solely on in-house technology reflects a broader trend in enterprise software. Many vendors now build on top of foundation models from a handful of AI labs, then fine-tune them with proprietary data and domain expertise. For cybersecurity, that domain knowledge — threat feeds, incident histories and detection rules — is the key asset that general-purpose models lack.
The service is aimed at security operations teams that face a daily flood of alerts, many of which turn out to be false positives. AI systems can summarise incidents, correlate signals across different tools and suggest remediation steps, allowing analysts to focus on the most serious threats. Vendors argue this can shorten response times and ease the pressure on skilled staff, though security leaders remain cautious about relying too heavily on automated decisions in high-stakes environments.
Palo Alto Networks has not disclosed detailed pricing or a full rollout schedule for the new service. The company has been investing heavily in AI across its portfolio, including tools for detecting anomalies in network traffic and automating policy enforcement. Its latest launch signals that the competition among security vendors is increasingly about which company can integrate the most capable AI models into reliable, enterprise-grade products.
The wider context is a cybersecurity industry under pressure on multiple fronts. Ransomware attacks, supply chain breaches and state-linked intrusions have raised the stakes for businesses and governments, while a persistent skills shortage leaves many organisations struggling to hire enough analysts. AI is seen by vendors and investors as one of the few ways to close that gap at scale.
Whether the new service delivers on that promise will depend on how well it performs in real-world environments, where attackers adapt quickly and false alarms carry a cost. Palo Alto Networks will need to convince customers that AI-generated recommendations are accurate, auditable and secure — particularly when the models themselves can be targets for manipulation. For now, the launch adds another significant name to the list of security firms betting that generative AI will reshape how enterprises defend their networks.
6
