As cybercriminals increasingly use generative AI to scale phishing, impersonation and social engineering attacks, security teams are looking for ways to bring trusted threat intelligence into the AI tools employees already use every day. Cybersecurity leader Red Sift today announced new integrations for Claude and ChatGPT that enable users to assess domains and investigate email security risks directly within the large language model (LLM) environments where they are already working.
“Attackers adopted AI faster than most defenders, and the cost gap is the whole story,” said Rahul Powar, co-founder and CEO at Red Sift. “When launching a phishing campaign costs almost nothing, getting the fundamentals right is the cheapest defense you have. Radar Lite is a free, simplified preview of what our broader platform can do: a quick check that your DNS and email security posture is where it should be for today’s threats, and now it runs inside Claude or ChatGPT, so the first step takes minutes instead of becoming a project.”
The launch comes amid growing concern about both the offensive and defensive implications of AI. IBM’s newly released 2026 Cost of a Data Breach Report found AI-driven cyberattacks surged 56% globally over the past year, while Gartner estimates AI can reduce the cost of launching phishing and spam campaigns by as much as 95%. At the same time, major AI developers are facing increased scrutiny after OpenAI and Anthropic recently disclosed incidents involving advanced AI models exceeding expectations during cybersecurity testing, including cases where models interacted with or gained access to real-world systems in controlled environments.
As organizations increasingly rely on AI assistants for everyday work, security leaders are looking to ensure those same platforms can help identify and mitigate emerging threats rather than inadvertently amplify them.
Rather than requiring security teams to pivot between specialized tools and browser tabs, the integration allows users to query a domain inside Claude or ChatGPT and receive an assessment of key email and domain security controls, including DMARC and other anti-spoofing protections. The goal is to make security intelligence as accessible as the AI assistants increasingly becoming a first stop for research, analysis and decision-making.
Security teams can use the capability to investigate suspicious domains, evaluate potential phishing risks and quickly understand configuration weaknesses. The results are delivered in plain language, highlighting what matters most and prioritizing remediation steps.
The announcement reflects a broader shift in cybersecurity: embedding security intelligence directly into AI-powered workflows. As AI becomes a standard interface for work, organizations are increasingly looking to place security insights at the same point where decisions are being made. The integrations are available immediately and can assess any domain, providing comparative context against industry peers as part of the analysis.
