Namit D’Cruz highlights short-lived credentials, least-privilege access and shared security visibility as key measures to strengthen cyber resilience
As artificial intelligence (AI) reshapes the cybersecurity landscape, organizations face growing pressure to identify vulnerabilities, secure cloud environments and respond to attacks faster. According to Namit D’Cruz, RVP India & SAARC at Datadog, AI is accelerating the discovery of long-standing software flaws, while giving attackers access to tools that can help them exploit weaknesses more quickly.
“Recent incidents show that AI has changed the clock, not the basics,” D’Cruz said, pointing to the growing speed at which vulnerabilities are discovered and exploited. He noted that vulnerabilities are increasingly being exploited within a day of disclosure, underscoring the need for organizations to strengthen their security practices rather than respond with panic or simply add more security tools.
Cloud Security Remains Central to AI Protection
As businesses integrate AI capabilities into cloud applications, managing credentials and controlling access have become critical security priorities. D’Cruz observed that many of the challenges associated with AI security are rooted in familiar cloud security questions: who holds a credential, what resources it can access, and when that access expires.
He cautioned that organizations connecting to AI services through permanent credentials risk creating security gaps when those credentials are forgotten or remain active longer than necessary. As businesses adopt new AI services and applications, the proliferation of access keys could create additional opportunities for attackers if credentials are not properly managed and retired.
“The software supply chain carries the same risk,” D’Cruz said, highlighting the security implications of third-party software packages used by cloud applications. Such packages may be incorporated into applications without teams explicitly selecting every dependency, creating potential exposure if malicious code is introduced into the software supply chain.
According to D’Cruz, attackers are developing malicious code designed to locate credentials, making effective credential management and supply-chain security increasingly important.
Short-Lived Credentials and Least-Privilege Access
To address these risks, D’Cruz urged organizations to maintain a clear inventory of their applications, credentials and AI agents. He recommended moving towards short-lived credentials and least-privilege access, limiting permissions to only what is necessary for a specific task and reducing the period during which compromised credentials can be used.
He also stressed the importance of monitoring the actual behaviour of applications and systems. Understanding how workloads operate in real-world environments can help security teams identify suspicious activity and investigate potential threats more effectively.
The approach requires organizations to look beyond individual security products and establish a broader understanding of their technology environments, including the relationships between applications, infrastructure, users and access permissions.
Shared Visibility Can Improve Incident Response
D’Cruz said cybersecurity cannot remain the responsibility of a single team or tool. Effective incident response requires security, engineering and operations teams to work from a common view of the environment, with sufficient context to understand the source and potential impact of an alert.
He cautioned that runtime alerts viewed in isolation can slow investigations and delay coordinated action. By connecting security signals with application, infrastructure and user context, organizations can improve their ability to assess incidents and determine appropriate responses.
Shared visibility, he argued, is essential to turning threat detection into a coordinated response, particularly as AI-assisted attacks increase the pressure on response times.
AI Can Also Strengthen Cyber Defences
While AI is expanding the capabilities available to attackers, D’Cruz said organizations should also use the technology to strengthen their defensive operations. AI can help security teams investigate incidents faster and process information more efficiently, while human experts retain responsibility for critical decisions.
He also called on leadership teams to prepare for severe cybersecurity incidents before they occur.
“Every leadership team should rehearse its worst 72 hours before they happen,” D’Cruz said, emphasising the importance of preparedness and coordinated decision-making during a crisis.
As enterprises expand their use of AI and cloud-based applications, the security challenge will increasingly involve managing identities and credentials, securing software dependencies, monitoring system behaviour and ensuring that teams can respond collectively. D’Cruz’s recommendations point to a security approach built on disciplined access controls, continuous visibility and incident preparedness rather than reliance on additional tools alone.
