AI-Driven Cybersecurity Outcomes: An Interview with Palo Alto Networks’ Tom Scully and Gopesh Maindola

Artificial intelligence (AI) is reshaping cybersecurity at a pace unmatched by previous technology shifts. As organizations across Asia-Pacific accelerate AI adoption-from copilots to domain-specific models-security teams are grappling with an explosion of alerts, rising attacker sophistication, and expanding digital risk.

In Frost & Sullivan's ongoing AI Transformation series, we have explored why AI is a business imperative, how enterprises can assess technical readiness, and why secure-by-design thinking is the foundation for sustainable adoption. This next phase focuses on how technology providers are evolving their solutions to help organizations defend themselves in an AI-driven world.

For this edition, we spoke with Tom Scully and Gopesh Maindola from Palo Alto Networks, who shared how they leverage AI for better security outcomes-and how enterprises can secure their AI initiatives.

Their perspectives reinforce a clear message: AI is no longer optional in cybersecurity. It is fundamental to achieving visibility, efficacy, and speed at modern scale.

How Palo Alto Networks Uses AI in Its Security Solutions

Palo Alto Networks has a long history of embedding AI and machine learning (ML) into its platform-from malware analysis to anomaly detection to behavioral analytics. These models have matured over years of threat intelligence collection and iterative improvement.

Tom Scully explains:

"AI has always been about increasing efficacy. With years of machine learning experience, we're now layering in more advanced models that help analysts make better decisions, faster."

In recent years, Palo Alto Networks has expanded this foundation by introducing AI copilots purpose-built for security operations. These copilots:

• Summarize alerts
• Provide root-cause explanations
• Reduce time spent on low-level tasks
• Consolidate visibility across complex environments
• Help analysts navigate overwhelming workloads

Frost & Sullivan's analysis shows that AI copilots are emerging as an impactful addition to SOC operations, especially as organizations face talent shortages and increasing complexity.

Why AI Matters Now More Than Ever

The pace of AI adoption across enterprises is accelerating-and so are the associated risks.

According to Scully, customers are now asking more urgent questions:

• How do we innovate responsibly with AI?
• How do we govern access to AI tools?
• How do we prevent shadow AI?
• How do we protect sensitive data when interacting with models?

This shift is driven by several factors:

• Pressure to adopt: Business leaders see AI's potential to transform operations and customer experience.
• Shadow AI emergence: Employees are using unapproved tools without security oversight.
• Digital risk expansion: AI models introduce new avenues for data leakage or misuse.
• Skills gaps: Security teams are still building expertise to evaluate and secure LLM pipelines.

Frost & Sullivan analysis confirms that organizations that have adopted AI face an expanded threat surfaces, making robust governance and controls essential from day one.

Advancing Security Outcomes Through AI-Driven Capabilities

Palo Alto Networks' approach reflects how the broader cybersecurity industry is evolving: leveraging AI not only to detect more threats, but to help analysts manage complexity and secure emerging AI-driven business environments.

Across the market, three categories of capabilities are proving essential for organizations seeking stronger outcomes:

1. AI for Threat Detection and Prevention

Years of curated telemetry enable AI models to:

• Classify threats with higher accuracy
• Detect zero-day attacks more effectively
• Automate prevention at machine speed
• Reduce false positives and alert noise

These capabilities help SOC teams respond faster and focus on the most critical risks.

1. AI for Analyst Productivity

AI copilots and natural-language interfaces are becoming foundational tools that:

• Accelerate investigations
• Automate repetitive tasks
• Summarize complex alerts
• Correlate signals across multiple systems

This supports analysts at all levels, helping overstretched teams do more with limited resources.

1. Security for AI Transformation

As organizations adopt AI internally, cybersecurity vendors are increasingly supporting:

• Governance frameworks and usage visibility
• Guardrails to prevent shadow AI
• Data loss prevention (DLP) for AI interactions
• Zero Trust controls around model access and sensitive data

These capabilities help enterprises build confidence in their AI systems and secure them responsibly.

Have AI Investments Improved Security Outcomes?

According to Maindola, the impact is clear:

• Higher detection accuracy
• Reduced investigation times
• More automation across workflows
• Greater productivity for analysts
• Better visibility across fragmented systems

These improvements align closely with Frost & Sullivan's industry benchmarks, which show sustained gains in MTTR, false positive reduction, and proactive threat discovery with AI-enabled systems.

The Future: Toward Agentic AI in Cybersecurity

The longer-term future points to agentic AI-systems capable of taking autonomous actions, coordinating across tools, and executing security tasks end-to-end.

Scully notes:

"Agentic AI gives AI arms and legs. But humans must stay in the loop. Authority needs oversight. Organizations must build foundations and guardrails before moving toward highly autonomous systems."

Key challenges include:

• Independent assessments
• Regulatory alignment
• Questions of liability
• Trust in automated decisions

Frost & Sullivan expects agentic AI to emerge gradually-first as assistive capabilities, then evolving into semi-autonomous operations under human supervision.

Best Practices: Securing AI Transformation in Your Organization

Drawing from Frost & Sullivan insights and Palo Alto Networks perspectives, we recommend the following best-practice principles:

1. Build AI governance before scaling AI usage

Define policies for acceptable use, prompt guidance, model access, and data classification.

1. Enforce Zero Trust for all AI interactions

Treat AI systems as high-value assets requiring strict access controls, continuous monitoring, and segmentation.

1. Deploy guardrails and DLP to prevent data leakage

Monitor both the inputs and outputs of AI tools to ensure sensitive information is not exposed.

1. Gain visibility into shadow AI

Inventory how employees are using AI tools-approved or otherwise-to identify unmanaged risk.

1. Keep humans in the loop-even as AI becomes more autonomous

Human oversight remains essential for ethical, accountable, and compliant decision-making.

Conclusion

Palo Alto Networks' insights highlight a broader industry reality: AI is now essential for both achieving stronger cybersecurity and securing AI-driven business transformation.

Organizations that invest in governance, enforce responsible controls, and leverage AI to enhance capabilities-not replace them-will be best positioned for the next era of digital risk, where attackers and defenders increasingly operate at AI speed.

About Kenny Yeo

Kenny Yeo currently leads Frost & Sullivan's cyber security practice across Asia Pacific. A current topic of interest is analysing how vital cyber security is today to enterprise digital transformation efforts to achieve secure DX outcomes. With 20 years of research, consulting, advisory, team management and business development experience, Kenny has expertise spanning cyber security, IoT, smart retail, industrial and e-government.

View all posts by Kenny Yeo

Kenny Yeo

Kenny Yeo currently leads Frost & Sullivan's cyber security practice across Asia Pacific. A current topic of interest is analysing how vital cyber security is today to enterprise digital transformation efforts to achieve secure DX outcomes. With 20 years of research, consulting, advisory, team management and business development experience, Kenny has expertise spanning cyber security, IoT, smart retail, industrial and e-government.