2026 Industry predictions from senior executives

Shailesh Shukla, Chairman and CEO:

Networks, security and AI are set to converge

CIOs and CISOs are having to manage a three way challenge – networks, security and AI. Next year will see them seeking a way to achieve this through a more converged approach that is designed to handle all three.

AI is creating a broader attack surface that older converged solutions can’t handle
A new style of convergence, delivered in the form of a platform, will be the chosen way forward

Unified SASE to take over from ‘stitched together’ solutions

Many SASE choices out in the market are made up of various different products that have been combined together to create a single solution. But this can present problems and complexities. Over the next 12 months, IT chiefs will prioritise searching for a more unified alternative.

A product-based approach will always be more complex than a service-based one
Fewer and fewer CIOs have the resources and in-house expertise to maintain piecemeal approaches

CIOs will stop throwing money at managing their own connectivity

Maintaining a secure network perimeter is an ever worsening headache for too many IT chiefs, especially given today’s working patterns and heavily distributed employee base. Enterprises instead will be looking to consume integrated services with minimal operational overhead.

Need a single pane of glass to ensure visibility of risk and network performance
Will demand more elastic services that scale with ease and extend to anywhere

Srinivasa Addepalli, Chief Technology Officer:

“AI Firewall” Becomes a Built-In Component of SASE

By the end of 2026, the “AI Firewall” will transition from a standalone product category to a core SASE building block. It will become a native control within the SASE data plane, tightly integrated with SWG, ZTNA, CASB, DLP, and WAAP, and consistently enforced across POPs, branches, the edge, cloud environments, and Zero Trust WANs. The term “AI Firewall” will appear as commonly and as natively in SASE architecture diagrams as SWG or CASB do today.

Every Major Enterprise App Will Have an Embedded Copilot

By the end of 2026, every major enterprise application will ship with an embedded copilot by default. Microsoft Office, Teams, Outlook, Salesforce, ServiceNow, Workday, SAP, and Atlassian will all rely on native AI copilots that touch sensitive enterprise data across multiple surfaces. That data will increasingly flow through embedded AIs in SaaS platforms, chat-based copilots, search and summarization assistants, browser extensions, and custom internal LLMs. As a result, SASE will evolve into the primary enforcement layer that sits between users, SaaS platforms, and their embedded copilots.

Semantic DLP Becomes the Default Standard for Enterprise Data Protection

For nearly two decades, DLP has been built around patterns — regexes, keywords, dictionaries, file fingerprints, and endpoint rules. But in 2026, enterprises will confront a reality that makes these mechanisms insufficient: AI models understand semantics, while traditional DLP only understands strings.

This mismatch will force a foundational shift. 2026 will mark the arrival of Semantic DLP as the new default for enterprise protection.

Renuka Nadkarni, Chief Product Officer:

Economics will drive AI security consolidation

Economics, not algorithms, will drive the great AI consolidation of 2026 in cyber. Even the federal warning that ‘only a few will survive’ isn’t hyperbole—it reflects the strange reality of an enterprise market where startups want to charge three times more to secure AI than the cost of service itself.

That imbalance is simply unsustainable, and it will force the industry towards consolidation into simpler, integrated, cost-rational models. AI Security will be a feature of overall network security, at best a product add-on but not a standalone business.

No single security solution can address the new attack surface introduced by AI.

AI adoption is creating entirely new classes of attack surfaces—spanning underlying infrastructure, sensitive data pipelines, and the models themselves. Each layer is vulnerable in different ways and demands its own defensive techniques. In practice, AI is simply a new class of traffic, and securing it calls for the same foundational controls we apply to any critical workload: access enforcement, threat protection, data-loss prevention, and continuous monitoring.

No single point solution can span this entire landscape. But by treating AI as a new traffic category, unified SASE architectures can address a broad portion of these risks. SASE will play a central role in the future of AI security—delivering multi-layered, distributed protections embedded throughout the security stack, rather than isolated in a standalone tool. SASE plays a significant role in the future of AI security with multi-layered, distributed, and embedded across the entire security stack—not concentrated in one tool.

Aditya K. Sood, VP of Security Engineering and AI Strategy:

The AI Risk Curve: What’s About to Change

Enterprises are on the verge of confronting a surge in AI-driven internal and external threats. As autonomous agents become deeply embedded in operations, adversaries will exploit prompt interference, context poisoning, and insecure tool/API access to manipulate agents. This manipulation can lead to data leaks, configuration alterations, or control bypassing, introducing new systemic vulnerabilities that enterprises are currently unprepared to manage. With models becoming more affordable and capable, the threat of offensive AI is becoming persistent. Attackers will leverage large-scale synthetic identities, AI-crafted malware that adapts for each target, and autonomous phishing ecosystems that monitor user behavior in real-time to optimize delivery.

Simultaneously, the misuse of enterprise AI and governance gaps will create their own set of risks. Employees may inadvertently leak sensitive data into unauthorized AI tools, rely on unverified outputs, or automate workflows without oversight—creating blind spots that attackers can exploit. The emergence of shadow AI, hallucinated security recommendations, and unchecked agent autonomy will compound operational risk. To stay ahead, organizations will need to deploy robust AI security solutions. These solutions will be crucial in transforming AI from an unpredictable liability into a controlled and secure asset.