Aryaka predicts AI-led overhaul of SASE & enterprise security

Security vendor Aryaka expects artificial intelligence to reshape enterprise networks and cyber security strategies over the next two years, with senior executives predicting a wave of consolidation, new AI-specific controls and a shift in how chief information officers buy connectivity.

The company's leadership team outlined a series of forecasts for 2026 that centre on the convergence of networking, security and AI, and on the growing role of secure access service edge, or SASE, in managing that shift.

Shailesh Shukla, Chairman and Chief Executive, said technology leaders now face three overlapping priorities.

"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," said Shukla, Chairman and CEO, Aryaka.

He said the spread of AI tools and workloads is changing the threat landscape.

"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," said Shukla.

Shukla also expects a shift in how organisations select SASE products. He contrasted unified services with multi-vendor combinations.

"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," said Shukla.

He said internal resourcing pressures will push that change further.

"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," said Shukla.

Shukla predicted that many enterprises will step back from managing their own network connectivity and edge infrastructure.

"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," said Shukla.

He said visibility and elasticity will become key buying criteria.

"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," said Shukla.

AI firewall emerges

Srinivasa Addepalli, Chief Technology Officer, said AI-specific controls will move into the core of cloud-delivered security architectures.

"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," said Addepalli.

He said generative AI will also permeate enterprise software.

"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," said Addepalli.

Addepalli expects this shift to expose limits in traditional data loss prevention tools.

"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," said Addepalli.

He said this gap will trigger a redesign of controls.

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

Consolidation pressure

Renuka Nadkarni, Chief Product Officer, focused on economic pressure in the AI security market and on the structure of AI-related defences.

"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," said Nadkarni.

She said current pricing levels will not endure.

"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," said Nadkarni.

Nadkarni said AI adoption is expanding the attack surface across infrastructure, data flows and models.

"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," said Nadkarni.

She said this will favour integrated network and security architectures.

"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," said Nadkarni.

Rising AI risk

Aditya K. Sood, Vice President of Security Engineering and AI Strategy, warned of a sharp increase in both external and internal AI-related threats.

"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," said Sood.

He said internal practices and governance will also create exposures.

"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," said Sood.

The executives expect these shifts in architecture, spending and threat patterns to play out across global enterprises by the end of 2026.