Deployment of an AI agent to complete a business task involves selecting a model, accessing data through an API and calling an internal tool, all within the same workflow. According to Dhananjay Ganjoo, Managing Director, F5, India & SAARC, this creates three immediate considerations for the enterprise. In an exclusive interaction with Rajneesh De, Group Editor, APAC Media & CXO Media, Ganjoo outlines how F5 AI Gateway addresses these three enterprise priorities – controlling the economics of AI, governing how agents interact with models, data and tools, and securing those interactions.
Under the F5 Distributed Cloud Services and F5 AI Data Fabric, what AI portfolio solutions and services are available in India?
Indian enterprises have full, localized access to F5’s complete Distributed Cloud Services platform, which runs on a highly connected global network anchored by local infrastructure in hosted across India & SAARC region. The platform is supported by the F5 AI Data Fabric an advanced data engine launched to accelerate telemetry-driven insights and train security models.
Because of this robust local footprint, organizations in India can deploy F5’s AI-powered Web Application and API Protection (WAAP) which integrates WAF, machine learning-based Bot Defense, and API security as well as multi-cloud networking and edge compute. This localized approach ensures that sensitive data traffic is processed and contained strictly within India’s borders, giving businesses the agility of the cloud without compromising regional regulatory compliance.
What will be the key functions of the F5 agentic AI gateway?
An enterprise AI agent completing a business task may need to select a model, access data through an API and call an internal tool, all within the same workflow. That creates three immediate considerations for the enterprise, which model is being used and at what cost, what data and tools the agent is authorised to access, and how sensitive information is protected throughout the interaction.
At its core, F5 AI Gateway addresses these three enterprise priorities – controlling the economics of AI, governing how agents interact with models, data and tools, and securing those interactions. The Model Gateway provides visibility into token consumption by model, provider, team and user, while using budgets, intelligent routing and semantic caching to manage costs. The MCP Gateway helps govern agent-to-tool interactions through fine-grained access policies and provides an audit trail of what an agent accessed and on whose behalf. AI Guardrails help secure prompts and responses against risks such as prompt injection and jailbreak attempts and while preventing sensitive data from being exposed through AI interactions.
The objective is to give enterprises one control point for the security, governance and economics of AI, rather than managing each through separate tools.
How are F5’s WAAP services geared to gather and analyse real-time data on emerging threats?
The challenge today is that emerging threats do not always arrive with a known signature. Attackers continuously change their behaviour, while applications and APIs are generating enormous volumes of traffic that security teams cannot assess manually.
F5’s WAAP approach is built around understanding what is happening across application and API traffic in real time, rather than relying only on known signatures. Distributed Cloud WAF combines signature-based detection with behavioural analysis and AI-powered risk scoring to correlate multiple signals and assess the risk of each request. This helps identify suspicious behaviour even when an interaction does not match a known attack pattern.
That visibility also extends to APIs, where continuous machine learning can establish behavioural baselines and flag unusual or malicious activity. F5’s threat intelligence and research help strengthen detection of emerging attack patterns, while the platform can apply policy and enforcement controls as those risks are identified.
The goal is to move from collecting more security data to using those signals to make faster, more informed decisions about which traffic should be allowed, challenged or blocked.
What are the verticals where F5 is witnessing maximum traction and what are the prevalent use cases?
The strongest demand tends to come from sectors where digital applications are central to the business and there is very little tolerance for downtime, fraud or security gaps. In India, this is particularly relevant across financial services, government and public services, technology and telecom.
The use cases vary by sector. In financial services, the focus is on protecting APIs and digital transactions, alongside bot and fraud prevention. Government and other regulated environments place greater emphasis on application availability, security and control across distributed infrastructure. For technology and telecom organisations, AI is creating newer requirements around inference traffic and infrastructure efficiency, alongside the established focus on application delivery, performance, and resilience.
AI is now adding another dimension. As organisations introduce AI into applications and workflows, the conversation is expanding to securing AI applications and APIs, governing access to models and agents, and managing inference efficiently. What we are seeing is that these are no longer separate infrastructure, security and AI conversations. Enterprises increasingly need to address them together. The common requirement remains straightforward: applications need to be available, secure and performant wherever they run.
What challenges are enterprise CISOs facing today regarding information overload and siloed, distributed data?
Enterprise CISOs today are dealing with complexity at an unprecedented scale. Applications, APIs and increasingly AI workloads are distributed across data centres, clouds and edge environments, each generating its own security signals. The challenge is no longer a lack of data; it is making sense of that data quickly enough to act.
For CISOs, complexity itself is increasingly becoming a source of risk, because every additional silo can slow detection, decision-making and response.
As AI adoption accelerates, this becomes even harder. F5 research shows that approximately 78% of organisations are already operating AI inference themselves, with an average of seven models in production.
This is where we see the security model evolving from managing multiple disconnected tools toward unified visibility and consistent controls across applications, APIs and AI workloads, so security teams can identify and respond to risks faster. Ultimately, the objective is not simply to give CISOs more information, but to give them the context they need to make faster and better security decisions.
How can F5 help enterprises control AI costs?
AI costs need to be managed while inference is happening, not after the bill arrives. F5 helps enterprises do this by giving them visibility and control over token consumption across providers, models, teams and users. Organisations can set budgets and monitor usage as it occurs, rather than discovering cost overruns later.
The next step is to reduce unnecessary consumption. With capabilities such as semantic caching, intelligent routing and model tiering, enterprises can avoid unnecessary model calls and direct workloads to the most appropriate model rather than sending every request to the most expensive option. F5 AI Gateway is designed to reduce token spend by up to 60% without application changes.
For enterprises running their own AI infrastructure, F5 also focuses on GPU efficiency, helping organisations get more useful output from existing capacity and lower the cost per token. This is why we see the conversation moving towards token economics. The real question for enterprises is not simply how much AI infrastructure costs, but how much useful AI output they can generate from every unit of compute they have already invested in.
What are the key pillars on which F5 GTM strategy rests in India? What are some of the key initiatives under this strategy?
India is one of the fastest-growing market for F5 in APCJ, but the opportunity is not simply about selling more technology. Customers are rethinking how they deliver and secure applications as workloads move across data centres, private clouds and public clouds, and AI is adding another layer to that complexity.
Our approach in India therefore rests on a few interconnected priorities: deepening our strategic customer relationships, capturing the infrastructure and AI refresh opportunity, expanding the conversation around application and AI security, and scaling these capabilities through our partner ecosystem
For us, this creates a natural opportunity to broaden the conversation with our existing customers. As infrastructure comes up for refresh, the discussion is increasingly about how application delivery, security and AI can work together rather than adding another point solution. With AI, that conversation is also moving towards economics: how much inference is costing, whether the right model is being used for the right workload, and how effectively compute capacity is being utilised.
Partners are critical to making that work. The opportunity now requires deeper architecture, AI delivery and runtime security skills, so partner enablement and specialisation will remain an important part of how we scale in India. The opportunity for F5 is therefore not only to participate in customers’ technology refresh cycles, but to become a strategic partner as they rethink how applications and AI workloads are delivered, secured and operated.
What is the overall channel structure that F5 currently follows in India? How are these differentiated between distributors, MSPs, ISVs and SIs?
More than 90% of F5’s global business is driven through channel partners, which reflects how important the ecosystem is to how we engage with customers. In India, we work across distributors, systems integrators and a broader ecosystem of technology and services partners, with each bringing different capabilities across reach, customer engagement, architecture, integration and services.
Broadly, distributors help us scale market reach and partner enablement, while systems integrators bring architecture, implementation and transformation capabilities into complex customer environments. MSPs can play an important role in taking these capabilities into managed services, while ISVs and technology partners can extend the broader ecosystem through integrations and complementary capabilities.
What is changing is the role partners play as customers modernise applications and adopt AI. The conversation is increasingly moving beyond deployment towards helping customers make decisions around architecture, AI readiness, application and API security, and hybrid multicloud environments. Through initiatives such as F5 Unity+ and the ADSP Partner Program, we are investing in enablement and technical expertise to help partners build these capabilities and create greater value for customers.
Ultimately, we see the role of the channel moving from technology fulfilment towards strategic advisory and services as customers navigate application modernisation and AI.
How does F5 today differentiate itself in the highly competitive cybersecurity solution provider landscape?
The cybersecurity market has no shortage of point solutions. The challenge for customers is managing them consistently as applications span data centres, clouds, APIs and now AI. F5 has spent three decades delivering and securing applications, and that gives us a strong foundation as this complexity increases.
F5 brings application delivery and security together through ADSP, allowing customers to apply consistent visibility and controls across hybrid multicloud environments while maintaining performance and availability. As threats accelerate, the ability to enforce those controls in the traffic path becomes increasingly important.
AI makes that proposition even more relevant because enterprises increasingly need to solve three problems together: how AI traffic is delivered, how models, agents and APIs are secured at runtime, and how inference is operated efficiently. We are bringing discovery, testing and runtime protection across models, agents and APIs into the platform. The conversation is also moving towards token economics, because customers will increasingly need to manage the security, performance and economics of inference together.
That is also why token economics is becoming part of the conversation alongside security and performance. Increasingly, customers will need to manage the security, performance and economics of inference together. For F5, these are not separate problems; they are becoming different dimensions of how modern applications and AI workloads are delivered and secured.
