Is SaaS Becoming an OS?

In this interview, Sreedhar Gade, VP of Engineering at Freshworks, explains this transition, reveals how Generative AI is changing product strategy, and moving to AI-native systems and conversation interfaces. He also discusses the consolidation trend with respect to enterprise SaaS, Freshworks journey to a combined platform, and his vision for an “agentic” world that emphasizes AI governance and trust.

How is GenAI changing the way Freshworks approaches product strategy today?

What we can uniquely do is serve both SMBs and enterprises, thus giving us a larger innovation canvas. In AI, we’re pulling together new forms of co-piloting, where AI helps agents by responding using the full context of the customer, with models we trained on Freshworks data to further enhance automation, and even agentic AI, where customers would work via AI agents for full resolution. Our goal is to move beyond automation to create truly intelligent self-learning systems.

In your view, is enterprise SaaS moving toward consolidation or modular expansion?

It depends on the maturity of the company. Startups tend to want best-of-breed tools that can do one thing for a use case. As businesses grow, typically they prefer unified platforms. The driver of this change is AI, and AI’s ability to accelerate product development as we speak. If you think about innovations at market level, the real differentiator is ecosystem. Enterprises are drawn to platforms that unify data, integrate with other functions seamlessly and deliver intelligence across the requisite functions. For sure, consolidation will be the greater trend at enterprise level.

How do you see this evolving? Are we heading toward an AI-native enterprise SaaS era?

We’re shifting from traditional software to intelligent systems. Until now, most SaaS products incorporated AI incrementally, tacked on to existing products. The future is truly AI-native, where conversational interfaces will be the primary way to interact regardless of the product segment, allowing users to ask and get information through natural dialogue.

We’re combining AI co-pilots with in-house models to boost automation and exploring agentic AI for end-to-end resolution—aiming to build intelligent, self-learning customer support systems.

Intelligence will also change as to where it resides. Previously, intelligence was built into the actual software—in the code itself. Now, intelligence is being offloaded to hyperscalers like OpenAI or Meta, where the core data continues to be owned by SaaS providers. Thus, we are beginning the hybrid era of SaaS, where SaaS companies represent the system of record, while intelligence becomes distributed and dynamic.

And what about verticalization? Will industry-specific SaaS dominate, or will horizontal platforms continue to lead?

In the enterprise play, your playbook is increasingly focused on platformization. Vertical niche solutions will have their use cases, but the value is in the construction of platforms that have flexibility. Platforms that support multiple use cases and allow deep integration into other ecosystems.

For example, and to build on our ESM (Employee Service Management) focus, customers can link their internal data sources—OneDrive, Google Drive, SharePoint—directly into Autopilot, and Freshservice indexes and pulls the intelligence across those data baskets, delivering an intelligent support experience. That is the beauty of a platform-led strategy—the data remains where it lives, but intelligence can be unlocked everywhere.

Freshworks started as a modular suite and is now evolving into a unified platform experience. How do you manage the tension between offering seamless integration across your suite and maintaining API-first flexibility for enterprises with diverse tech stacks?

That’s a great question, and it’s something we’ve been very intentional about. While we began with standalone products—starting with Freshdesk and then expanding across ITSM, CRM, and more—we knew that to deliver true value at scale, we had to rebuild our platform from the ground up.

Over the past few years, our engineering teams have rearchitected the backend to ensure interoperability between all modules. Whether it’s IT service management or customer support, all products now share a common data model, event framework, and intelligence layer. This allows them to seamlessly exchange information, respond to triggers, and provide a unified user experience.

From a customer’s perspective, moving from, say, ITSM to ESM feels natural—it’s the same interface, same design system, and even shared integrations. You might, for example, discover a SharePoint integration appearing consistently across modules, without needing to configure it multiple times. The short answer: We solved the integration challenge by rethinking the platform architecture itself to unify modular products without losing flexibility.

As you move from mid-market to enterprise, do your customers lean toward a single-vendor model, or are they still asking for openness and best-of-breed integration?

Honestly, it’s a mixture of both. Larger enterprises often have legacy systems or best-of-breed solutions already in place, so they definitely value API-first flexibility. That said, many are also looking to reduce vendor sprawl and consolidate around platforms that can do more, especially if it offers consistent UX and deeper integration out of the box.

Internally, we’ve also adopted this mindset. Whether we’re building tools in-house or integrating third-party solutions, our teams now design everything with an API-first approach. It’s not about either/or anymore—it’s about ensuring flexibility within a unified framework.

Do you think the future of SaaS will look more like an enterprise operating system or a loosely connected network of AI-powered services?

I’d lean toward the latter. The underlying operating system, if you will, will likely be powered by large language models or agentic frameworks. We’re already seeing examples—Amazon’s recently launched Paragraph Core, Microsoft’s Fabric, and OpenAI’s emerging stack are all pointing toward this trend.

On top of these foundational layers, we’ll likely see a modular ecosystem of services—each AI-powered, each specialised, but all loosely connected within a shared infrastructure. It won’t be monolithic. It’ll be intelligent, interoperable, and purpose-built.

Startups prefer best-of-breed tools, but as they scale, unified platforms become key. With AI speeding development, enterprises now prioritize ecosystems that unify data, integrate seamlessly, and deliver cross-functional intelligence.

What, in your view, will differentiate future-ready SaaS companies from legacy players over the next 3–5 years?

In the early days, the product was the IP. Building software used to take months or years. But now, with tools like Cursor or emerging platforms like Tiro.dev, you can spin up working software in hours.

The real differentiator is no longer the product—it’s the system of record you establish and the ecosystem you build around it. For instance, in our case, if a customer uses five different internal systems—HR, finance, payroll, etc.—we want Freshservice to offer a unified layer that interacts across all of them. The employee should be able to initiate requests or access services through Slack or Teams without worrying about what’s happening behind the scenes.

That’s where the next generation of SaaS will win: By delivering integrated, intelligent, and invisible experiences powered by modular systems but presented as one cohesive platform.

Are you seeing renewed interest in SaaS solutions from regulated sectors like finance or healthcare?

Yes, absolutely. Traditionally, sectors like healthcare have been more conservative in adopting SaaS, especially compared to industries like HR tech or customer support. But that’s changing fast. We’re seeing growing interest not only from healthcare but also from pharma, automotive, and even agriculture. Some of these sectors are experimenting with plain generative AI use cases, while others are exploring robotic process automation integrated with AI.

To support this momentum, we’re working on building industry-specific AI models. While general-purpose models like GPT-4 can handle a wide range of tasks, verticalised models will have domain-specific knowledge—like understanding CT or MRI scans, parsing prescriptions, and correlating drug data. That’s the direction we see the industry heading.

What will the next generation of enterprise SaaS look like in the next three to five years?

Three to five years is a long horizon in AI terms, but even in the next two years, we’ll see massive transformation. The future is agentic. Multi-agent orchestration will become the norm—where intelligent agents can independently perform complex workflows and collaborate with other agents.

Think of booking an entire holiday package—flights, hotels, airport transfers. Instead of a human coordinating each step, different agents will interact and complete these tasks autonomously. And it won’t stop there. In sectors like healthcare, agents could assist in managing patient care or treatment plans.

Another critical area will be AI governance. Just as SaaS evolved to include robust cybersecurity, AI must now embed trust, explainability, and safety. Governance will be key to ensuring responsible AI adoption, especially for enterprise use.

On that note, how does Freshworks ensure that its generative AI is explainable, secure, and enterprise-ready?

At Freshworks, we’ve built a dedicated framework called PRETI AI to address AI safety, governance, and enterprise readiness. It stands on five key pillars:

Safety – We rigorously scan for any bias, toxicity, or harmful patterns in our AI systems to ensure responsible outcomes.

Privacy – Personally identifiable information (PII) is protected at all levels. We mask or anonymize sensitive data like credit card details or names to ensure nothing leaves the network improperly.

Role-Based Access Control (RBAC) – Only authorized personnel can access specific data or AI outputs, enforcing data security.

Traceability – Every AI-generated action is traceable. If something goes wrong, we can review the entire decision-making path and correct it.

Security – End-to-end encryption ensures the integrity of all transactions and data exchanges.

We have a dedicated team and robust tooling in place, and we work closely with partners like Microsoft and AWS to evolve these safeguards continuously.

aanchalg@cybermedia.co.in