The model is not the moat

The question I hear most from investors has changed. A year ago, it was simple: “Do you have AI?” Today it’s harder and more interesting: “Where is AI actually driving resolution in production?” That shift is the most important signal in enterprise tech right now. It means the hype cycle is giving way to a serious reckoning – and organisations are finding that capability and value aren’t the same thing.

 

Caught between two inadequate options

 

Mid-market and enterprise organisations today face a frustrating choice. On one side sit legacy platforms carrying years of customisation, cost and operational complexity. On the other, a wave of AI point solutions that perform well in controlled demonstrations but lack the data depth, contextual awareness, governance controls and operational integration required to function reliably in production. Neither is a sustainable answer. Yet many organisations are being sold one or the other.

 

Context is the differentiator

 

What determines whether AI delivers value isn’t the model itself. Models are commoditising fast and any advantage built on model selection alone will erode. The real moat is context: the accumulated operational knowledge, service history, asset relationships and workflow logic that sits beneath the model and tells it what to actually do.

 

For AI to resolve issues – not merely summarise them – it needs to understand the specifics of a business: its assets, service history, change approvals and the interdependencies between systems. Without that grounding, even the most technically sophisticated model is essentially guessing.

 

This is the design principle behind Freshworks. Rather than layering AI on top of an existing system, we’ve embedded it directly inside the operational environment itself – one built on over a decade of workflows and real-world actions across 75,000 customers. The result is a system of context, not just a system of record. It manages the interactions between people, software and infrastructure, and in doing so creates the foundation AI needs to actually resolve problems rather than generate plausible-sounding responses.

 

Where the momentum is clearest

 

The organisations benefiting most from this approach are mid-market and lower-end enterprise companies – and that’s not coincidental. These businesses – companies with between 500-20,000 employees such as New Balance, TaylorMade, Vermeer and Nucor – operate in genuinely complex environments but don’t have the luxury of large internal teams or the budget for consultant-heavy implementations.

 

Vermeer, a manufacturer with nearly a century of operational history, is a representative example. By automating IT operations rather than bolting on a separate AI layer, it freed capacity to invest in growth instead of maintaining infrastructure. That principle scales. With AI operations within the actions and workflows already embedded in the system, rather than wrapping around them from the outside, accuracy and resolution rates are meaningfully higher than those produced by wrapper solutions. The architecture doesn’t guess. It knows.

 

A word on the SaaS debate

 

There’s been a real debate in recent weeks about whether enterprise software is under existential threat from AI. I’ll give you my honest read: I don’t think enterprise software will disappear, but I do think a lot of it deserves to. The platforms that are too expensive to maintain, too slow to deliver value and too dependent on armies of consultants to function are exactly what AI should replace. The question isn’t whether software survives. It’s whether your software earns its place.

 

As CIOs plan their next wave of AI investment, the winners won’t be determined by which vendors have the most-discussed models. They will be determined by which companies help IT leaders deal with the operational problems they actually face every day: compounding complexity, slow time to value, and the mounting cost of maintaining systems that were never designed to work together.

 

The genuine measure of AI’s value in enterprise is not the sophistication of the model. It is the ability to automate low-value operational work and give skilled IT teams the space to focus on problems that actually need them. That is the standard against which enterprise AI should now be judged.

---

To learn more about how Freshworks is building AI into the operational layer of enterprise IT, visit freshworks.com

---

Dennis Woodside, CEO, Freshworks