Frontier AI companies are just discovering what services firms have known all along

By Nitesh Bansal, Managing Director & CEO, R Systems International

Over the past few weeks, there has been significant conversation around OpenAI and Anthropic expanding aggressively into enterprise AI deployment and services, alongside broader concerns that frontier AI firms could begin eating into traditional IT services revenue models.

But in many ways, these developments reinforce a reality the industry has known for years: enterprise AI transformation cannot succeed through models and tools alone. Real enterprise value still depends on implementation, workflow integration, governance, data readiness, change management, and sustained operational oversight.

The emergence of deployment and services arms from leading AI firms reflects the growing complexity of enterprise AI adoption itself. While model capabilities are advancing rapidly, enterprises are still struggling to scale AI beyond pilots, integrate it into legacy environments, establish governance structures, and align deployments to measurable business outcomes.

In fact, OpenAI and Anthropic together reportedly committed nearly $5.5 billion toward deployment-led enterprise AI initiatives in May 2026 alone, signalling how central human-led implementation and operationalisation have become to enterprise AI adoption.

Historically, even the world’s largest enterprise software platforms have relied heavily on implementation ecosystems to drive adoption and scale. Platform companies build the core technology, but the larger burden of deployment, integration, transformation, and workflow redesign has traditionally been carried by services and consulting ecosystems.
Salesforce and Adobe, for instance, generate only a relatively small share of revenue from professional services, with broader partner ecosystems driving most enterprise deployment and transformation work.

Salesforce derives roughly 6% of its revenue from professional services, while Adobe’s services contribution stands at about 3.5% – reinforcing how enterprise technology adoption has historically depended on implementation ecosystems beyond the platform providers themselves.

Enterprise AI adoption is increasingly becoming less of a model-access challenge and more of an operational transformation challenge. Enterprise environments are largely brownfield in nature that are built on years of legacy infrastructure, fragmented systems, customised workflows, and operational complexity that cannot be transformed through plug-and-play AI deployments alone.

This is also where the industry conversation is beginning to shift. Instead of focusing purely on whether AI can automate services work, enterprises are increasingly evaluating: How AI systems can be operationalised responsibly? How governance and human oversight are maintained in production environments? How multiple models and AI tools are integrated into enterprise architectures? How workflows need to be redesigned to derive measurable business value?How organisations manage change, accountability, and risk in AI-led operations?

The emergence of deployment-focused services arms from AI companies simply reinforce the growing recognition that enterprise AI adoption requires embedded expertise and operational context and not just access to advanced models.

The long-term opportunity will ultimately favour firms that can combine technology expertise with operational and contextual understanding, governance depth, and vendor-independent integration capabilities. As AI adoption matures, competitive differentiation is likely to depend less on access to frontier models alone and more on the ability to embed AI effectively within complex enterprise ecosystems and deliver measurable outcomes at scale.