Our industry has become obsessed with building ever more capable AI models, while enterprises have become equally obsessed with gaining access to them.
Every week brings another benchmark, another frontier model, and another announcement that promises to reshape the AI landscape. Just last week, Moonshot AI’s Kimi 3 was the latest reminder that intelligence at the model layer is becoming more capable, more affordable, and increasingly interchangeable.
That’s great news for enterprises, but it also raises a more important question: If every enterprise rents increasingly powerful intelligence from the same handful of providers, from where will lasting competitive advantage actually come?
The answer is not yet another foundation model, but building intelligence that is unique to the enterprise.
Every organization possesses decades of accumulated expertise that no public model can replicate. It exists in the judgment of experienced employees, the way critical decisions are made, the nuances of customer relationships, industry-specific operating practices, and the battle scars and countless lessons learned through years of execution. That knowledge has always been the real differentiator, yet most organizations have never treated it as a strategic asset. Instead, it remains scattered across people, processes, applications, and documents, becoming increasingly difficult to capture, improve, and retain. As we recently researched, there is $18 trillion in trapped value just wasting away in our Global 2000 organizations. Are these really enterprise debts, or is this the value that needs to be unleashed that keeps them unique?
Services-as-Software™ defined: transforming human expertise into enterprise intelligence
This is precisely why HFS has evolved the concept of Services-as-Software™ (SaS). What began as a way for services and software providers to transform expertise into scalable digital capabilities has become something much broader:
Services-as-Software™ is the HFS operating framework that enables enterprises to build Sovereign Enterprise Intelligence by capturing and codifying human expertise, then continuously improving it through execution.
Net-net, SaS combines AI, business context, enterprise data, and governance to create continuously learning digital capabilities that remain owned by the enterprise rather than becoming part of someone else’s intelligence.
Three principles underpin the SaS approach
Capture and codify human expertise. Organizations must transform human expertise into reusable digital capabilities rather than allowing critical knowledge to remain trapped within individuals, documents, or consulting engagements.
Retain sovereignty over enterprise intelligence. AI should be informed by enterprise context without enterprises surrendering the knowledge, operating logic, and business expertise that differentiate them. Enterprise intelligence must remain an enterprise asset, not become part of someone else’s competitive advantage.
Continuously learn from execution. Every workflow, customer interaction, and business outcome should strengthen the enterprise itself. SaS creates a continuous learning cycle in which execution improves the operating model, enriches enterprise intelligence, and elevates the skills and judgment of the people who work within it.
Traditional software automates transactions, and traditional services apply human expertise. Services-as-Software™ continuously transforms that expertise into enterprise intelligence that grows increasingly valuable through execution:
Building Enterprise Intelligence requires a new enterprise operating model
Our Services-as-Software™ Enterprise Intelligence Model connects leadership intent with business execution through six tightly integrated layers, all guided by a clear strategic objective. Every organization begins with an outcome it is trying to achieve, whether improving customer experience, accelerating product innovation, reducing risk, or transforming operational performance. That strategic intent becomes the organizing principle for every layer that follows.
Compute provides the infrastructure that powers enterprise AI, while foundation models contribute increasingly interchangeable reasoning capabilities. Agent orchestration coordinates models, workflows, and governance across complex business processes, ensuring intelligence is applied consistently rather than in isolated pockets across the enterprise.
At the heart of the model sits OneOffice Execution, where AI, people, enterprise data, and business operations converge into a single execution environment. This is where enterprise context is embedded in every decision, enabling AI to understand not only language but also the organization’s operating policies, customer relationships, regulatory obligations, commercial priorities, and institutional knowledge. Rather than automating individual tasks, OneOffice continuously aligns strategic intent with operational delivery while transforming every interaction into enterprise intelligence that remains under enterprise ownership.
Surrounding this execution layer is Governance and Enterprise Intelligence, which protects, validates, and continuously enriches the knowledge that differentiates the organization. Governance ensures AI operates securely, ethically, and in compliance with enterprise policies, while enterprise intelligence captures the insights generated during execution keeping them owned by the business, rather than becoming fragmented across applications or external AI platforms.
The Activation Layer then turns intelligence into organizational change. The activation layer bridges the gap between experimentation and production by combining Forward Deployed Engineers, Global Capability Centers, business leaders and services partners to industrialize AI across the enterprise. It is where isolated pilots become repeatable operating capability. This is where leaders, employees, and business functions adopt new ways of working, redesign processes, develop new skills, and embed AI into everyday decision-making. Without activation, even the most sophisticated AI operating model remains a technical capability rather than a business transformation.
Finally, Business Outcomes complete the learning cycle. Every customer interaction, operational workflow and business result feeds new knowledge back into the operating model, refining enterprise context, strengthening enterprise intelligence and continuously improving future execution.
Unlike traditional technology stacks, the Services-as-Software™ Enterprise Intelligence Model is designed as a continuous learning system. Strategic intent shapes execution, execution generates enterprise intelligence, activation embeds new capabilities across the organization, and business outcomes continually strengthen every layer of the model. The objective is not simply to automate work, but to create an enterprise that continuously learns, continuously adapts and continuously builds intelligence that competitors cannot rent, replicate or buy.
A smarter insurer illustrates how the model works
Consider a global insurance company seeking to reduce claims costs while improving customer experience.
Leadership begins with a clear business objective: settle claims faster, reduce fraud and improve customer trust. Compute provides the infrastructure to run AI at enterprise scale, while foundation models contribute reasoning capabilities for analyzing documentation, summarizing conversations, and supporting decision making. Agent orchestration coordinates specialized AI agents responsible for fraud detection, policy validation, regulatory compliance, and customer communications.
The real transformation occurs inside OneOffice Execution, where those agents operate with full enterprise context. Policy rules, customer history, underwriting practices, regulatory obligations and decades of claims expertise are combined into a single operating model that allows AI to make decisions consistent with how the insurer actually runs its business rather than relying solely on generic model knowledge.
Governance ensures every decision is explainable, compliant and aligned with enterprise policies, while every claim outcome feeds back into the operating model. Fraud patterns become easier to identify, customer interactions become more personalized, and underwriting decisions become more accurate because the insurer is continuously building enterprise intelligence rather than simply processing transactions more efficiently.
The technology improves, but more importantly, the business becomes smarter.
The real debate isn’t open versus closed AI… it’s rented versus owned intelligence.
Alex Karp recently ignited debate by warning that enterprises risk giving away the intelligence that differentiates their business as they increasingly depend on frontier AI providers such as OpenAI and Anthropic. His delivery was characteristically provocative, but the strategic issue deserves far more attention than the headlines it generated.
Every prompt, workflow and business process submitted to an external AI platform raises an important question. Is the enterprise simply consuming intelligence, or is it helping create intelligence that eventually benefits someone else?
This is why the growing focus on open models matters. The debate is not ideological. It is strategic. Open models give enterprises greater control over where inference runs, how models are tuned and how proprietary knowledge is protected. They reduce dependence on any single model provider while allowing organizations to combine multiple models within the same operating environment.
This is also why the Palantir-NVIDIA partnership is more significant than many observers realize. The announcement was never simply about combining GPUs with software. It reflects a broader shift toward enterprise-controlled AI environments where compute, models, orchestration and governance operate together while enterprise intelligence remains under the organisation’s ownership rather than becoming another external dependency.
As foundation models become increasingly interchangeable, the competitive advantage will no longer come from choosing the smartest AI model, but from building the smartest operating model.
Bottom line: Enterprise Intelligence is becoming the defining strategic asset of the AI era
The first generation of enterprise AI was about gaining access to intelligence. The new generation is about creating intelligence that competitors cannot buy because it is built from an organization’s own expertise, operating context, and continuous learning.
Services-as-Software™ provides the operating model for making that possible. It transforms human expertise into Enterprise Intelligence, protects the knowledge that differentiates the enterprise, and ensures every business outcome strengthens the organization rather than an external platform.
Every enterprise will soon have access to increasingly capable AI. The organizations that are already leading their industries are those that continuously build enterprise intelligence through Services-as-Software™, because that is the one asset their competitors can neither license nor replicate.
Posted in : Agentic AI, AGI, Artificial Intelligence, Business Process Outsourcing (BPO), Buyers' Sourcing Best Practices, GCCs, Governance, IT Outsourcing / IT Services, Large Language Models (LLMs), Leadership, Sovereign data
