Why Software and Services Stayed Separate

Historically, software companies have largely avoided building large consulting, systems integration, or outsourcing organizations. Likewise, services firms generally avoided owning and developing significant software platforms.

There have always been exceptions. IBM built substantial software assets while maintaining a major consulting business. CGI successfully combined software and services capabilities. Yet these examples remain exceptions rather than the rule.

The reasons are straightforward. From a software company perspective, services are often viewed as margin dilution. Services businesses require more labor, lower scalability, and a fundamentally different operating model.

At the same time, services firms have traditionally focused on delivering expertise rather than building and maintaining technology platforms. The capital requirements, product development cycles, and risk profiles are very different. As a result, software and services evolved as complementary but distinct industries.

AI Changes the Equation

The emergence of AI raises an important question: Will AI change the relationship between software and services?

For traditional enterprise software, I believe the answer is largely no.

The existing software stack, whether it is ERP, CRM, productivity platforms, or enterprise databases, will remain in place for many years. I do not expect companies such as Microsoft, Oracle, or SAP to suddenly build large-scale outsourcing organizations around their traditional offerings. The economic logic that separated software and services still applies to these environments.

However, agentic-native systems are different.

Much of the technology investment flowing into the market today is focused on building agentic-native environments. Venture capital funding has shifted heavily toward these models, while established software providers are increasingly developing companion agentic offerings that either integrate into or operate alongside their existing platforms.

The reason is simple: these systems promise significantly greater business value than traditional software alone.

Agentic-native Systems Behave More Like Living Systems

Traditional enterprise software is designed for stability. Organizations implement the software, configure it to fit their environment, connect it through APIs, and then manage change through controlled release cycles. The architecture is intentionally structured to minimize disruption.

Agentic-native systems operate differently. These environments are often built around an ontology, a knowledge graph, and extensive data assets that provide context and understanding rather than simply storing information. Many are paired with digital twins that allow organizations to model future scenarios and anticipate change. On top of these foundations sits an AI and agentic layer capable of taking real-time actions based on what the system observes and predicts.

What makes these environments unique is their constant evolution. The ontology expands and adapts. The system continuously learns from new business data. Agents may be created, modified, used once, and then retired. New capabilities are absorbed from human processes and embedded directly into the agent ecosystem. Unlike traditional software, these systems are never truly finished.

Why Agentic AI Requires a New Services Model

Because agentic-native systems are dynamic, they require a different kind of support. Traditional software follows a "build, implement, and maintain" model. Agentic environments require continuous adjustment, governance, and optimization.

One of the most important innovations in this area came from Palantir's concept of forward-deployed engineers. Rather than remaining distant from the client, these engineers operate directly within the business, helping govern and refine the system as conditions change. This is fundamentally a services model.

The same organization that builds the technology increasingly places people inside the client environment to guide and evolve it. The software is no longer a static product delivered to a customer; it becomes a living operational system that requires ongoing stewardship.

As a result, software providers entering the agentic-native market increasingly need service capabilities as part of their architecture and delivery model.

The Coming Convergence

This shift creates something we have rarely seen in enterprise technology: a structural convergence between software and services.

Software companies developing agentic-native platforms will increasingly need embedded service organizations capable of continuously shaping outcomes for clients.

At the same time, services firms will likely seek greater ownership of intellectual property, platforms, and agentic technologies. We are already seeing signals of this trend as service providers invest more aggressively in proprietary assets and technology-enabled offerings.

This convergence will inevitably create new competitive tensions. Relationships that were once straightforward partnerships may evolve into forms of co-opetition. Software vendors and services providers will continue to work together, but they may also find themselves competing for ownership of the client relationship and the underlying technology platform.

The Strategic Implication

The most important implication is that agentic AI may force a rethinking of industry boundaries that have existed for more than four decades.

The traditional software stack will continue to operate under the old rules. However, the emerging agentic-native stack appears to require a fundamentally different model, one where software and services become increasingly inseparable.

For enterprise leaders, this means evaluating technology providers differently. Organizations will no longer focus on who has the best software or who provides the best services. Instead, the winners may be the providers that can successfully combine both into a single, continuously evolving operating model that delivers business outcomes in real time.