Service-as-a-Software: The More Things Change, the More They Stay the Same

The concept of Service-as-a-Software is on the rise, thanks to advancements in Generative AI (GenAI). Many are calling this a groundbreaking new paradigm. But is it truly new? 

As someone who worked on the idea of Service-as-a-Software in the Big Data Analytics space years ago, I can confidently say this is not new. 

The core challenges, opportunities, and misconceptions surrounding it remain as relevant today as they were back then. 

While many people have explored the potential applications areas of Service-as-a-Software, in this blog I explore from my perspective the idea of Service-as-a-Software, its evolution, and what GenAI brings to the table.

‍

The Earlier Framings of Service-as-a-Software

The concept of Service-as-a-Software has existed earlier though not very well known and adopted in the market. It has been around in various forms for years. My own experience dates back to my time at Mu Sigma, a Big Data Analytics unicorn, where the founder championed the idea of Service-as-a-Software (as far back as 2016)  in the context of big data analytics services and solutions.

At that time, the idea was suggested as an alternative to the traditional SaaS paradigm that the venture capital (VC) ecosystem was forcing upon analytics solutions startups. Traditional SaaS works well in relatively static contexts, where product requirements don’t vary significantly across customers and evolve only incrementally over time.

But the reality of business problem-solving with analytics was (and still is) far messier:

• Dynamic Data: The data landscape changes rapidly. New sources, formats, and volumes emerge regularly.
• Muddy & Fuzzy Problems: Business challenges are seldom static. They evolve as leaders frame and reframe questions and problems.
• Unique Contexts: Each organization’s data, culture, and objectives require tailored solutions that don’t fit into a one-size-fits-all SaaS model.

Traditional SaaS models for analytics fell short because they couldn’t keep up with these dynamics. 

This gap was often filled by analytics services firms, consulting firms, or platform-based companies like Palantir and C3.ai, which introduced concepts like “forward-deployed engineers” to customize solutions for specific use cases.

The Service-as-a-Software model we proposed at the time was built around the idea of:

1. Increased Software-ization of Services: Using modular, configurable blocks of software that could be dynamically stitched together workflows to solve specific business problems.
2. Meta-Software: Purpose-built abstracted software modules designed to enable this dynamic assembly, allowing services to evolve alongside changing business needs.

This approach aimed to address the inherent fuzziness of business problem-solving by combining the flexibility of services with the scalability of software. 

VCs were intrigued but skeptical, often returning to their SaaS-first playbooks.

‍

The Current framing of Service-as-a-Software

Fast forward to today, Service as a Software is essentially the ability to deliver as software those offerings that were traditionally delivered through a services (manual, human intensive) operating and business model.

And the conversation has shifted to Generative AI (GenAI) as the enabler of Service-as-a-Software. 

GenAI undeniably advances the concept by introducing the promise of:

• Dynamic AI Agents: Capable of addressing specific tasks contextually and almost autonomously.
• Improved Modularity: AI-driven systems can adapt workflows based on context and objectives.
• Multi-Agent Orchestration: Coordinating multiple AI agents to execute complex workflows

‍

The Challenges Then and Now

Despite these advancements, the challenges that plagued Service-as-a-Software remain. Here’s a look at where GenAI may hit the same roadblocks:

1. Problem Framing is still required

Even with GenAI, business problems remain fuzzy and ill-defined. While AI can assist in contextual process automation, data processing, and recommendations, the initial framing of problems still requires a mix of art and science—something GenAI has not mastered yet.

2. Dynamic Workflows are Limited

The dream of GenAI autonomously stitching together “agents” to create dynamic workflows is compelling. But in practice, this capability is still in its infancy. Current AI systems work best with well-defined objectives and workflows, leaving more complex or ambiguous scenarios to human intervention.

3. High Gross Margins are a Myth

In data-heavy fields like analytics, even AI-powered solutions often rely on significant professional services components for customization and implementation. This limits scalability and impacts gross margins—a problem SaaS models were supposed to solve.

‍

What GenAI Brings to the Table

Despite these limitations, GenAI does push Service-as-a-Software forward in meaningful ways:

1. Scalability for Defined Problems: AI is highly effective for business processes with clear objectives and defined steps.
2. Real-Time Adaptability: AI can process and respond to data changes in real-time, enabling more dynamic solutions than traditional SaaS.
3. Personalization at Scale: GenAI enables tailored user experiences, making it easier to adapt software to individual customer needs without manual customization.

‍

So what? Selectively Preserve the Past while Adopting the New

The potential of GenAI in Service-as-a-Software is real, but it won’t eliminate the challenges inherent in business problem-solving. For companies operating at the frontiers of analytics and AI, success will depend on balancing these dynamics:

• Leveraging AI to handle repeatable, scalable tasks.
• Embracing Services where human expertise is indispensable.
• Developing Modular Frameworks that bridge the gap between AI’s capabilities and the messy realities of business needs.

‍

Conclusion: A Familiar Revolution

The rise of Service-as-a-Software is a classic case of “the more things change, the more they stay the same.” While GenAI introduces exciting possibilities, the core challenges of dynamic workflows, fuzzy business problems, and gross margin pressures remain.

For companies at the cutting edge, success will depend on blending the best of AI, modular software, and services into solutions that evolve with customer needs. Service-as-a-Software isn’t just a new trend—it’s a necessary evolution for businesses navigating complexity in a rapidly changing world.

Are you ready to embrace the next wave of Service-as-a-Software? Let’s discuss how AI and modular design can transform your approach to business problem-solving.