Businesses should be clear about what they’re deploying AI for

However, standalone subscription software is now under severe pressure. Despite rapid product and feature launches, XaaS companies are struggling to sustain revenue growth. The rise of Generative AI (GenAI) is altering the economics of software, leading to a shift where AI is no longer a product but an embedded layer. This change is visible globally and Indian businesses must adapt.

Also Read: How AI is transforming the role of venture capital for startups

The success of XaaS stemmed from its ability to create digital moats. Adobe eliminated piracy by making Photoshop subscription-based. Salesforce unbundled Customer Relationship Management software into separate offerings, like its sales cloud and marketing cloud. SaaS firms introduced more modular pricing, offering micro-services separately to maximize revenue per customer. 

But GenAI has launched an era of instant copyability—new features can be copied quickly. OpenAI’s ChatGPT, once a breakthrough, is now one of many AI chatbots, with free alternatives emerging just months after each new iteration.

Moreover, switching costs have crashed. Earlier, SaaS businesses relied on the difficulty of moving data to retain customers. Now, automation lets users switch easily. In enterprise software, large clients that once hesitated to shift platforms are now open to exploring cheaper AI-driven options.

The open-source threat and subscription dilemma: The rise of open-source AI models has been disruptive. DeepSeek, Llama and hundreds of others have made high-quality AI models freely available, leaving even AI giants in a scramble to justify their charges. 

Despite a wave of new capabilities, premium pricing is hard to sustain. Salesforce, Adobe and Zoom, all of which once thrived on SaaS pricing, are feeling the impact. Salesforce’s revenue growth has slowed to low double digits in spite of its AI-powered Einstein GPT. Adobe faces competition from free creative tools, forcing it to increasingly consider bundling AI with existing products.

Also Read: Jaspreet Bindra: Expect Agentic AI to make big waves in 2025

The pivot—AI as an embedded layer: With declining pricing power with chatbots, firms are loading existing platforms and ecosystems with AI. Google includes GenAI within Google One without an extra charge (so far). Amazon will integrate AI into Prime services. 

Microsoft’s Copilot AI, initially priced at a monthly $30 per user, has struggled to gain mass adoption. China’s Baidu, Alibaba and Tencent are embedding GenAI into the cloud, away from standalone AI subscriptions. BYD is bundling self-driving AI into cars for free, unlike US-based Tesla. Charging extra for AI will probably get difficult across industries around the world.

Agentic AI and the domestic challenge: The buzz now is around AI agents or automated systems that perform tasks on behalf of users. Those dazzled by what agents can do tend to underestimate the ease of building them. Creating most AI assistants is almost as easy as playing street cricket—widely accessible, rapidly replicable and requiring little. This is not true for all agents, but their democratization because of low creation barriers is both good and problematic.

For an Indian software enterprise, the key question is whether the AI agents it is building can help users increase their revenues. Many agents will deliver cost savings, which is a valid reason to adopt them. However, users expecting big top-line boosts because of agents—especially in service industries—need to think again. Remember, AI-driven features are no longer a differentiator now that these capabilities can easily be replicated.

Agents with unique and valuable characteristics could create remarkable new opportunities, no doubt, but they cannot count on AI by itself. AI needs to be integrated appropriately into core business processes, which goes well beyond just adding chatbots and workflow automation. 

A legal tech firm, for example, may find that an AI agent can draft contracts and cut costs, but if every competitor has the same feature, it will not enlarge revenues. Instead, firms must explore proprietary data advantages, deeper integration with existing enterprise software and AI-enhanced human decision-making processes.

Also Read: Redundancy alert: Here’s how AI assistants are threatening Indian software code factories

Adapt with a clear path to profits: XaaS revolutionized software, but GenAI is now pushing all players to a different paradigm. The key to survival is thoughtful adaptation. While AI adoption is inevitable, blindly chasing new features without a clear revenue strategy may not help. Companies that persist with old subscription models for incremental AI capabilities will face diminishing returns. The winners will be players that embed AI effectively within broader ecosystems, ensuring that its value is inseparable from the core product or service.

In this new era, differentiation will rely less on having the most advanced AI and more on integrating it in support of established business strengths. Factors such as physical product innovation, brand equity, distribution networks, relationships and proprietary data will again become primary business moats. AI alone is not a business differentiator. How it is applied within these foundational assets will determine long-term success.

The author is a Singapore-based innovation investor for LC GenInnov Fund.