Between $250 and $400 billion annually, that’s the estimate for the global SaaS market, depending on the sources and the scope considered. Fueled by the rise of cloud-native infrastructures, software-as-a-service has spread across all application verticals.
In the 2025 edition of its Technology Report, Bain & Company dedicates an analysis to how generative and agentic artificial intelligence is redrawing the economic contours of this strategic sector. A technological shift raises a fundamental question: will we witness market expansion or destructive value cannibalization?
The evolution of AI model costs is the first indicator of an economic disruption. OpenAI’s reasoning model (o3) saw its cost drop by 80% in just two months. A drastic fall in costs, combined with ongoing improvements in precision, is redefining the sector’s economic equation.
Over a three-year horizon, Bain predicts that every routine, rule-based digital task could migrate from the traditional “human + application” model to an “AI agent + API interface” model. This transition represents a major shift in value creation and capture, and thus a vital challenge for SaaS publishers.
Four Economic Scenarios for a New SaaS
Bain’s analysis identifies four distinct economic trajectories depending on the automation potential of user tasks and the extent to which AI can penetrate workflows.
1. AI enhances SaaS
In the “AI enhances SaaS” scenario, incumbent publishers maintain their competitive edge. These segments are characterized by low automation and low AI penetration. Flows rely on human judgment, strict oversight, and deep domain knowledge (e.g., Procore’s project-cost accounting or Medidata’s clinical trial randomization). In this case, publishers use AI to boost productivity, protect unique data that differentiates the offering, and price time-savings at a premium.
2. Spending compresses
The “spending compresses” scenario exposes established players to new economic risks. The human role persists, but third-party agents can connect to APIs and siphon value (e.g., HubSpot list creation or Monday.com task boards). In this scenario, publishers must quickly launch their own agents, deepen partner integrations to raise switching costs, and limit access to critical endpoints.
3. AI eclipses SaaS
In the “gold mines” where AI surpasses traditional SaaS, companies gain an edge thanks to exclusive data and rules enabling full automation (e.g., Cursor’s AI-powered code editor). The publishers’ strategy will be to build end-to-end agent-driven solutions, train sales teams to sell outcomes, and move from seat-based pricing to outcome-based pricing.
4. AI cannibalizes SaaS
These are the battlegrounds. Tasks are readily automatable and duplicable (e.g., first-line support with Intercom, Tipalti’s invoice approvals, or ADP timesheet approvals). In this scenario, the aim is to proactively replace SaaS activity with AI. The publisher must choose between becoming a neutral agent platform or supplying the data that fuels it. The winners will be those who orchestrate agents most effectively.
The Emergence of a New Value Architecture
According to Bain, we are witnessing a fundamental restructuring of the SaaS value chain around a three-tier architecture.
> Record-keeping systems form the base, storing critical business data and managing access. Their economic advantage lies in unique data structures and regulatory logic that are costly to replicate.
> Agent operating systems orchestrate actual work, scheduling tasks and invoking the right tools. The current competitive edge comes from the scarcity of GPUs and the stack of proprietary tooling.
> Result interfaces translate natural-language requests into agent actions. Their economic power stems from embedding them in daily routines and user trust.
The Strategic Stakes of Semantic Standards
A crucial element of the analysis concerns the emergence of inter-agent communication standards. Protocols like MCP from Anthropic and A2A from Google aim to standardize exchanges, creating network effects with rapid tipping points and “winner takes most” dynamics.
Bain identifies a major economic challenge: the first semantic standard capable of establishing a shared vocabulary at industrial scale could redefine the AI ecosystem and drive a significant wave of value creation.
For established SaaS publishers, it’s a unique leadership opportunity but it requires high-risk strategic bets, notably selective open-sourcing and evolving monetization models.
How to Preserve Value Creation
In the face of these transformations, Bain lays out five key economic recommendations.
> Centralize AI in the product roadmap by identifying repetitive, automatable tasks and implementing solutions before customers look elsewhere. The objective: transform the product into a “do-it-for-me” experience with demonstrable ROI.
> Turn unique data into a durable competitive edge, because even if models like GPT-4o are ubiquitous, the value lies in proprietary data: usage patterns, specialized content, transactional history.
> Rethink pricing for an AI-first world by gradually moving away from per-user models toward value-based pricing: tasks completed, tickets resolved, AI outputs generated.
> Build AI mastery across the organization by making AI a core capability rather than a side project, requiring specialized talents and a culture of innovation.
> Shape the standards ecosystem by standardizing key object definitions within your platform and selectively publishing schemas where the company already excels.
Optional Obsolescence, Mandatory Disruption
Bain’s conclusion rests on a striking economic paradox: disruption from AI is inevitable in the SaaS sector, yet obsolescence remains optional. This disruption will sometimes expand the market, sometimes commoditize it, favoring incumbents or new entrants depending on the circumstances.
The economic message is clear: today’s SaaS leaders can shape the future rather than endure it, provided they adapt their investments and strategy to the specific context of each workflow, anchor themselves to the new platform layers, and invest in the semantic gaps affecting their developers.
In this race for transformation, speed of execution and the precision of strategic bets will determine who writes the next chapter of the SaaS economy before competitors step in.
