A persistent question has risen to the top of conversations among technology and finance leadership over the past 18 months: Is Software-as-a-Service (SaaS) dead?
Amplified by prominent industry voices, the position argues that agentic artificial intelligence (AI) will collapse the traditional software stack, that large language models (LLMs) will enable any company to build its own systems on demand, and that the seat-based subscription model has reached the end of its useful life.
Broader market signals also add fuel to the narrative. Slower sales cycles, increased scrutiny on software spend and return on investment, and a reset in SaaS valuations have led some to interpret current conditions as a structural decline rather than normalization.
Despite this, SaaS is not dead — it’s evolving. The SaaS industry is entering a period of meaningful evolution shaped by AI, economics and changing expectations from enterprise buyers.
How Does AI Disrupt the SaaS Model?
The argument against SaaS when it comes to AI rests on three connected ideas:
- Traditional SaaS sells access to a workflow through a user interface and charges per seat. If an AI agent can execute the same workflow directly against the underlying database or application programming interface, the interface and the seat may become unnecessary.
- Much of what mid-market SaaS provides is functionally a database with forms and reports on top. AI-assisted development now makes it dramatically cheaper to build custom versions of that pattern.
- AI agents shift value capture toward the model and orchestration layer. This threatens to commoditize the underlying application as a data source.
The sharpest version of the argument against SaaS is that a buyer no longer cares whether the workflow runs on a commercial platform or a system built in-house — only that the work gets done.
How AI Is Reshaping SaaS Across the Market
Large Enterprise Platforms
Large enterprise platforms like Salesforce, Microsoft, SAP, Workday and ServiceNow are positioning themselves as agent platforms — owning the data, identity and workflow primitives so that agents run on top of their suites rather than around them.
Pricing is shifting toward consumption layered on top of seat licenses, and the larger providers are investing aggressively in agent orchestration, governance and integration tooling.
Mid-market Horizontal SaaS
Saas providers in the mid-market are embedding AI deeply into existing workflows to increase stickiness and expand per-customer value before agents can disintermediate them.
Many are opening agent-friendly interfaces and model context protocol endpoints, recognizing that being the preferred system of record for an agent may be more valuable than controlling the interface itself.
Vertical SaaS
Vertical SaaS providers in healthcare, hospitality, construction and field services look the most defensible.
Deep industry-specific workflows, regulatory requirements and proprietary data create moats that generic agents do not easily cross. These providers are adding AI features incrementally without existential repositioning.
Point Solutions
Single-purpose tools like meeting transcription, simple scheduling and lightweight analytics are the most exposed. Their value proposition was often a polished interface over commoditizable functionality, which is exactly what a capable agent can replicate. These providers should expect consolidation, acquisition or repositioning as agent-callable services with usage-based pricing.
Why SaaS Is Stronger, Not Weaker, in an AI-driven Model
The counter-argument is that this framing confuses the interface with the system of record. AI agents have to read from and write to something — a trusted, permissioned, audit-ready foundation of structured data, integrations and controls.
That groundwork is precisely what SaaS providers have spent two decades building, and it does not get easier to construct simply because the model layer has improved. In an agent-driven world, the system of record arguably becomes more valuable, not less, because every agent action depends on it. AI actually needs SaaS.
Enterprise buyers also continue to demand security reviews, compliance certifications and operational resilience. These requirements favor established providers and disadvantage bespoke or do-it-yourself alternatives.
The seat-based pricing model is a billing convention, not a law of physics; consumption, outcome, and agent-based pricing models preserve the underlying business while adapting to how value is delivered.
SaaS vs. AI-led Internal Development: What Organizations Are Actually Deciding
While AI lowers the technical barrier to building software, the decision between SaaS and internal development is not purely technical or financial.
There is a case for internal build in specific scenarios — particularly in well-defined domains such as accounting and finance. Under Accounting Standards Codification (ASC) 350-40, certain development costs can be capitalized and amortized, which may create a more favorable financial profile than expensing SaaS subscriptions.
However, that financial view is incomplete on its own. When an organization buys SaaS, it acquires more than software — it acquires a counterparty accountable for uptime, security patches, regulatory updates, disaster recovery and product roadmap.
When that same organization builds internally, all of that accountability transfers in-house. It extends beyond IT and reaches the controller, CFO, audit committee and board.
Owning a core system internally introduces ongoing responsibilities that include:
- Sarbanes-Oxley (SOX) control design and testing
- Real-time response to regulatory changes from the Financial Accounting Standards Board (FASB) and the Internal Revenue Service (IRS)
- Security and incident response obligations under evolving Securities and Exchange Commission (SEC) disclosure rules
- Talent continuity and system maintenance as engineers who built the system move on
- Ongoing investment to support growth roadmaps (acquisitions, geographies and other integrations)
Commercial SaaS providers absorb these costs across their customer base. Internal teams must sustain them independently, competing for resources against other priorities.
AI-assisted development lowers the cost of building. It does not change the organizational readiness required to operate, secure, audit and maintain systems over time. That readiness — covering information technology operations, security and compliance, finance process maturity, and governance culture — is the rate-limiting factor for most organizations considering whether to build core systems in-house.
What SaaS Providers and Buyers Should Do Next
Providers
The pressure on SaaS providers is real, but it is more about pricing model disruption and margin compression than wholesale displacement. To remain competitive, providers should:
- Lean into compliance, audit readiness and accountability as durable moats — particularly for finance, human resources and regulated industries where the buyer personally signs certifications.
- Shift pricing toward consumption, outcome and agent-based models that mirror how internal-build alternatives would be costed and that align provider revenue with customer value.
- Embed more deeply in adjacent workflows so that replacing the core means rebuilding a dozen integrations, not just one system.
- Offer managed or hybrid platforms where customers can customize and extend — and potentially capitalize on that work — while the provider retains operational, security and regulatory accountability.
Providers that own systems of record and regulatory accountability are likely to strengthen their position, while point solutions face more immediate disruption.
Buyers
Layering organizational readiness onto the segment analysis produces a more grounded view of where SaaS pressure actually lands.
Large enterprises may have the capability to build, but generally extend internal development around a commercial core rather than replacing it. Mid-market and private equity-backed organizations face stronger incentives to build, but often lack the operational depth to sustain those systems. Highly regulated industries will move more slowly regardless of cost advantages.
For buyers, it’s clear AI is changing the cost of building, and accounting treatment can make internal development appear attractive. But the accountability for delivery — operational, regulatory and reputational — does not move with the cost curve. Buyers should not only ask whether building is cheaper, but also whether the organization is structured to own what it builds.
The Future of SaaS Is Evolution, Not Extinction
The headline "SaaS is dead" makes for a memorable conference slide. The reality is more measured and, for thoughtful providers and buyers, more consequential.
The SaaS category is splitting into two paths:
- Providers that own data, workflows, compliance and accountability will emerge stronger.
- Providers whose value proposition was a thin layer of interface over commoditizable functionality face real disruption, and in real cases, consolidation.
SaaS is not ending. It is being asked to prove, more clearly than ever, what it is accountable for — and to price accordingly.
How Cherry Bekaert Can Help
Cherry Bekaert advises technology providers and the organizations that buy from them on the strategic, financial and operational dimensions of this transition.
Our Digital Advisory professionals work with software companies on revenue model evolution, mergers and acquisitions, and compliance posture, and with corporate buyers on build-versus-buy analysis, software cost treatment, and the control implications of internally developed systems.
To discuss how these themes apply to your organization, contact us today.
