By Alexander Jones, International Banker
On January 12, the US artificial intelligence (AI) company Anthropic announced a new product, Claude Cowork, which includes a suite of plugins that enable companies to automate key processes and workflows across various business divisions, including sales, legal and finance. The launch has since triggered a rapid devaluation of software-as-a-service (SaaS) stocks, especially of companies with solutions that most closely overlap with Cowork’s. With AI agents increasingly able to perform tasks and workflows that subscription-based software has supported, some now foresee a “SaaSpocalypse”, with much of the industry facing an existential threat.
Cowork allows users to connect Claude directly to external tools, data sources and workflows, enabling them to retrieve data, generate outputs and execute tasks through a single conversational interface rather than navigating separate software environments. The release effectively positions Claude not only as a standalone assistant but also as an orchestration layer capable of interacting directly with third-party tools, pulling structured data and executing multi-step workflows.
Having widespread access to such features, therefore, threatens the traditional SaaS interface. Tasks that would otherwise require navigating multiple applications—whether that be querying a CRM (customer relationship management), generating a report or updating records—can be handled through a single layer that sits above those systems. According to Forrester, investors are exiting their software positions in part due to fears that SaaS products are fundamentally too complex, and users struggle to manage the sprawl of hundreds of SaaS applications “that don’t talk to each other”.
The strategic implication is that if the necessary software can now be at the user’s fingertips, the rationale for paying recurring SaaS subscriptions becomes more difficult to justify.
The strategic implication is that if the necessary software can now be at the user’s fingertips, the rationale for paying recurring SaaS subscriptions becomes more difficult to justify. “In a build versus buy decision, most companies decided buy—buying enterprise software—made more sense. By buying the software, solutions could be implemented more quickly and at a lower cost because development costs were spread across all the customers of an enterprise software company,” supply-chain researcher Steve Banker explained in a March 26 article for Forbes. “The ability of AI-assisted development has changed that equation. Today, teams can move from idea to prototype in hours instead of months.”
Equity markets reacted tellingly, particularly on February 3, after Anthropic released its legal plugin for Cowork, which demonstrated Claude’s ability to navigate structured legal workflows—a domain traditionally dominated by specialised SaaS providers. “Legal automates contract review, NDA triage, compliance workflows, legal briefings, and templated responses—all configurable to your organisation’s playbook and risk tolerances,” Claude explains on its website. “Built for commercial counsel, product counsel, privacy/compliance, and litigation support teams.”
Shares of legal-tech and documentation-heavy software firms were hit hard, along with adjacent enterprise SaaS names, including Thomson Reuters, Sage and Wolters Kluwer. “Anthropic launched new capabilities for its Cowork to the legal space, heightening competition,” Morgan Stanley wrote in a note about Thomson Reuters, a provider of business and legal information. “We view this as a sign of intensifying competition, and thus a potential negative.”
According to reporting from Bloomberg, the release triggered “a $285 billion rout in stocks across the software, financial services and asset management sectors” on the same day. “A Goldman Sachs basket of US software stocks sank 6 percent, its biggest one-day decline since April’s tariff-fuelled selloff, while an index of financial services firms tumbled almost 7 percent,” Bloomberg also noted. “The Nasdaq 100 Index fell as much as 2.4 percent at one point before trimming losses to 1.6 percent.”
The S&P 500 Software and Services index, meanwhile, dropped by almost 4 percent, before losing another 0.73 percent the following day—the sixth consecutive session of losses—which Reuters estimated on February 4 had wiped out a staggering $830 billion in market value since January 28. “The selloff, which arguably started last quarter, is a manifestation of an awakening to the disruptive power of AI,” James St. Aubin, chief investment officer at Ocean Park Asset Management in Santa Monica, California, told Reuters. “The seemingly wide moats of these companies feel a lot more narrow today as competition from AI-created products intensifies. Perhaps this is an overreaction, but the threat is real, and valuations must account for that. My biggest fear is that this is a canary in the coal mine for the labour market.”
That competition may be beyond the reach of even the biggest players in the SaaS space, especially given how the behemoths of big tech are now integrating AI into popular existing tools to enhance functionality and improve user experience at scale.
Microsoft’s Copilot stack is embedded across its core enterprise footprint—from Word and Excel to Teams, Dynamics and GitHub—allowing users to generate documents, analyse data, write code and automate workflows directly within the applications they already use. Similarly, the company’s Azure OpenAI services enable enterprises to build custom AI layers on top of their internal data, effectively transforming the cloud into both the infrastructure and the orchestration layer for enterprise software. Instead of competing with standalone applications, Microsoft is absorbing functionality into a horizontally integrated environment where AI sits at the centre of user interaction.
Largely through its Gemini ecosystem, meanwhile, Google is embedding AI across its productivity solution, Workspace, as well as its cloud services and developer tools. Within Workspace, AI is increasingly capable of drafting documents, summarising communications, generating spreadsheets and coordinating tasks across Gmail, Docs and Sheets. Google is also crucially linking these capabilities to its cloud platform, where enterprises can deploy AI models that interact directly with proprietary datasets. Rather than purchasing discrete SaaS tools for specific functions, users can thus increasingly execute those functions within a unified AI-facilitated environment.
When AI is embedded at the platform level, the need to access separate, standalone applications that replicate similar functionality declines. Indeed, features that once supported entire SaaS categories, including reporting, workflow automation, basic analytics and internal tooling, can instead be incorporated into a broader ecosystem. In such environments, the point of control shifts away from individual applications and toward the platforms that own the interface, the data layer and the distribution channel.
Again, if organisations can generate bespoke tools quickly and cheaply, the economic case for paying recurring subscription fees for SaaS solutions becomes weaker. This trend has been growing since at least 2024, when payments fintech (financial technology) firm Klarna confirmed it had stopped using Salesforce’s popular CRM product in favour of its own in-house AI system based on OpenAI’s ChatGPT, aptly demonstrating the growing movement away from purchasing standardised applications to generating internalised, context-specific ones.
According to Forrester, however, not all SaaS companies will necessarily fare badly in an AI-first world. The US advisory firm predicted that:
- Horizontal-point-solution SaaS vendors will be challenged. “Vendors with low switching costs and weak, non-embedded enterprise workflows won’t scale beyond their current customer base,” Forrester explained in a February 11 report.
- Vertical- or domain-specific SaaS vendors will have a greater chance of survival. With market growth expected from roughly $133.5 billion (2025) to $194.0 billion (2029 forecast), Forrester sees vertical software continuing to thrive if SaaS vendors can offer differentiated solutions that address complex industries, such as healthcare and manufacturing, or that control unique, proprietary data.
- Enterprise SaaS companies racing to become AI companies still risk disintermediation. The shift to AI-first companies from such vendors as Oracle, Salesforce and ServiceNow will “attempt to preserve their moat using what’s hard for AI-only companies to replicate”, including their deep vertical experience-building specialised solutions; deep bench of consulting partners; access to vast customer data for benchmarking and machine learning; and the integration of people, process, tech and governance. “As agentic control planes, orchestration layers, and semantic data layers increasingly sit above these suites, however, even AI-first SaaS vendors risk disintermediation,” Forrester warned.
In the meantime, investors are increasingly differentiating between SaaS companies with value anchored in proprietary data, embedded workflows and/or regulatory complexity and those with value more closely tied to interface and usability. The former group retains a degree of defensibility even in an AI-facilitated environment. The latter faces a more direct challenge, however. If an AI layer can replicate or bypass the interface, then the economic moat shifts upward, away from the application and toward the orchestration layer that controls access.