signüll: All SaaS Are Facing Dual Pressure from AI
signüll believes that SaaS companies like Figma, Airtable, and Canva are facing dual pressure from AI-native startups and model labs, and that relying solely on "ambient programming" is not the core threat.
Their assessment is that small AI-native teams are entering the existing SaaS market with lower R&D costs and faster iteration speeds from single-point workflows; model labs are integrating independent product capabilities in design, spreadsheets, and knowledge management into general dialogue and agent interfaces, compressing product categories and entry value from above.
Enterprise software buyers will reduce the purchase of standardized tools by seat licenses and instead buy AI workflows that are delivered by task and can execute across systems; SaaS with thin functionality, low migration costs, and a lack of proprietary data will face renewal and pricing pressure. Model vendors, AI-native vertical applications, and platforms that control enterprise data entry will compete for budgets, while SaaS with deep workflow integration, compliance certification, and high switching costs still has defensive space.
Source: Public Information
ABAB AI Insight
The previous moat for SaaS came from cloud delivery, subscription billing, and database lock-in. Salesforce, ServiceNow, and Adobe do not just sell features, but embed organizations through customer data, permission systems, workflow configurations, and third-party integrations; this is the part that is still difficult to replace overnight in the AI era. Klaviyo's acquisition of AI customer success company Agency, and appointing its founder as Chief Product Officer to advance AI agents, indicates that mature platforms are integrating AI capabilities into existing data and customer relationships through acquisitions.
The offensive from model labs is essentially a competition for user interfaces and task orchestration layers. If users can invoke the company's knowledge base, CRM, spreadsheets, emails, and design tools within a general agent, SaaS that originally relied on independent application interfaces will be downgraded to replaceable data sources or execution interfaces; Microsoft CEO Satya Nadella has previously suggested that AI agents may move business logic out of vertical, database-oriented software.
AI-native companies attacking from below follow a different capital path: they do not first replicate complete products, but instead take over a high-frequency, quantifiable work outcome, charging based on results instead of headcount. Forrester assesses that horizontal point SaaS with low switching costs, not deeply embedded in enterprise processes, and unable to prove immediate ROI are most vulnerable; software in industries like healthcare and manufacturing has a higher survival rate due to proprietary data, regulation, and business complexity.
This represents a restructuring of the supply chain. In the past, value primarily belonged to those "who owned the application interface"; in the future, value will be redistributed to those "who control data, schedule agents, and bear responsibility for outcomes." SaaS will not disappear entirely, but profits at the pure functionality level will be compressed; companies that can integrate data governance, industry rules, and agent execution authority into irreplaceable workflows may transform from being called tools to becoming a new control layer.
ABAB News · Law of Cognition
- Features will be copied, but workflows will be relied upon.
- When entry points are unified, categories begin to depreciate.
- Low-cost offensives seize features, while higher-level platforms seize relationships.