How to Talk to Clients About AI Disruption in Software

The conflict with Iran sent shockwaves through the world and markets, requiring significant investor attention. While this has caused anxiety, there are other important dynamics in the markets that remain relevant for advisors. Before the conflict, software-as-a-service (SaaS) stocks began to struggle, with the selloff accelerating in late January and February following the release of Anthropic's Claude Cowork, a desktop agent that automates workflow across files and tasks. This and other AI advancements appeared to threaten the competitive advantages of widely used SaaS products, raising concerns about AI structurally disintermediating SaaS companies in the coming years. The IGV ETF, tracking the U.S. software industry, fell more than 20% in weeks and remains over 25% off its September high, despite some recovery.

This market action has caused anxiety about U.S. tech exposures. The following are key takeaways from a conversation with Maria Karahalis, Investment Director and Portfolio Strategy Manager at Capital Group, along with additional insights, to provide advisors with context and talking points for client conversations about these concerns.

Market Concern: AI-assisted coding (“vibe-coding”) allows companies to create their own software

• ​The bear case: If AI can generate functional, customized software from plain-language prompts, the value proposition of off-the-shelf SaaS products erodes. Why pay ongoing subscription fees for a mass-market product when you can build something tailored to your workflow quickly? This is especially concerning for niche, single-function software products, which are most vulnerable to being replicated by AI agents with minimal technical oversight.
• ​Is this economical? For example, a basic Shopify subscription costs $29/month and serves millions of customers. High-quality software often has low barriers to entry, and companies focused on growth may find it more cost-effective to use robust software services rather than spend time and resources building their own.
• ​Companies like Salesforce and ServiceNow are deeply embedded in enterprises that have invested years and millions into their tech stacks. While their status isn't permanent, companies are unlikely to replace their software overnight. The focus may be on integrating new technology with existing infrastructure rather than reinventing it.

Market Concern: AI allows companies to be more productive with fewer employees, damaging software companies’ seat-based pricing models

• ​Most major SaaS companies charge customers based on the number of users. If AI agents can handle tasks previously performed by many employees, companies may reduce the headcount interacting with these platforms. Even a modest reduction in seats across millions of contracts could lead to significant revenue decline. Pivoting away from seat-based pricing involves friction—repricing contracts, retraining sales teams, and rebuilding financial models around usage-based revenue is a long-term process, with potential revenue shortfalls in the interim.
• ​However, companies are projected to spend more on software in 2026, not less. A February report from Gartner estimates $1.4 trillion will be spent on software in 2026, a 15% growth from 2025. AI adoption may be driving new software purchases, such as security tools, AI management platforms, and cybersecurity software, which could offset potential seat losses.
• ​The job displacement narrative may be overblown. In fact, job postings for software engineers are spiking.

Market Concern: Software companies are racing to create their own AI systems but may not keep pace with AI-native architectures

• ​Foundation model companies like Anthropic and OpenAI are building AI systems from the ground up, with reasoning and autonomy at their core. Traditional software companies are retrofitting AI onto platforms built for a pre-AI world, raising concerns that incumbents can't match the pace of foundation models. The historical parallel is the cloud transition, where on-premise giants scrambled to build cloud products while born-in-the-cloud companies outpaced them. The worry is that today's SaaS incumbents are in the same position as yesterday's on-premise vendors.
• ​However, incumbents have advantages: proprietary data, distribution, customer relationships, and high switching costs. Proprietary data is valuable for training AI models in specific business contexts, which AI-natives can't easily replicate. Incumbents who leverage their data effectively could build better domain-specific AI than foundation model companies.
• ​Oracle and SAP, despite being challenged during the cloud transition, remain large, profitable companies. The transition created winners and losers among incumbents rather than eliminating them entirely.

Key Takeaways

• ​There is genuine excitement about the capabilities of new AI technology. Change can cause some businesses to fail, but also allows others to adapt and emerge stronger.
• ​The narrative of a broad-based “SaaS-pocalypse” is likely unfounded. There will be winners and losers, and markets are still determining who can adapt. This shakeout may take months or years, potentially favoring active managers in the interim.
• ​The future of software will likely differ from the past, but equity investors with long time horizons and disciplined plans are more likely to benefit from structural change than those who abandon the sector entirely.