Salesforce is navigating the AI transition by championing a hybrid model of "apps and agents." This strategy positions its traditional software ("apps" for humans) as the foundation, which is now extended and made more powerful by AI ("agents"). This narrative preserves the value of their core offerings while embracing AI's productivity gains.

Marc Benioff asserts that the true value in enterprise AI comes from grounding LLMs in a company's specific data. The success of tools like Slackbot isn't from a clever prompt, but from its access to the user's private context (messages, files, history), which commodity models on the public web lack, creating a defensible moat.

Experts argue Salesforce's AI strategy is flawed. Instead of building competing models, it should focus on making its platform indispensable for agents from OpenAI and Anthropic. This positions Salesforce as the essential 'venue' where humans and AI interact, increasing the value of its core subscription without competing directly with frontier labs.

As users increasingly interact with CRM data via external tools like Slack and AI, the core value shifts from the UI to the data structure. This could prompt new companies to choose cheaper, flexible databases over expensive, full-featured CRMs, threatening Salesforce's market position.

For an existing SaaS company to become a credible AI player, it must do more than add features. It requires a fundamental reinvention, akin to 'open heart surgery.' This means re-architecting to be agent-first, shifting to outcomes-based pricing, and even rebranding—as seen with Intercom becoming Finn—to attract new funding.

The CEOs of two similarly valued companies, Salesforce and Verizon, hold opposite views on AI's impact. Marc Benioff, whose SaaS business is seen as vulnerable, argues AI enhances his product's value. Conversely, Verizon's Dan Schulman, whose telco business is resilient to AI disruption, predicts staggering 20-30% unemployment, creating a puzzling dichotomy in executive outlooks.

When a SaaS company successfully launches a new AI product, it creates a second, conflicting business. It must manage the legacy SaaS model (seats, predictable metrics) alongside the new AI model (outcomes, unpredictable metrics), creating tension in strategy, branding, and operations.

To succeed in the AI era, SaaS companies cannot just add AI features. They must undergo a 'brutal' transformation, changing everything from their org chart and GTM strategy to their core metrics and pricing model. This is a non-negotiable, foundational shift.

Salesforce's announcement of a "headless" product was primarily a marketing move to acknowledge the rise of AI agents. In reality, it was a rebranding of their existing APIs, with no substantial change to the underlying product. This highlights a trend of incumbents using new jargon to appear innovative.