The results show how AI can change the way established software companies capture value. The impact will vary across the SaaS market.
AI Is Becoming Part Of Software Economics
Salesforce's Agentforce numbers offer one of the clearest examples.
Agentforce ARR exceeded $1.5 billion in Q2, while combined Agentforce and Data 360 ARR reached nearly $3.9 billion, up more than 210% year over year. Salesforce also reported 3.2 billion Agentic Work Units in the quarter, bringing its cumulative total to 7 billion.
The company defines Agentic Work Units as tasks completed by AI agents across Agentforce and Slack. The metric measures the amount of work performed by agents.
Data 360 ingested 104 trillion records during Q2, up 355% year over year, including 82 trillion records through Zero Copy, which increased 731%.
Agents need access to enterprise data and business context to perform useful work. Salesforce is expanding the commercial role of the data already sitting inside its platform.
Snowflake is pursuing a similar strategy from the data infrastructure side. Its Q2 results showed product revenue accelerating for the third consecutive quarter, while management pointed to AI workloads as an increasingly important driver of platform consumption.
Snowflake CEO Sridhar Ramaswamy described AI as creating a flywheel across the business. AI products can bring new users and workloads onto the platform, increasing consumption of the underlying data infrastructure.
The Software Model Is Moving Beyond Seats
The change is also visible in how enterprise software companies are selling AI.
Salesforce reported that bookings from its Agentforce One Edition and Agentforce for Apps premium offerings more than doubled sequentially in Q2. Slackbot users also grew more than 150% quarter over quarter.
ServiceNow provides another example. The company reported $3.88 billion in subscription revenue for Q2 2026, up 24.5% year over year, while ServiceNow AI crossed $1 billion in annual contract value. Agentic deployments increased ninefold in nine months.
These results point toward a model in which AI capabilities become part of the commercial value of the platform. Agent usage can also increase demand for the data, workflows and infrastructure underneath it.
Snowflake is building around this change. Its platform supports multiple AI models and includes tools for routing workloads according to factors such as performance and cost, giving customers more control over the economics of running AI workloads at scale.
Seat-based subscriptions remain an important part of the software market. The value of software can increasingly be tied to the work performed through it.
The Real Pressure Is On Vulnerable SaaS
That distinction matters for the SaaS debate.
Atlassian reported $1.59 billion in revenue for its fiscal Q2 2026, up 23% year over year, while remaining performance obligations rose 44% to $3.8 billion. Its cloud products include AI capabilities across its platform and work-management products.
The key dividing line may be what sits underneath a software product. Companies with important enterprise data, established workflows and systems that businesses already depend on have assets AI agents still need.
Salesforce's expansion of AI agents into customer-service workflows illustrates how an incumbent can use that position to extend its role.
NVIDIA CEO Jensen Huang made a similar argument during Cisco's AI Summit in February. He argued that AI systems would continue to use software tools as part of their work.
The SaaS market therefore faces a more complicated future than the phrase "SaaS is dead" suggests.
Software companies whose value rests on easily replicated functions face greater pressure. Companies that own critical data, workflows and enterprise infrastructure can make their platforms the places where AI agents perform work.
Salesforce and Snowflake are showing what that transition can look like in financial results. SaaS may not be disappearing. Its economic unit is changing.
