Enterprise Software Watch: Salesforce Tests a New SaaS Model for the AI Era
AI Agents, SaaS Economics and the New Enterprise Software Stack
Enterprise Software Is Moving From Applications Employees Use to Platforms That Execute Work
Salesforce used its investor day to outline an "agentic enterprise" strategy centered on AI-powered interfaces, enterprise data management, agent governance and expanded monetization models—reaffirming a $63 billion revenue target for fiscal 2030. The company introduced AIforce Max at $550 per user per month and demonstrated Claudeforce, which connects Salesforce capabilities to Anthropic's Claude through MCP servers. Meanwhile, Snap announced enterprise partnerships with Salesforce, NVIDIA and AWS to bring Agentforce, AI visual interpretation and workplace assistants into its $2,195 Specs AR glasses—a signal that enterprise software is extending beyond the screen into physical workflows.
The throughline is structural: enterprise software vendors are building agent orchestration layers, governance frameworks and outcome-based monetization models on top of—and increasingly in place of—traditional seat-based applications. Gartner estimates that up to $234 billion of enterprise application spending is exposed to agentic arbitrage through 2030—roughly 20% of SaaS application spending. As Gartner's George Brocklehurst put it, "Agentic systems deliver outcomes directly, bypassing traditional UX-heavy applications and making the software invisible. This breaks the link between user growth and revenue growth."
Salesforce Tests the Next SaaS Business Model
Salesforce's investor day made one thing clear: the company is no longer defending the CRM seat. It is building the infrastructure for a post-seat enterprise. The company reaffirmed its fiscal 2030 revenue target above $63 billion, with current trailing-twelve-month revenue at $43.94 billion. AI and data products—Agentforce and Data Cloud—are the growth engines. Agentforce and Data 360 annual recurring revenue has surpassed $2.9 billion, up roughly 200% year-on-year.
The architectural shift is significant. Salesforce introduced what it calls an "interface revolution"—agents sit between users and enterprise software. Rather than navigating separate applications, users express intent to an AI agent that accesses connected tools and returns answers, dashboards or applications. Claudeforce, currently in open beta with about 40 customers, connects Salesforce capabilities to Anthropic's Claude through MCP servers. Siemens is running a 25-user pilot and believes the technology could eventually reach 18,000 users.
Salesforce also introduced AIforce Max at $550 per user per month, including Slack and Slackbot. The company unveiled Guardian for agent identity controls, data classification and risk monitoring, and Agent Fabric—an extension of MuleSoft that discovers, registers, monitors and manages agents from multiple providers including Microsoft Azure, AWS and Google. Agent Fabric can monitor agent latency, error rates, policy violations and costs while recommending changes to reduce token consumption.
Why this matters: Salesforce is attempting to monetize the agent layer without abandoning the CRM seat entirely. But the company acknowledged it is "following startups rather than leading the change" on pricing innovation. The $550 AIforce Max price point and the Claudeforce beta suggest Salesforce is hedging: keep the seat, add the agent, and let the market decide which model wins.
Microsoft, ServiceNow, SAP and Oracle Build the Agent Layer
Across the enterprise software stack, the competitive battlefield has shifted from "who has AI features" to "who controls the agent layer." Each major vendor is building an orchestration, governance and identity framework—and each is positioning that framework as the enterprise's control plane.
Microsoft detailed new Agent 365 admin controls in mid-September, including a cross-vendor agent registry spanning Microsoft, AWS Bedrock, Google Cloud, Databricks Genie and Anthropic Claude agents. The company introduced Execution Container isolation to block unsanctioned local or "shadow" agents, reusable security policy templates, and agent cost monitoring with spend limits. Agent 365 is priced at $15 per user per month standalone or bundled in Microsoft 365 E7 at $99 per user per month—positioning Microsoft's agent registry as the enterprise control plane for multi-cloud agent estates.
ServiceNow has expanded its AI Control Tower to discover, monitor, govern, secure and measure AI agents across AWS, Azure, Google Cloud, SAP, Oracle, Workday and 25 more enterprise systems. The platform now includes real-time agent monitoring via the Traceloop acquisition, a kill-switch capability that can shut down agents that go off-script, and cost tracking and ROI dashboards. ServiceNow has also opened its Action Fabric to external agents through MCP—Anthropic's Claude Cowork became the first design partner, connecting Claude to ServiceNow's governed execution layer. The company has set a $30 billion-plus subscription revenue target for 2030, with AI expected to represent more than 30% of annual contract value.
SAP CEO Christian Klein told a Goldman Sachs conference that SAP will release only AI agents capable of solving the most complex enterprise problems—not simple tasks any large language model can handle. "Our ERP wants the most complex business processes in the world, and we need to solve the most complex challenges of our customers with our agents," Klein said. SAP is betting that its data, ontology and Business Data Cloud provide the context that general-purpose agents lack. The company acknowledged a headwind: 60% of SAP's installed base has not yet migrated to the cloud, a prerequisite for its agentic roadmap.
Oracle has launched an AI-native builder experience for Oracle AI Agent Studio, enabling customers and partners to create and run Fusion Agentic Applications natively within Oracle Fusion Cloud Applications. The environment supports no-code, low-code and professional development workflows and inherits Fusion Applications' existing security, access permissions, approval, audit trail and governance capabilities. Oracle has launched 22 Fusion Agentic Applications across HR, finance, supply chain and customer experience.
Workday unveiled its next generation of Illuminate Agents for HR and finance, including agents for contingent sourcing, supplier contracts, self-service and frontline workflows. Workday positions itself as "the AI platform for managing people, money and agents," with Flex Credits as its consumption currency—metered per agent skill, from 1 credit for information retrieval to 750 credits for talent pool lead identification.
What Happens to the Traditional SaaS Seat?
The SaaS seat is not disappearing overnight. But its economic foundation is eroding. IDC predicts that seat-based pricing will be largely obsolete by 2028. Industry analysts estimate AI agent deployments will eliminate 20–35% of enterprise SaaS seats by the end of 2027. The share of SaaS companies on pure per-seat pricing fell from 21% to 15% in a year.
The pricing models replacing seats vary by vendor. Salesforce charges $2 per autonomous resolution. HubSpot charges 50 cents per resolved conversation. Intercom charges nothing unless the outcome lands. Microsoft prices Copilot Credits from $0.01 each. SAP bundles roughly 200 AI actions per Advanced Full User Equivalent, with agent runs costing $0.40 to $1.80 past the pool. Workday meters Flex Credits per agent skill.
The common thread is that every vendor is experimenting with a new pricing currency—and no two are the same. As IT Pro reported, enterprises increasingly have to model tokens, credits, actions and variable AI consumption rather than simply counting seats. Procurement is becoming harder, not easier.
Why this matters: The "SaaSpocalypse" debate—which wiped nearly $2 trillion in software valuations in early 2026—was about whether AI would destroy SaaS. The more accurate framing, as Gartner suggests, is metamorphosis, not apocalypse. SaaS is not dying. It is being restructured around agents, outcomes and consumption.
Enterprise Software Moves Into Physical Workflows
Two developments this week illustrate how enterprise software is extending beyond the desktop into physical operations. Snap announced enterprise partnerships with Salesforce, NVIDIA and AWS to integrate workplace tools into its Specs augmented-reality glasses. Salesforce will add Agentforce to Specs, allowing workers to access company data and automate tasks through the glasses. NVIDIA will power AI that interprets a worker's visual environment and retrieves relevant company data. AWS will offer its Amazon Quick AI assistant for voice-command tasks like checking product availability.
Separately, INVE Technologies—formerly Identiv—completed the sale of its IoT assets to Trackonomy and is pivoting to a "physical AI compliance SaaS" strategy under new interim CEO James Greenwell, effective September 21. The company received $50 million in Trackonomy preferred equity and plans to pursue compliance SaaS acquisitions that combine software with physical-world data from Trackonomy's battery-powered smart labels.
Why this matters: Enterprise software has traditionally been confined to screens—desktops, laptops, phones. AR glasses and physical AI platforms represent an expansion of the enterprise software surface area into factories, retail outlets and field service operations. If agents can perceive the physical environment and act on enterprise data, the application layer expands into territory it has never occupied.
AI Changes the Enterprise Software Competitive Map
The competitive landscape is splitting into three layers: systems of record, agent orchestration, and outcome monetization.
Incumbents with deep data and process logic—SAP, Oracle, Workday—retain an advantage in systems of record. Their agents operate within transactional systems as native capabilities rather than external add-ons. But their challenge is cloud migration: SAP's 60% on-premises installed base is a structural headwind.
The orchestration layer is where Microsoft, ServiceNow and Salesforce are competing most aggressively. Microsoft's Agent 365 registry spans multi-cloud agents. ServiceNow's AI Control Tower governs agents from 25+ enterprise systems. Salesforce's Agent Fabric manages agents from Azure, AWS and Google. The vendor that becomes the default agent registry captures the control plane—and the customer relationship.
Outcome monetization is the least mature layer. Salesforce's $2 per resolution and outcome-based pricing for Agentforce are early experiments. But attribution disputes loom large—Stripe has warned that sales conversions may result from factors other than the software itself. Only 4 of 65 enterprise software companies analyzed by AlixPartners have fully adopted outcome-based pricing.
What CIOs and IT Buyers Should Watch
- AI attach rates: What percentage of enterprise software contracts now include AI agents or consumption commitments—and how do those attach rates vary by vendor?
- Agent adoption vs. seat decline: Are enterprises reducing seat counts as agent deployments increase? Watch net retention and seat expansion metrics at Salesforce, ServiceNow and Workday.
- Usage and action-based revenue: What share of vendor revenue now comes from non-seat sources? ServiceNow says half of its net new business is non-seat-based—the clearest proof point that the transition is real.
- Cloud backlog and AI ACV: SAP's 60% on-premises installed base and its cloud migration rate will determine whether its agentic roadmap can scale. ServiceNow's AI ACV crossed $1 billion; Workday's exceeded $100 million.
- M&A involving AI-native software: Salesforce acquired Fin for $3.6 billion. ServiceNow acquired Sweep. TrueFoundry acquired Seldon AI. Expect more consolidation as incumbents fill agentic capability gaps.
- Agent governance standards: Microsoft's multi-cloud registry, ServiceNow's AI Control Tower and SAP's AI Agent Hub are competing to become the enterprise's agent control plane. The winner will shape procurement decisions for years.
- Outcome pricing attribution: The first major enterprise contract dispute over AI-attributed outcomes will test whether outcome-based pricing can scale beyond early adopters.
Key Metrics to Track
| Metric | Why It Matters | Where to Find It |
|---|---|---|
| AI-related ACV | Measures how much revenue is tied to AI products rather than legacy seats | Earnings calls: Salesforce, ServiceNow, Workday |
| Net revenue retention | Indicates whether agents are expanding or cannibalizing account revenue | Quarterly filings |
| Agent adoption | Number of customers deploying agents in production, not pilots | Vendor disclosures, earnings calls |
| Usage/action revenue | Share of revenue from consumption or outcomes rather than seats | ServiceNow non-seat ACV disclosure |
| Cloud backlog | Future revenue committed but not yet recognized—a leading indicator | SAP, Oracle, Salesforce filings |
| Enterprise software spending | Gartner's $234B agentic arbitrage estimate provides the baseline | Gartner, IDC, Forrester |
| AI-native software M&A | Acquisition activity signals where incumbents see gaps | Reuters, Bloomberg, CRN |
The enterprise software battle is no longer about which application has the best interface. It is about who controls the agent layer that executes the work.
Salesforce's investor day, Snap's AR glasses partnerships, and the agent governance frameworks from Microsoft, ServiceNow, SAP and Oracle all point to the same conclusion: enterprise software is expanding from applications employees use to platforms that execute work autonomously. The seat is not dead, but its economic centrality is fading. Consumption, actions and outcomes are becoming the new pricing currencies.
Gartner's $234 billion agentic arbitrage estimate is not a prediction of SaaS's death. It is a prediction of SaaS's restructuring. The companies that thrive will be those that design for the new economics—agents, governance, orchestration and outcomes—rather than those that try to preserve the old ones. The question for every enterprise software company is no longer "How do we add AI?" It is "Where in the value chain do we capture economic value when agents perform the work?"
News Sources
- Nasdaq — Salesforce Investor Day Unveils AIforce, Claudeforce and $63B FY2030 Revenue Target (Sept 17, 2026)
- Reuters — Snap targets enterprises with Salesforce, Nvidia AI tools for augmented-reality glasses (Sept 16, 2026)
- RFID Journal — Identiv Completes Sale of IoT Assets, Renames Company INVE Technologies (Sept 18, 2026)
- Gartner — $234 Billion in Enterprise Application Software Spend Is at Risk from Agentic AI (July 1, 2026)
- EyeOn.ai — Microsoft Agent 365 adds Execution Container isolation and centralized agent governance controls (Sept 20, 2026)
- BankInfoSecurity — ServiceNow's New Platform Also Governs Everyone Else's AI (May 7, 2026)
- Constellation Research — SAP CEO Klein: Our AI agents want to solve the complex business processes (Sept 10, 2026)
- DigitalToday — Oracle AI Agent Studio native builder experience (Sept 7, 2026)
- HRTechFeed — Workday Unveils Next Generation of Illuminate Agents (Sept 8, 2026)
- IT Pro — Why enterprise software is becoming harder – not easier – to buy (Sept 15, 2026)
- WSJ — Salesforce Targets Fiscal 2030 Revenue at Over $63 Billion
- CXToday — Salesforce's Outcome-Based Pricing: Strategic Masterstroke or Defensive Scramble Against AI Seat Compression? (Aug 27, 2026)
- IDC — Seat-based pricing predictions
- AlixPartners — Outcome-based software pricing: Hype or reality? (July 29, 2026)
The CODEW Stat
$234 billion of enterprise application software spending is at risk from agentic AI through 2030—roughly 20% of SaaS application spending. Salesforce's Agentforce and Data 360 ARR surpassed $2.9 billion, up 200% year-on-year. ServiceNow's AI ACV crossed $1 billion, with half of net new business now non-seat-based. And IDC predicts seat-based pricing will be largely obsolete by 2028. Stay updated, and thanks for visiting The CODEW. This is Erwin —see you in the next Software Watch stories!
Reviewed by Erwin Castro
on
Monday, September 21, 2026
Rating:
