Enterprise Software Watch: Agentic Platforms, Pricing Experiments, Governance Race Reshape Enterprise Software

Enterprise Software Watch | September 23, 2026

The enterprise software landscape has shifted from agent announcements to agent infrastructure, governance frameworks, and consumption-based monetization.

Agentic Platforms, Pricing Experiments and the Governance Race Reshape Enterprise Software

Market Signal

Enterprise Software Is Being Rebuilt Around Agents, Governance and Consumption

Three weeks after Dreamforce, the enterprise software landscape has shifted from agent announcements to agent infrastructure, governance frameworks, and consumption-based monetization. Salesforce unveiled AIforce, a live interface layer that pulls CRM context into Claude, Slack, and external tools—while introducing Core, Advanced, and Max editions priced at $195, $395, and $550 per seat per month, each bundled with Flex Credits. ServiceNow launched a completely reimagined AI Agent Studio with a visual node canvas and one-click agent creation, and closed its acquisition of Sweep to expand cross-platform agentic capabilities. SAP began production deployment of more than 200 AI agents across finance, supply chain, and spend management. Workday expanded its agentic focus toward complex HR and finance processes while being named a Leader in Gartner's inaugural Magic Quadrant for Workforce Management Technology.

The throughline across all these developments: every major enterprise software vendor is racing to control the agent orchestration and governance layer while simultaneously experimenting with pricing models that decouple revenue from human seats. Gartner estimates that up to $234 billion of enterprise application spending is exposed to agentic AI disruption through 2030—roughly 20% of SaaS spending. The question is no longer whether AI agents will change enterprise software. It is which vendors will control the agent layer, and whether consumption-based pricing can replace seat revenue at scale.

Salesforce's AIforce: CRM Beyond the CRM

Salesforce's most consequential Dreamforce announcement was not a new agent—it was AIforce, a live interface layer that brings Salesforce context, permissions, workflows, and actions into external surfaces including Anthropic's Claude, Slack, Tableau, and custom experiences. The strategy is explicit: Salesforce wants its CRM to work beyond the CRM, embedding its intelligence into whichever interface the user prefers.

The pricing architecture is equally significant. Salesforce simplified its Agentforce Sales, Service, and Industries offerings into three editions: Core at $195 per seat per month with 500,000 Flex Credits, Advanced at $395 per seat per month with 1 million Flex Credits, and Max at $550 per seat per month with 2.75 million Flex Credits. All editions bundle Slack, Tableau Next, Premier Success, and embedded AI agents. The Flex Credit model introduces consumption-based pricing on top of seats—a hybrid approach that keeps the seat as an anchor while monetizing agent usage.

Salesforce also formalized its Trusted Enterprise AI Harness, consolidating governance features from MuleSoft Agent Fabric, Informatica, and Data 360 into a single platform for agent context, security, model management, and auditability. The company says it has delivered 7 billion Agentic Work Units across Agentforce and Slack, including 3.2 billion in Q2 alone.

Why this matters: AIforce is Salesforce's answer to the "SaaSpocalypse" debate. By making CRM an interface-agnostic intelligence layer and monetizing through both seats and consumption, Salesforce is hedging across business models. The company reaffirmed its $63 billion revenue target for fiscal 2030, but the mix—how much comes from seats versus credits versus outcomes—remains the open question.

ServiceNow's Reimagined AI Agent Studio and the Control Tower Strategy

ServiceNow's September 2026 release delivered a completely reimagined AI Agent Studio—a faster path from ideation to deployed agent with a visual node canvas, side-by-side build and test, one-click agent creation from recommended opportunities, and modality-specific agent types for chat and voice. The goal, per ServiceNow's documentation, is to "compress time-to-first-agent for both technical and non-technical builders alike".

The strategic context is ServiceNow's AI Control Tower, which the company has expanded to discover, monitor, govern, secure, and measure AI agents across AWS, Microsoft Azure, Google Cloud, SAP, Oracle, Workday and more than 25 enterprise systems. ServiceNow has also opened its Action Fabric to external agents through MCP servers—Anthropic's Claude Cowork became the first design partner, connecting Claude to ServiceNow's governed execution layer.

ServiceNow's own operational data illustrates the scale of agentic deployment: the company has 20 level-one AI agents "badged" into its own organization, handling 90% of customer-service and employee-service requests. ServiceNow says its level-one agents have generated 2.5 million hours of work from employees. The company raised its 2026 AI ACV target from $1 billion to $1.5 billion, supported by consumption-driven demand.

ServiceNow also closed its acquisition of Israeli AI startup Sweep on September 1, adding cross-platform agent capabilities that work across ServiceNow, Salesforce and HubSpot. The deal, estimated at hundreds of millions of dollars, sharpens ServiceNow's CRM push against Salesforce while expanding its agentic reach.

SAP: From Agent Announcements to Production Execution

SAP moved from Sapphire announcements to production execution with a September 22 webinar demonstrating AI agents handling finance, supply chain, and spend scenarios in SAP Cloud ERP Private. The demos included a financial closing assistant resolving goods receipt/invoice receipt items blocking period close, a batch management agent identifying downstream impact of a supplier miss, and a spend agent guiding employees to contract-compliant buying.

The scale of SAP's agentic commitment is substantial: more than 200 agents across five domains, over 50 Joule assistants and roughly 2,400 Joule skills. SAP CEO Christian Klein has been explicit about the strategy: SAP will release only agents that solve complex business processes, 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 at a Goldman Sachs conference.

Klein acknowledged a structural headwind: 60% of SAP's on-premises customers have not yet migrated to the cloud—a prerequisite for the company's agentic roadmap. SAP's bet is that its data, ontology, and Business Data Cloud provide the context that general-purpose agents lack. The company is also working with Snowflake and Databricks to handle external data.

Workday: Agentic Focus Shifts to Complex HR and Finance Processes

Workday has shifted its agentic AI strategy from volume to value—consolidating simple agents into more capable agents targeting complex HR and finance processes. "We've killed a lot of agents since I've come back, or we've rolled small, simple agents up into bigger, more complex agents," said CEO Aneel Bhusri. Agent adoption is up 35% in a single quarter, with more than 5,500 customers now using one or more Workday agents.

Workday's Sana Enterprise is positioned as an AI workbench for HR, finance, and IT—orchestrating Workday agents behind the scenes to reduce routine service requests. Purpose-built agents include a Talent Acquisition Agent for recruiters and hiring managers, a Payroll Agent for payroll teams, and a Workforce Management Agent that automates schedule changes, time tracking, shift swaps, and demand-based staffing.

Workday was named a Leader in Gartner's inaugural 2026 Magic Quadrant for Workforce Management Technology, and Sana was named a 2026 Top HR Product of the Year by HR Executive. The company is also expanding its Chennai engineering center as a key hub for agentic AI development.

Oracle: Agentic AI Inside the Database

Oracle's approach to agentic AI is distinctive: rather than building a separate agent layer, the company has embedded agent capabilities directly into its database and applications portfolio. Constellation Research's "Agentic AI Washing" report singled out Oracle for having "a unique, differentiated architecture that delivers on agentic AI capabilities across its Oracle AI Database portfolio". The report concluded that Oracle's "breadth and depth of production-ready functionality enable customers to build, deploy, and operationalize AI agents directly against their enterprise data—no AI washing necessary".

Oracle has launched 22 Fusion Agentic Applications across HR, finance, supply chain and customer experience. The company introduced an AI-native builder experience for Oracle AI Agent Studio, supporting no-code, low-code and professional development workflows while inheriting Fusion Applications' security, access permissions, approval and audit trail capabilities. Oracle also launched a Clinical AI Agent for nurses on September 14, expanding its AI documentation tool from physician workflows to inpatient nursing.

Oracle's partnership with Google Cloud brings Gemini models into Oracle AI Agent Studio for Fusion Applications, giving customers access to frontier models within Oracle's governed environment. The company also launched an Oracle AI Database Agent that lets users query databases in natural language without SQL, offered through Google Cloud Marketplace.

Asana's Agentic Work Management: A Category Shift

Asana is repositioning its entire platform around what it calls Agentic Work Management—a category shift rather than an AI add-on. "We have been a company that has really been focused on seats and subscriptions, and now the new Asana is really around both seats and subscriptions and consumption and outcomes," said CFO Aziz Megji. The company's AI product adoption is showing measurable results: 25% of Asana customers with annual spending of at least $100,000 now use an AI product, and those customers have stronger retention and expansion patterns. All three of the company's reported net retention rate cohorts improved, with its largest customers' four-quarter average net retention rising to 98% from 96%. Current remaining performance obligations accelerated to 11% growth from 7% in the prior quarter.

Agentic Work Management brings Asana AI Studio and Asana AI Teammates into a unified core experience. AI Studio and AI Teammates will no longer be treated as add-ons—customers will receive AI Teammates requests and AI Studio credits when they procure Asana. The company is also launching Command by Asana (aimed at development personas), Asana Service Management (IT service management) and Asana Client Management (professional services firms). Asana acquired StackAI in the prior quarter, adding cross-workflow and cross-system orchestration capabilities.

Why this matters: Asana provides a useful mid-market counterpoint to the enterprise giants. The company's early evidence—AI adoption correlating with stronger retention—offers a data point that agentic features may strengthen SaaS economics rather than undermine them, at least in the mid-market.

The Pricing Experimentation Landscape

Every major enterprise software vendor is experimenting with pricing models that blend seats, consumption, and outcomes. No two are identical:

Vendor	Seat Component	Consumption Component
Salesforce	$195–$550/user/month	Flex Credits (500K–2.75M included); $2 per resolution
ServiceNow	Non-seat ACV growing	Consumption-based Assist SKUs; AI ACV $1.5B target
SAP	Advanced Full User Equivalent	AI Units (200 bundled actions ≈ 20 agent runs); $0.40–$1.80/agent run
Workday	Subscription included	Flex Credits (1–750 per agent skill)
Microsoft	$15/user/month Agent 365 standalone; $99/user/month M365 E7	Copilot Credits from $0.01 each
Asana	Seats + subscriptions	AI Teammates requests + AI Studio credits included

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.

The Agent Governance Race

Governance has become the central competitive battleground for enterprise agent platforms. As AI agents gain system access, tool use and computer-use capabilities, the security surface expands dramatically. An agent that can read email, query databases, execute code and interact with enterprise applications is effectively a non-human user with broad permissions—and if compromised, can cause damage at machine speed.

The vendor responses are converging on similar patterns. Microsoft introduced Execution Container isolation in Agent 365 to block unsanctioned shadow agents, along with a cross-vendor agent registry spanning Microsoft, AWS Bedrock, Google Cloud, Databricks Genie and Anthropic Claude agents. ServiceNow's AI Control Tower includes a kill-switch capability that can shut down agents that go off-script, plus real-time monitoring via the Traceloop acquisition. Salesforce's Trusted Enterprise AI Harness consolidates context, governance, security, and model management into a single platform. SAP AI Agent Hub (via LeanIX) gives enterprises a single control plane to discover, govern and monitor agents across all vendors.

The regulatory environment is also tightening. Representatives Mike Lawler and Josh Gottheimer introduced the bipartisan Stop Rogue AI Act on September 15, designed to ensure businesses and federal agencies can see rogue and dangerous AI agents operating inside their own networks and stop them before they cause damage. An EY survey released the same day found that autonomous AI implementation is outpacing oversight, with 26% of respondents whose organizations use agentic AI reporting AI agents operating internally and 36% concerned about risks including data loss and financial damage.

Why this matters: Agent governance is not a feature that can be added later. It is a prerequisite for enterprise deployment. The vendors that solve identity, permissions, auditability, isolation and kill-switch capabilities will win the enterprise agent market. Those that don't will be limited to low-risk experimentation.

M&A Activity Signals Where Incumbents See Gaps

September has been an active month for enterprise software M&A, with acquisitions signaling where incumbents see capability gaps:

ServiceNow → Sweep (closed Sept 1): Cross-platform AI agents working across ServiceNow, Salesforce, and HubSpot. Estimated at hundreds of millions.
Cyera → Oasis Security (completed Sept 3): $1 billion for the leading access management platform for the agentic enterprise—non-human identity security.
Bending Spoons → Miro (Sept 10): $1.355 billion for the workplace collaboration platform, following the $1.285 billion acquisition of Airtable.
Valsoft/TAG Software Group → Square 9 (Sept 8): AI-powered intelligent document processing.
Progress Software → Domo's AI and Data Platform Business (Sept 23): Expanding analytics and AI capabilities.

These deals describe a map of the agentic enterprise stack—from agent orchestration (Sweep) to non-human identity (Oasis) to collaboration surfaces (Miro, Airtable) to document intelligence (Square 9). The pattern suggests that incumbents are buying rather than building capabilities they view as critical to their agentic strategies.

What to Watch Next

Salesforce's FY2027 guidance: The first earnings report covering the AIforce launch will reveal whether the hybrid seat-plus-credits model is gaining traction with enterprise buyers.
ServiceNow's AI ACV trajectory: The company raised its 2026 target from $1 billion to $1.5 billion. Whether it hits that number—and how much comes from non-seat sources—will test the consumption pricing thesis.
SAP's cloud migration rate: 60% of SAP's on-premises customers have not yet migrated. The pace of conversion to S/4HANA and cloud ERP will determine whether SAP's 200+ agent roadmap can reach its addressable market.
Workday's agent consolidation: CEO Aneel Bhusri has killed simple agents and consolidated them into more complex ones. Whether this strategy produces measurable ROI for customers will validate the "quality over quantity" approach.
Agent governance standards: Microsoft's multi-cloud registry, ServiceNow's AI Control Tower and Salesforce's Trusted Enterprise AI Harness are competing to become the enterprise's agent control plane. The winner will shape procurement decisions for years.
The Stop Rogue AI Act: If passed, the bipartisan bill would create regulatory requirements for agent visibility and kill-switch capabilities—potentially accelerating enterprise adoption of governance platforms.
Outcome-based pricing attribution: The first major enterprise contract dispute over AI-attributed outcomes will test whether outcome-based pricing can scale beyond early adopters.
Asana's Agentic Work Management launch: The company's Q3 results will provide the first data on whether bundled AI credits accelerate adoption in the mid-market.

The CODEW Analysis

Enterprise software has moved past the agent announcement phase. The competition is now about orchestration, governance, and pricing models that decouple revenue from human seats.

Salesforce's AIforce, ServiceNow's AI Agent Studio and Control Tower, SAP's production agent deployments, Workday's consolidation toward complex processes and Asana's Agentic Work Management all point to the same conclusion: every major vendor is building the infrastructure layer that controls how agents are created, governed, monitored and monetized. The seat is not dead, but its economic centrality is fading. Consumption, actions and outcomes are becoming the new pricing currencies—and no two vendors are pricing them the same way.

The M&A activity reinforces this thesis. ServiceNow bought Sweep for cross-platform agents. Cyera bought Oasis Security for non-human identity. Bending Spoons bought Miro and Airtable for collaboration surfaces. Progress bought Domo's AI and data platform. These are not feature acquisitions—they are infrastructure acquisitions. The companies that win the next phase of enterprise software will be those that control the layers through which agents operate: data, identity, orchestration, and governance. The application remains important. But the control plane is where the value is migrating.

Sources

Salesforce — Salesforce Expands Agentforce With a New Portfolio of AI Agents Built for High-Value Work (Sept 11, 2026)
Salesforce — New Salesforce Editions Bundle Everything Businesses Need for Agentic Transformation (Sept 3, 2026)
Nasdaq — Salesforce Unveils AIforce, Expands Agentforce at Dreamforce (Sept 17, 2026)
Moor Insights & Strategy — At Dreamforce 2026, Salesforce Goes All In on Agentic AI (Sept 18, 2026)
UC Today — Dreamforce Roundup: 5 Takeaways on AIforce, Slack, and Agent Trust (Sept 17, 2026)
SiliconANGLE — On theCUBE Pod: Salesforce embraces AI agents (Sept 21, 2026)
ServiceNow — Reimagined AI Agent Studio [September 2026 release] (Sept 11, 2026)
Nasdaq — ServiceNow Maps AI Control Tower Strategy as Agents, Security Drive Growth (Sept 10, 2026)
CRN — ServiceNow Sharpens Its Salesforce Attack With Sweep Acquisition (Sept 8, 2026)
DigitalToday — SAP CEO says he will launch only AI agents that solve complex problems (Sept 11, 2026)
SAP — Powering the Autonomous Enterprise with AI Agents and SAP Cloud ERP Private (Sept 22, 2026)
Constellation Research — SAP CEO Klein: Our AI agents want to solve the complex business processes (Sept 10, 2026)
Workday — Workday Named a Leader in Inaugural 2026 Gartner Magic Quadrant for Workforce Management Technology (Sept 17, 2026)
Workday — Sana from Workday Named a 2026 Top HR Product of the Year by HR Executive (Sept 15, 2026)
CloudWars — Workday Agentic AI Targets Big, Complex Processes w/Big ROI (Sept 8, 2026)
Constellation Research — Agentic AI Washing: The Oracle Data Report Card (Sept 1, 2026)
Investing.com — Oracle launches AI agent for nurses to cut documentation time (Sept 14, 2026)
Nasdaq — Asana Unveils Agentic Work Management as AI Fuels Retention and ARR Growth (Sept 21, 2026)
IT Brief — Google named Gartner leader in enterprise AI assistants (Sept 11, 2026)
UC Today — Microsoft Teams & Copilot: What's New in September 2026 (Sept 14, 2026)
TechTarget — AI agent autonomy puts CIO controls to the test (Sept 18, 2026)
EY — EY survey finds that autonomous AI implementation outpaces oversight, yielding an AI governance gap (Sept 15, 2026)
Lawler.house.gov — Reps. Lawler, Gottheimer Introduce Bipartisan Bill to Stop Rogue AI Agents (Sept 15, 2026)
Business Wire — Cyera Completes Acquisition of Oasis Security (Sept 3, 2026)
Business Wire — Bending Spoons enters into a definitive agreement to acquire Miro for $1.355 billion (Sept 10, 2026)
Taiwan News — Progress Software Completes Acquisition of Domo's AI and Data Platform Business (Sept 23, 2026)
IT Pro — Why enterprise software is becoming harder – not easier – to buy (Sept 15, 2026)
Gartner — $234 Billion in Enterprise Application Software Spend Is at Risk from Agentic AI (July 1, 2026)

The CODEW Stat

$234 billion of enterprise application software spending is at risk from agentic AI through 2030. Salesforce has delivered 7 billion Agentic Work Units across Agentforce and Slack. ServiceNow has 20 level-one AI agents handling 90% of its own employee and customer service requests, generating 2.5 million hours of work. Workday's agent adoption is up 35% in a single quarter. And 25% of Asana's largest customers now use an AI product—and show stronger retention as a result.

The CODEW