Every enterprise will have AI agents. The question is whether they operate as a coordinated workforce, or become the next generation of shadow IT. MuleSoft is uniquely positioned to answer that question.
Enterprises are deploying AI agents at pace, but without coordination. Gartner projects 40% of enterprise apps will embed agents by end of 2026. The result: fragmented agent estates across Bedrock, Vertex AI, Copilot Studio, and Agentforce with no unified governance, no visibility, and no way to measure value.
This mirrors the API sprawl challenge MuleSoft solved a decade ago. The difference is the speed: agent proliferation is happening 10x faster than API growth, and the governance gap is already visible.
Most MuleSoft customers have significant dormant estates, APIs and connectors built but underutilised. The shift to agentic architectures creates a natural activation moment: every agent needs APIs, every API becomes an agent-callable tool, and every interaction generates consumption.
The Consumption Value Delivery Manager role sits at exactly this inflection point, turning latent platform investment into active, measurable, recurring value through Flex Credits at $0.10 per action.
The reframing question: What if the biggest risk is not choosing the wrong AI strategy, but having no strategy for governing the agents you have already deployed?
MuleSoft is evolving from connecting systems to orchestrating intelligent digital labour.
Connect apps, data, and devices. API-led connectivity. Build once, reuse everywhere.
Discover, orchestrate, govern, and observe autonomous AI agents across every platform and protocol.
A unified control plane for the agentic enterprise, purpose-built on a decade of API-led connectivity.
Agent Scanners detect and register agents across Agentforce, Bedrock, Vertex AI, and Copilot Studio into the centralised registry.
A context-aware broker that routes tasks to the right agent or tool for reliable multi-step workflows.
Enterprise-grade policies, authentication, rate limits, and compliance controls on every agent interaction.
Visual maps of how agents, systems, and LLMs interconnect, with performance monitoring and flow tracing.
Customer's agent sends a prompt
Registry identifies available agents and tools
Broker routes to the right resources
Policies enforced, identity verified
Action logged, value measured
The full Anypoint Platform is the backbone of Salesforce's integration and automation strategy.
Single platform for iPaaS and full lifecycle API management. 1,500+ pre-built connectors. The foundation everything else runs on.
Curated library of reusable APIs, connectors, and templates. The reason MuleSoft achieves 445% ROI: build once, reuse everywhere.
AI IDE that builds integrations in natural language or agentically via the MuleSoft MCP Server.
Drag-and-drop integration for business users. Pre-built connectors for Salesforce, NetSuite, Workday, and more.
Ultrafast gateway extending Anypoint governance everywhere: on-premises, cloud, edge. Now extended to agent traffic.
Discover, orchestrate, govern, and observe AI agents across every platform. The "air traffic controller" for the digital workforce.
Intelligent routing, cost controls, and policy enforcement across every AI provider from one control point.
Deploy into Docker/Kubernetes on AWS, Azure, GCP, or on-premises. Full control with enterprise-grade resilience.
Purpose-built scripting language for transforming data between any format: JSON, XML, CSV, flat files.
A European financial services company with 200 APIs, 60% dormant. Click each stage to walk through the transformation.
Baseline audit
Dormant to active
Fabric deployment
Policy enforcement
Measurable ROI
Agent Scanners identify 200 APIs across the estate. 120 are dormant, built but not actively consumed. 14 shadow agents are detected running without governance across three cloud providers.
200 APIs catalogued60% dormant14 shadow agents found1 billion+ agents projected by 2029. Without governance, enterprises face the same fragmentation crisis that APIs created a decade ago, but at 10x the speed.
74% of companies cannot track which agents are active, what data they access, or what actions they take.
95% of GenAI pilots fail to deliver P&L impact. The barrier is not the technology, it is the absence of enterprise controls.
Without consumption tracking, agent investments become cost centres rather than value drivers. Flex Credits change that equation.
Agents on Bedrock, Vertex AI, Copilot Studio, and Agentforce operating in silos. No interoperability, no shared context.
Unregistered agents making decisions without audit trails. In regulated industries, this is a board-level risk.
40x growth in four years. By the time you build a governance strategy manually, the agent estate has already doubled.
Integration is table stakes. The differentiator now is who can discover, orchestrate, govern, and observe agents at enterprise scale, across every ecosystem.
| Capability | MuleSoft | Boomi | Workato | Tibco | SnapLogic |
|---|---|---|---|---|---|
| API-led connectivity | Industry originator. 500,000+ reusable assets on Exchange. Full lifecycle management. | Low-code integration with Atom architecture. Simpler but less reuse depth. | Recipe-based automation. Strong for workflow, weaker on API product strategy. | Tibco Cloud Integration with API management. Strong in event-driven, less API-first than MuleSoft. | SnapGPT for AI-assisted pipelines. Growing but limited enterprise footprint. |
| Agent discovery and registry | Agent Scanners detect agents across Agentforce, Bedrock, Vertex AI, Copilot Studio. Centralised registry on Exchange. | No native agent discovery. Relies on manual cataloguing or partner tools. | No agent scanning capability. Focused on recipe and connector library. | No agent discovery capability. Event-driven focus, not agent-aware. | AgentCreator announced. Single-platform, no cross-ecosystem scanning. |
| Multi-agent orchestration | Agent Broker routes tasks dynamically across agents and MCP servers. Topic-based domains with intelligent delegation. | Flow-based orchestration for integrations. No agent-to-agent routing layer. | Workflow orchestration is strong, but designed for app-to-app, not agent-to-agent. | Event-driven orchestration and BPMN support. Not positioned for agentic workloads. | AutoPilot for pipeline automation. No multi-agent broker architecture. |
| Governance and compliance | Flex Gateway for agent-to-agent policy enforcement. Full audit trail, rate limiting, identity verification. | API gateway available. Governance is integration-focused, not agent-aware. | Enterprise security and compliance. Governance is connector-level, not agent-level. | API management via Mashery. Governance is API-focused, no agent policy layer. | Basic API management. No dedicated agent governance layer. |
| Protocol support (MCP and A2A) | Native MCP server support. A2A protocol interoperability. Open standard commitment. | No MCP or A2A support announced. Proprietary connector model. | No MCP support. Strong on pre-built connectors (1,200+) but closed ecosystem. | No MCP or A2A. Focused on messaging protocols (MQTT, AMQP, Kafka). | No MCP or A2A support. Proprietary Snap architecture. |
| Ecosystem breadth | Salesforce + open ecosystem. Works with any cloud, any agent platform, any protocol. | Broad connector library (2,000+). Strong in hybrid cloud but vendor-neutral positioning. | 1,200+ connectors. Strong SaaS-to-SaaS, weaker on-prem and custom. | Strong in event-driven and real-time analytics. Less breadth in SaaS connectors. | Growing connector set. Enterprise adoption still maturing. |
| Consumption pricing | Flex Credits at $0.10/action. Consumption-based model aligned with digital labour economics. | Connection-based pricing. Less predictable for high-volume agent workloads. | Task-based pricing. Competitive for automation but complex at agent scale. | Traditional enterprise licensing. No consumption-based model for agent workloads. | Task-based pricing. Transparent but limited flexibility at scale. |
No competitor has an equivalent to Agent Fabric's four-pillar approach: discover, orchestrate, govern, observe. MuleSoft is not retrofitting agentic capabilities onto an integration platform, it is extending a decade of API-led connectivity into the agent layer natively.
MCP server support and A2A interoperability mean MuleSoft customers are not locked into a single agent ecosystem. Agents built on Bedrock, Vertex AI, Copilot Studio, and Agentforce all participate equally.
The true competitor is the status quo: fragmented agents, no governance, no visibility. Every enterprise will have agents. The question is whether they operate as a coordinated workforce or as the next generation of shadow IT.
Every value conversation needs a factual foundation. Sourced and verifiable.
Model how converting dormant MuleSoft assets drives measurable consumption growth.
A phased approach that delivers measurable value at every stage.
Deploy Agent Scanners. Catalogue the full API and agent estate. Identify dormant assets. Establish consumption baseline and define success metrics with stakeholders.
Prioritise highest-value dormant APIs for reactivation. Deploy Agent Fabric for initial agent orchestration. First Flex Credit consumption begins. Quick wins reported to leadership.
Roll out governance policies via Flex Gateway. Onboard additional agent platforms. Expand value calculator model with real consumption data. Build internal champions.
Full observability across the agent estate. Consumption-based value reporting to C-suite. Identify expansion opportunities. Build the case for next fiscal year investment.
20 years on the supplier side at HPE, Atos, Microsoft, and Salesforce, managing enterprise accounts including Philips across multiple engagements. A career built on translating complex technology into measurable business outcomes for customers who need clarity, not complexity.
At Salesforce, identified EUR 300M in EBITDA improvement potential for a global client, demonstrating the kind of value forensics this role demands. Now transitioning that supplier-side intelligence into consumption value delivery, where the goal is not selling more, but activating what customers already own.
99.5% distinction. Strategic leadership with academic rigour.
Starting May 2026 at University of Hertfordshire. 24-month programme.
AWS Cloud Economics, Google Cloud, GenAI, and agentic AI credentials.
Built the BVF tool: a structured methodology for scoring AI initiative value.
Weekly thought leadership newsletter on AI transformation in the enterprise.
EBITDA improvement potential identified for a global client at Salesforce.
HPE, Atos, Microsoft, Salesforce. Deep knowledge of enterprise buying and selling.
Managed the Philips account directly across multiple supplier engagements.
Given that agent sprawl is accelerating faster than API sprawl ever did, what if the biggest risk for MuleSoft customers is not choosing the wrong platform, but having no governance strategy for the agents they have already deployed? The reframing, from platform selection to agent governance, is where the consumption value story really begins.