Vendors are racing to become the central AI hub, and organizations are adopting multiple tools. The question is no longer which AI to deploy. It is how to coordinate agents across the enterprise to ensure control, governance, and risk mitigation. It also raises the fundamental question of who should own that control as agents spread across business processes.
The debate around AI orchestration is intensifying. And not just in technology circles. Within CRM and CX strategy discussions, a clear tension is emerging:
Should organizations rely on cross-enterprise AI platforms (such as Microsoft Copilot, Google Gemini or IBM watsonX), or on CRM-specialised AI (as Salesforce Einstein) to drive value?
The discussion has entered a particularly intense phase:
- A recent Barron’s CIO survey reports that 76% of CIOs are already using Microsoft Copilot across enterprise environments. This shows how horizontal AI tools are scaling rapidly.
- Salesforce has expanded support for the Model Context Protocol (MCP) and third-party integrations. With the Agentforce Command Center, it positions Agentforce as a coordination hub for AI agents across systems.
- Market attention is high: My recent LinkedIn post on this topic drew over 130,000 impressions, 600 likes, and 100 comments: This is clear evidence that people are actively searching for clarity.
What often goes unchallenged in this debate is an implicit assumption: That even as AI agents multiply across functions and processes, they should ultimately be controlled by a single platform or central authority.
With so much noise, I felt the need to slow the conversation down and take a step back for a calmer and more balanced reflection. That is the purpose of this newsletter.
Why Are We Talking About AI Orchestration Now?
Many organizations have 20 pilots, 7 disconnected AI tools, no clear ROI, and a number of unlisted shadow AI usages. This is today’s reality. AI is increasingly distributed across models, agents, workflows, and systems. Uncontrolled.
In this context, the orchestration and governance of AI agents becomes unavoidable.
AI orchestration is about how organizations coordinate multiple AI agents. It is not simply about deploying more AI. It is about controlling how intelligence flows across the enterprise.
In practice, orchestration involves:
- Integration and interoperability: Coordinating multiple AI agents, tools, databases, APIs, and systems working together or in parallel.
- Context continuity Preserving the context (data, history, state) across across interactions and systems.
- Automation: Enabling dynamic handoffs to automate end-to-end processes and manage the flow of tasks between components. Users no longer need to switch between multiple tools which is a frequent source of frustration in many projects.
- Governance, monitoring, and lifecycle management: Enforcing governance, data security, and tracking models, versioning, compliance, access controls, and performance over time.
As AI moves deeper into customer journeys, its value depends on how well CRM AI interacts with ERP, finance, HR, and supply chain AI systems. This, thereby, influences end‑to‑end customer experience delivery.
For regulated sectors (finance, healthcare), where compliance and context matter most, this integrated model aligns AI insight with audit-ready execution pipelines.
But, as AI agents move closer to execution, orchestration becomes an ownership question as much as a technical one.
Who is accountable when AI acts: A central AI platform, or the business leaders responsible for the process and its outcomes? or both? A RACI approach?
This is where many AI strategies become fragile. As AI scales, orchestration introduces strategic trade-offs around platform choices, integration complexity, context preservation, governance, scalability, and ultimately, ownership and customer outcomes.
Executive Perspective: Why It Matters Now
The real question for C-level executives today is not “Which AI is better?” any longer. It is where AI should live, how it should be coordinated, and who owns AI-driven decisions when they sit inside specific business processes.
This is a matter of strategic significance and the orchestration of multiple AI agents is becoming a board-level topic.
Because, at the end of the day, the legal and financial liabilities remain very human. Strictly.
And for board governance, it presents really complex challenges. Because it directly influences:
- Enterprise agility and cross-functional coordination,
- Operational risks and resilience,
- Customer experience delivery,
- Organizational transparency and governance.
The challenge is no longer simply governing AI systems, but governing decision rights, accountability, and risk in a world of autonomous agents.
Why This Is Relevant for CRM and CX
CRM AI tools influence frontline experiences (sales reps, service agents, and marketers). Tools that misinterpret context or provide inaccurate advice can degrade user adoption and customer outcomes, not improve them.
In this context, CRM and CX leaders are increasingly becoming owners of AI agents embedded in their processes. They may gain accountability for customer impact, compliance, and risk by integrating AI agents directly into processes.
This raises a crucial question:
- Should AI remain confined to siloed productivity gains, or
- Should it become an integral part of end-to-end enterprise workflows?
Vendor Perspectives: Competing to Be the Hub
Vendors are increasingly competing to become the central point of AI orchestration within the enterprise. Each aims to position its platform as the “hub” through which other AI tools, workflows, and data should flow.
AI orchestration for vendors is about interoperability (connecting components), context (understanding and preserving state), workflow continuity (coordinating steps across tasks), and governance, monitoring, and lifecycle management.
At the heart of the debate is a structural technological choice:
- Horizontal platforms offer broad reach and governance but are often too generic.
- CRM-specialized AI provides deep context but can be siloed.
Some vendors leverage broad enterprise reach, embedding AI across productivity and operational systems to become the default orchestration backbone.
Others focus on deep CRM and customer-context intelligence, aiming to be the central source of insights for sales, service, and marketing workflows.
For example:
- Microsoft Copilot integrates broadly into Outlook, Teams, and Excel, offering reach and adoption. For organizations who have largely invested in Microsoft 365, this is compelling. But this approach may also lack the deep CRM context that is needed, and risk too generic recommendations.
- Salesforce, by contrast, embeds AI directly into CRM workflows. This enables domain-specific insights and automation. The risk is not quality but fragmentation if CRM AI remains disconnected from enterprise processes.
What Executives Should Pay Attention To
No single vendor today fully combines enterprise-wide coordination with deep domain and customer context.
As a result, most organizations will need to operate with more than one AI hub. This is one of there reasons we are talking of AI ecosystems.
But coordination does not automatically imply ownership.
Platforms may orchestrate AI agents, but accountability for AI-driven decisions may sit with process owners who understand business context, regulatory exposure, and customer impact.
In practice, the most effective strategies will likely admit this reality and combine platforms rather than forcing a single AI layer to do everything. They will also clearly define who controls, supervises, and is accountable for AI behavior.
5 questions for executive teams:
- How do we regularly integrate AI into board-level discussions on strategy, risk, and performance?
- How do we select and supervise partners within a growing AI ecosystem?
- Which AI decisions should be centrally governed, and which should be owned by process leaders?
- How do we assign accountability when AI agents operate across multiple processes and platforms?
- How do we govern autonomous, multi-agent AI systems operating across the organization?
Key Takeaways
- AI orchestration is a business transformation imperative, not just a technology decision. It is a question of ownership and accountability.
- Cross-enterprise AI platforms serve broad purposes and can accelerate adoption but need CRM context to unlock customer value. CRM‑specialized AI remains vital for deep personalization and customer insights, but must integrate with enterprise AI workflows to avoid siloed automation.
- Vendor lock-in and ecosystem silos: Heavy reliance on one AI vendor can create long-term constraints.
- Governance & compliance: Governance, data integrity, and risk management must be designed in from the start, as orchestrated AI amplifies both value and risk. Even with MCP and enterprise controls, configuring permissions, audit trails, and compliance workflows at scale can be non-trivial and resource-intensive. This requires specialized governance expertise.
- Value measurement: AI ROI remains elusive when tools are not tied to end-to-end outcomes.
- Executive alignment across CIO, CMO, COO, and CX leadership is critical to scale AI responsibly, and steer investment across AI capabilities and orchestration models.
Final word
The real work for executives is how they transform their organization around AI. Strategically, responsibly, and with data and people at the centre.
The hardest decision may not be which AI platform to choose, but how to distribute control and accountability as AI agents increasingly act inside the business.