Your Employees Already Orchestrate. Your Company Doesn’t. Why the $2.3 Trillion Digital Transformation Gap Starts with Operating Model Design.

ORCHESTRATE OR OBSOLETE The newsletter for technology and organizational leaders navigating the organizational transformation in the age of AI. Learn how to orchestrate global expertise, measure what matters, and build flexible organizations that win (OrchestrateCo).

March 2026 • Orchestrate or Obsolete Newsletter • Issue #2

Why the $2.3 trillion transformation gap starts with an operating model problem — not a technology one.

80% of your workforce is already using AI tools you didn’t approve (Gartner, 2025). They’re not rebelling. They’re orchestrating — stitching together ChatGPT, personal automations, and workarounds because the official stack doesn’t serve them. Meanwhile, $2.3 trillion has been wasted globally on failed digital transformations (Statista, 2025), and 56% of CEOs report zero measurable ROI from their AI investments (PwC, 2026).

Your people are already composable. They orchestrate agents, external tools, even unofficial teams and pods. Your operating model isn’t. And it should be — not just for AI agents, but for every external capability: human skills, offshore teams, specialist pods. The gap isn’t technology. It’s that no one designed the organisation to let them.

Everyone Is Already Orchestrating (Badly)

Three versions of accidental orchestration are happening in every enterprise right now. None of them are intentional. All of them are expensive.

Your employees are bottlenecked by governance, not by capability. They want to expand their resource pool — bring in a contractor for a sprint, spin up an AI agent for data processing, tap offshore specialists for a surge. But they can’t. IP risk, NDA concerns, no framework. There’s no governance that enables dynamic expansion — only governance that blocks it. If they could orchestrate external capability safely, they would. It’s easier to get resources from outside than from inside. But the organisation isn’t designed to let that happen.

DYNAMIC CAPABILITY BUDGET Here’s the concept that’s missing: a dynamic capability budget on top of headcount. Employees should have a governed orchestration pool — a budget to expand and contract their capability set based on project needs, within an overall envelope. Human, AI, offshore. Scaling up for a three-week sprint, scaling down when it’s done. This doesn’t exist in most organisational designs. It should.

Your vendors orchestrate you. At an EGN Digital Transformation roundtable I recently chaired in Singapore, Maria Singson, PhD — founder of SMEMojo and former IBM Client Success Management lead across APAC — described the vendor side from the inside. She managed 170 CSMs deployed post-sale, whose job was to figure out after the purchase what the client actually wanted to do with the technology. The vendor doesn’t just sell you software. They sell you a roadmap. You conform to their mould.

Single-vendor buying IS orchestration — just the wrong kind. I know because I built it. At a NASDAQ-listed infrastructure firm, your data sat in our system, and that’s how you accessed it. Going with one vendor is a valid capability decision — you’re outsourcing that function to a partner. The problem starts when you hand over the orchestration layer with it. Their roadmap becomes yours. Their pricing changes become your budget crisis — just ask the enterprises hit by VMware’s 10x price hikes post-Broadcom acquisition, or the UK Cabinet Office, which spent £894 million remediating a single vendor dependency (UK National Audit Office, 2024).

“You join their ecosystem, you conform to their mould of what transformation looks like.” — Maria Singson, EGN roundtable, March 2026

The Composable Operating Model for Digital Transformation

So if everyone is already orchestrating — just badly — what does intentional orchestration actually look like? Not a platform. Not a vendor stack. An operating model with three layers. And this applies to your organisation the same way it applies to your software architecture. Manage your company like software: composable, modular, governed by interfaces not hierarchies.

The Capability Layer. Every function is a module with a defined interface — whether it’s a dev team in Ho Chi Minh City, a vendor delivering your CRM, an AI agent running your data pipeline, or a specialist pod on a three-week engagement. The “API” isn’t code. It’s the operating model contract: what the capability does, what it needs, what it outputs, what governance applies. This works for a Python developer the same way it works for a GPT-4 agent. The interface is the same. The module is swappable.

The Orchestration Layer. You own this. This is where build/buy/partner decisions happen. Where “orchestrate the orchestrators” becomes real — your offshore team builds on your codified practices, your AI agents follow your playbooks, your vendors deliver through your integration standards. This is the 5=20 model: a team of 5 delivering like a team of 20, because the orchestration layer multiplies capability, not headcount. And the builds themselves can be executed by external building capabilities — offshore resources augmented by AI, working on your codified practices. You don’t build alone. You orchestrate the build.

The Intelligence Layer. Steven Nunez, Head of AI at vitX, who joined the roundtable, described backward chaining from goal states — start from what you want to achieve, let the knowledge graph compute the path through available capabilities. This isn’t just data infrastructure. It’s the simulation engine that lets you model “what if we swap this vendor” or “what if Vendor X fails mid-delivery on our SAP implementation.” When a vendor changes pricing overnight — as VMware did — the intelligence layer lets you simulate the switch before you’re in crisis mode. That’s the difference between reactive vendor management and strategic orchestration.

THE COMPOSABLE OPERATING MODEL Composable architecture isn’t just a software pattern. It’s an organisational design principle. Gartner projects 60% of enterprises will adopt composable architecture by 2026, with early adopters seeing a 37% TCO reduction. But the framework applies to your org chart, your team structures, your resource allocation. Modularity. Autonomy. Discovery. Orchestration. Your team structure should be as swappable as your tech stack.

“But Aren’t You Just Locking Yourself in the Build Cycle?”

At the same roundtable, Swati Saxena — VP of Global Strategic Client Solutions at Unison Consulting and former Head of D&A Enterprise Applications at the BBC — pushed back with the question every pragmatist is thinking: aren’t you locking yourself into a build cycle if you move away from vendor platforms?

It’s the right challenge. And the answer is: building doesn’t mean building from scratch. It means assembling from components using external building capabilities. Offshore resources augmented by AI, working on your codified practices. The orchestration layer governs the build the same way it governs the buy.

Clorox invested $580 million in a single-vendor ERP transformation and called it a success. That’s because ERP is a standard capability — it doesn’t differentiate Clorox from its competitors. Buy standard. Build differentiating. Orchestrate the integration. The real risk matrix isn’t build vs. buy. It’s controlled orchestration vs. uncontrolled dependency.

The Dynamic Capability Budget

The real barrier isn’t technical. It’s budgetary. Organisations have two levers: headcount and vendor contracts. Nothing in between. But your people’s needs are dynamic. Projects spike and dip. A data migration needs three extra engineers for six weeks. A market entry needs a UX researcher for a sprint. A regulatory deadline needs a compliance specialist for a month. The capability exists — globally, on demand, augmented by AI. The organisation just has no mechanism to access it.

What if every team had a governed orchestration budget — a pool to expand capability dynamically? Bring in an offshore specialist for three weeks. Spin up an AI agent for a data pipeline. Contract a UX researcher for a sprint. All within governance rails: IP protection, NDA frameworks, quality standards, codified practices that travel with the work.

The military calls this mission command: centralised intent, decentralised execution. A 12-person special operations team outperforms battalion-sized conventional units — not through headcount, but through the autonomy to compose capability for the mission. The commander sets the objective and the standards. The team composes capability. No one asks for a headcount increase. The 5=20 model for companies works the same way.

44% of enterprise SaaS licences are paid but never used (Zylo, 2025). That’s not waste. That’s misallocated orchestration budget. The money exists. It’s locked into the wrong capability slots. McKinsey estimates $211–512 billion in potential economic benefits across six APAC markets from SME digital adoption — but only if those organisations retain agency over their technology choices. The Carlyle Group achieved 90% AI adoption across the firm within one year by giving teams autonomy over tool selection within governance rails. That’s a dynamic capability budget in practice, even if they didn’t call it that.

At vitX, we see this daily. We operate as the trust layer between enterprises and Vietnamese technology teams. The hardest lesson: the governance interface is harder than the technical interface. APIs for data are solved. APIs for people — managing IP boundaries, delivery standards, cultural context across borders — that’s where orchestration actually lives. The composable model sounds clean in theory. In practice, it requires codified practices that travel with the work, not with the vendor. That’s what keeps sovereignty with the client.

“Every employee should be able to create their own army of bots and AI agents to work for them.” — Maria Singson, SMEMojo

Three Shifts to Make This Quarter

1.     Audit your capability map. For every function in your organisation, answer three questions: Who controls the interface? Could you swap this module in 90 days? Do the people using this capability have any budget to expand it when demand spikes? If the answer to the third question is “no” for every team, your operating model has a single point of failure — and it’s your own rigidity.

2.    Pilot one dynamic capability budget. Pick one team. Give them a governed pool — even 10–15% of their annual headcount cost is enough to test the model — to orchestrate external capability on demand. Offshore specialists, AI agents, specialist contractors. Set the governance rails: IP protection, NDA framework, quality standards, reporting. Measure what happens to output, speed, and cost per deliverable. You’ll have your proof case within a quarter.

3.    Define your APIs — for people, not just systems. Every external capability you bring in — whether it’s a vendor, an offshore team, or an AI agent — needs a defined interface: what it does, what it needs, what it outputs, what standards it follows. Stop managing external capability through ad-hoc relationships. Start managing it through designed interfaces. The organisations that do this are composable. The ones that don’t are locked in — whether they know it or not.

You’re already orchestrating. Your employees proved that when 80% of them started using tools you didn’t buy. The question isn’t whether to orchestrate. It’s whether your operating model is designed for it — or whether your employees, vendors, and competitors are making that decision for you.

---

What’s the one capability in your organisation that should be orchestrated differently?

Join the conversation on Linkedin

Hit comment and tell me, or hit a DM - I read every response.

Upcoming: The 5=20 Playbook — how to build a team of 5 that delivers like 20, using orchestrated capability across internal teams, global talent, and AI agents.

References List (Expand)

1. Gartner (2025) "Top Strategic Technology Trends: Shadow AI and Employee-Led Innovation." Gartner Research. 2. Statista (2025) "Value of Failed Digital Transformation Projects Worldwide, 2018-2025." Statista Research Department.

3. PwC (2026) "27th Annual Global CEO Survey: AI Investment and Returns." PricewaterhouseCoopers.

4. MIT Sloan Management Review (2025) "Why 95% of Generative AI Pilots Fail to Scale." MIT Sloan.

5. McKinsey & Company (2024) "The State of Digital Transformation: Success Factors and Failure Patterns." McKinsey Digital.

6. BCG (2024) "Digital Transformation: From Aspiration to Outcome." Boston Consulting Group.

7. McKinsey & Company (2024) "Unlocking Success in Digital Transformations: The Role of Organisational Resistance."

8. Gartner (2025) "CIO Survey: Vendor Lock-In and Strategic Technology Decisions." Gartner Executive Programs.

9. Bain & Company (2025) "Digital Transformation Success Rates: A Global Benchmark." Bain & Company.

10. Zylo (2025) "SaaS Management Index: Enterprise Software License Utilisation Report." Zylo Research.

11. IDC (2025) "Worldwide Digital Transformation Spending Guide, 2024-2028." International Data Corporation.

12. Everest Group (2025) "Digital Transformation State of the Market: Success Rates and Key Drivers."

13. Gartner (2026) "CIO Agenda: Geopolitical Factors in Vendor Engagement Strategy." Gartner Research.

14. McKinsey Global Institute (2025) "AI Infrastructure Investment: The $5.2 Trillion Horizon." McKinsey & Company.

15. Flexera (2025) "State of IT Visibility Report: Software License Waste and Shadow IT." Flexera.

16. IBM Security (2025) "Cost of a Data Breach Report: Shadow AI and Unapproved Tool Usage." IBM.

17. Gartner (2025) "Composable Business Architecture: Market Adoption and TCO Impact." Gartner Research.

18. UK National Audit Office (2024) "Digital Transformation in Government: Managing Strategic Supplier Dependency." NAO.

19. European Commission (2025) "Cloud and Digital Sovereignty: NIS2 and DORA Regulatory Framework."

20. Singson, M. (2026) "SMEMojo: RIC Framework for SME Digital Sovereignty Scoring." EGN Digital Transformation Roundtable, Singapore, 3 March.

21. Nunez, S. (2026) "Knowledge Graphs and Backward Chaining in Enterprise Orchestration." EGN Roundtable Discussion, Singapore, 3 March.

22. McKinsey Global Institute (2024) "The Economic Potential of Generative AI and Digital Adoption in APAC." McKinsey & Company.

23. Broadcom (2024) "VMware Licensing Changes: Enterprise Impact Assessment." Multiple industry reports.

24. Microsoft (2024) "European Commission Competition Ruling: $1.12 Billion in Penalties." European Commission Press Release.

25. Carlyle Group (2025) "Enterprise AI Adoption: From Pilot to Full Deployment in 12 Months." Internal case study reported in Financial Times.

26. Clorox Company (2024) "Digital Transformation Case Study: $580M ERP Modernisation." Harvard Business Review.

27. Gartner (2025) "Composable Enterprise Framework: Modularity, Autonomy, Discovery, Orchestration."