Re‑Architecting Capability for AI: Governance, SMEs, and the Talent Pipeline Paradox

by Muhammad Zaid Musaddeq

Abstract

AI adoption is accelerating across industries, yet most enterprises remain early in their maturity journey. While organizations are investing heavily, only a small fraction possess the governance, architectural capability, and leadership readiness required to scale AI responsibly. From an EA perspective, AI introduces new governance objects- models, data pipelines, and agent behaviors- that must be managed with the same rigor as any architectural asset. SMEs provide contextual judgment and oversight that governance frameworks depend on. This paper argues that SMEs remain indispensable for sustainable capability development and outlines how enterprises can evolve their EA frameworks, operating models, and governance structures to adopt AI responsibly.

1. Introduction: AI Adoption as an Enterprise Architecture Challenge

Empirical evidence shows that while AI investment is rising, most organizations lack the maturity required to scale it responsibly. Real‑world signals reinforce this gap: reporting from the Los Angeles Times (2025) shows even Stanford graduates struggling to secure entry‑level roles as AI automates junior tasks, while Nartey’s (2025) analysis highlights 85 million jobs displaced globally. These trends expose a weakening talent pipeline essential for future SMEs and governance stewards. From an EA standpoint, AI introduces new architectural risks, and capability demands across business, data, application, and technology domains- requiring a governance‑driven approach to adoption. These realities establish the foundational axioms explored in Section 2.

2. Foundational Axioms for AI Adoption in Enterprise Architecture

2.1 AI Requires Human‑Led Governance

Research from del Rio‑Chanona et al. (UCL, ILO, Oxford) shows that while AI substitutes simple tasks, it struggles with complex, contextual work. McKinsey similarly notes that while employees are ready for AI, leadership and governance lag behind. TOGAF 10 emphasizes that new technology components must be governed through clear accountability, decision rights, and lifecycle controls. Regulatory frameworks increasingly mandate human oversight, reinforcing the architectural principle that technology must remain subordinate to governance.

2.2 SMEs Are Core Components of Architecture Capability

Deloitte’s research shows that AI’s effectiveness depends on human orchestration, contextual interpretation, and alignment with business architecture. SMEs validate AI outputs, ensure ethical use, and maintain alignment with enterprise objectives. In EA terms, SMEs are essential capability enablers embedded within the operating model.

2.3 Expertise Emerges Through Experience, Not Algorithms

TOGAF 10 stresses that architectural capability develops through accumulated experience. AI systems rely on statistical patterns and historical data but cannot replicate tacit knowledge, architectural judgment, or experiential wisdom. The systematic review by Bositkhanova & Dadaboyev reinforces that AI enhances efficiency but cannot replace human expertise in complex decision environments.

Together, these axioms show that AI adoption transforms the EA capability framework itself. Human‑led governance, SME judgment, and experience‑based expertise form the non‑negotiable foundations for scaling AI safely. Without these pillars- governance, skills, and experiential maturity- enterprises risk deploying AI into architectures they cannot control or sustain.

3. The Talent Pipeline Paradox: Architectural Risks of Replacing Junior Roles

A growing trend is the replacement of junior developers, analysts, and architects with AI agents. While attractive from a cost perspective, this shift creates structural risks within the enterprise architecture capability framework. As established in Section 2, AI effectiveness depends on human‑led governance, SME oversight, and experience‑based architectural judgment- capabilities that cannot develop without a healthy talent pipeline.

Junior roles form the experiential base for future architects, engineering leaders, and governance stewards. Eliminating these roles disrupts the capability‑maturity progression emphasized in TOGAF’s Architecture Capability Framework. Without early‑career practitioners gaining exposure to architectural methods, governance cycles, and cross‑domain decision‑making, enterprises lose the human expertise required to supervise AI systems or design and evolve complex architectures.

This creates several EA‑critical risks:

• Erosion of the architecture capability pipeline
• Over‑dependence on AI systems without adequate human oversight
• Loss of institutional knowledge and architectural continuity
• Inability to evolve or govern AI‑enabled landscapes through TOGAF’s iterative governance cycles

Deloitte warns that workforce planning must shift from static substitution models to dynamic human–AI ecosystems. Without a sustainable talent pipeline, enterprises risk creating architectures they cannot govern- undermining both capability-maturity and long‑term resilience.

4. Engineering a Sustainable AI Journey Through Enterprise Architecture

AI adoption cannot be approached as isolated automation choices. As Section 2 established, human‑led governance, SME oversight, and experience‑based capability are foundational to architectural maturity. Section 3 showed that removing junior roles weakens these foundations by disrupting the talent pipeline required for governance cycles and capability evolution. To move forward sustainably, enterprises must strengthen their EA capability across investment planning, operating‑model design, governance structures, and talent development.

4.1 Investment Planning Beyond Short‑Term ROI

McKinsey reports that while 92% of companies plan to increase AI investments, only 1% consider themselves mature. EA provides the structural lens to evaluate AI initiatives not only for cost efficiency but for capability development, risk mitigation, and architectural coherence. Investment decisions must therefore prioritize long‑term capability maturity rather than short‑term automation gains.

4.2 Designing Hybrid Human–AI Operating Models

Deloitte’s research shows that workforce planning is shifting toward dynamic, always‑on orchestration. EA must design operating models where AI augments human roles, SMEs supervise AI outputs, and decision‑making remains human‑led. This aligns with TOGAF’s principle that business architecture anchors all technology decisions and ensures that AI adoption strengthens- not bypasses- governance structures.

4.3 Establishing Governance and Ethical Frameworks

AI introduces new governance requirements and EA must define governance structures that ensure AI systems remain aligned with enterprise values, regulatory expectations, and architectural integrity. This includes embedding AI governance checkpoints into TOGAF’s iterative lifecycle.

4.4 Cultivating the Next Generation of SMEs and Architects

Sustaining architectural expertise requires intentional career pathways that blend AI‑augmented learning, mentorship, and supervised exposure to real architectural work. Without this, enterprises cannot maintain the SME pipeline needed for long‑term governance and capability maturity. EA leaders must therefore design talent models that preserve experiential learning while integrating AI into day‑to‑day workflows.

5. Human‑Critical vs. Autonomous AI Scenarios in Enterprise Architecture

As governance and operating models mature, a key architectural question emerges: which activities require human oversight, and where can AI operate autonomously? Clarifying these boundaries is essential for designing sustainable operating models, allocating SME oversight effectively, and ensuring that AI adoption reinforces- rather than weakens- the enterprise architecture capability framework.

5.1. Scenarios Where Human Architects/SMEs Are Essential

Scenario	Explanation	EA Implication
Contextual reasoning and architectural judgment	AI struggles with ambiguity, cross‑domain tradeoffs, and nuanced interpretation.	Target‑state design, capability modeling, and risk trade‑offs require SME oversight.
High‑stakes governance and accountability	Autonomous AI can introduce compliance, safety, and ethical risks.	Governance checkpoints and lifecycle controls must remain human‑led.
Cross‑domain architectural integration	Humans detect misalignment, bias, and ethical gaps that AI cannot reliably identify.	Enterprise‑wide architecture and strategic alignment require human expertise.
Validation in regulated or sensitive domains	AI generated documentation may contain errors or fabricated content.	SMEs must validate output in finance, healthcare, and public‑sector architectures.

5.2. Scenarios Where AI Can Operate Autonomously (Low‑Risk, Bounded Contexts)

Scenario	Explanation	EA Implication
Routine, deterministic workflows	AI reliably executes structured, rules‑based tasks.	Automate repetitive operational processes.
Low‑risk operational tasks	AI performs well in marketing automation, customer‑service triage, and workflow orchestration.	Use autonomous agents where risk is minimal.
Predictable environments with clear constraints	AI excels when goals and boundaries are predefined.	Apply autonomy to infrastructure scaling and pattern‑based optimizations.

These scenarios show that AI autonomy is viable only in low‑risk, bounded contexts, while human architects remain indispensable for governance, contextual judgment, and cross‑domain coherence. For EA leaders, capability maturity depends on intentionally balancing human‑critical oversight with selective automation- ensuring that AI strengthens architectural integrity rather than eroding it.

6. Conclusion

Enterprise AI will succeed only where strong governance foundations exist. This paper showed that AI adoption is an EA capability challenge, requiring- human‑led oversight, SME judgment, and experience‑based maturity. Section 5 demonstrated that while AI can operate autonomously in low‑risk, bounded scenarios, human architects remain essential for contextual reasoning and governance accountability.

For EA leaders, the mandate is clear: integrate AI through architecture‑driven practices that strengthen capability maturity and preserve the talent pipeline. Sustainable advantage will come from intentional Human-AI symbiosis- where AI amplifies human capability, and human judgment ensures alignment, ethics, and architectural integrity.

References

• McKinsey & Company. (2025). AI in the workplace: A report for 2025.
• Deloitte. (2025). Autonomous workforce planning: Agentic AI and the future of workforce strategy.
• Bositkhanova, N., & Dadaboyev, S. M. U. (2025). Revolutionizing workforce planning: The strategic role of AI in HR strategy. Discover Global Society, 3(100). Springer Nature.
• The Open Group. (2022). TOGAF Standard, 10th Edition.
• Gartner. (2025). AI Engineering: Architecting AI for Scalability, Trust, and Sustainability.
• Nartey, J. (2025). AI Job Displacement Analysis (2025–2030). SSRN.
• del Rio‑Chanona, R. M., Ernst, E., Merola, R., Samaan, D., & Teutloff, O. (2025). AI and Jobs: A Review of Theory, Estimates, and Evidence. ILO/UCL/Oxford.
• Los Angeles Times. (2025, December 19). They graduated from Stanford. Due to AI, they can’t find jobs.