Enterprise technology leaders face a stark reality in early 2026: AI agents promise transformative efficiency, yet most initiatives languish in ‘pilot purgatory’ without robust enterprise resource planning integration. A McKinsey report published January 9 warns that resources poured into generative AI are starving core ERP capabilities, with only 40% of companies reporting enterprise-level EBIT impact from AI, and most citing less than 5% gains.

“Resources focused on AI are coming at the expense of enabling ERP to provide the system capabilities AI needs to thrive,” write McKinsey partners Bjørnar Jensen, Darwin Deano, Michael Allison, and associate partner Talha Bin Asad. ERP systems, often dismissed as legacy baggage, house a company’s ‘operating DNA’—process knowledge, data structures, and business logic essential for scaling AI agents in high-volume transactions.

AI’s projected $17 trillion to $26 trillion global economic impact hinges on this synergy, yet 80% of firms using gen AI in at least one function struggle with tangible outcomes, per McKinsey’s November 2025 analysis.

The Investment Chasm Widens

IT budgets tell the tale: Nearly half of organizations plan gen AI investments, while core infrastructure and architecture funding plummets. This ‘great divide’ breeds unsupported experiments, as AI agents falter without ERP’s end-to-end workflows. High performers achieving 5%+ EBIT impact redesign workflows, scale rapidly, and bet big on transformative AI, according to the same McKinsey study.

In manufacturing, AI-enhanced ERP delivers 30-40% efficiency gains via autonomous production scheduling and supply chain optimization, reports Top10ERP.org on January 3. SAP’s Sapphire 2025 updates evolved its Joule copilot into an autonomous agent with Joule Studio for skill-building, signaling vendor maturation.

Oracle NetSuite’s SuiteWorld 2025 unveiled NetSuite Next and Autonomous Close, embedding agentic AI into finance and operations for continuous workflows, per ERP Today December 29.

Vendor Playbooks Emerge

McKinsey outlines a six-step playbook: Clarify workflow value by mapping ERP data, transactions, and rules to outcomes like margins; decide buy-versus-build, favoring standardized agents from SAP or custom edges; define shared ontology grounded in ERP; embed agents in steps like approvals; balance flexible architectures with stable ERP cores via APIs and orchestration; and refine via ‘value mission control’ dashboards.

SAP’s tools like Business Data Cloud and Signavio exemplify this, enabling cross-functional ‘agent squads.’ Celonis aids process mining. Meanwhile, Gartner’s forecast sees 40% of business applications embedding task-specific AI agents by 2026’s end, up from under 5% in 2025, as noted in ERP Software Blog December 26.

Microsoft Dynamics 365 advances with the ERP Model Context Protocol (MCP) server, shifting from static actions to dynamic frameworks adapting to business needs, announced at Ignite 2025 per the Microsoft Dynamics 365 Blog December 9.

Domain Transformations Take Hold

High-impact domains like finance and operations demand holistic change over isolated use cases. McKinsey charts core ERP realms—finance, supply chain, HR—versus partial or non-core areas, urging domain focus for synergies in data, tech, and change management. AI agents extend ERP in ‘long-tail’ exceptions like dynamic inventory or intelligent sourcing.

Infor and Oracle deployments signal 2026 architectures for autonomous multi-step workflows, not mere chatbots, transforming manufacturing ERP into systems of autonomous operations, writes ERP Today January 6. Agentic AI slashes unplanned downtime via predictive maintenance.

IFS leverages TheLoops’ Agent Development Lifecycle for asset-intensive orchestration, while Workday augments HR with personalized paths under human oversight, per AIMultiple research.

Governance and Risks in Focus

Scaling demands ‘human-in-the-loop’ for high-stakes decisions, robust data controls, and testing amid integration risks like model drift. McKinsey’s rule: $3 in change management per $1 in model development. Constellation Research predicts 2026 ‘skirmishes’ over data access fees as agents proliferate, per its January 4 trends report.

Forbes contributor Patrick Moorhead notes in an August 2025 piece on event-driven agentic ERP that future systems evolve from reactive to dynamic platforms autonomously executing workflows. Oracle’s Fusion AI Agent Marketplace deploys no-code agents, handling exceptions in real-time supply chains.

Aaron Levie, Box CEO, tweeted July 2025: “The last mile of making AI Agents work in real, highly variable and hostile environments, is insanely hard. And increasingly it’s the most valuable part.” X discussions echo integration as key, with Rimini Street touting 10-20% cost cuts via agentic AI on existing ERP.

2026 Roadmaps Accelerate

SAP’s January 12 news center outlines five AI themes, including relational models like SAP-RPT-1 for ERP forecasting and agentic governance for hundreds of specialized agents. ERP Today December 11 cites Versori’s whitepaper: 100% of 20 senior ERP leaders prioritize AI automation, 82% see it driving innovation.

Microsoft Research envisions 2026 agentic systems collaborating as ‘digital coworkers,’ per ERP Today December 30. The Linux Foundation’s late-2025 Agentic AI Foundation pushes standards, addressing 2025’s interoperability leaps via Anthropic’s Model Context Protocol.

Enterprises must elevate ERP in AI strategies, tying CIO/CTO initiatives to P&L. As McKinsey concludes: “Unlocking AI’s full potential requires treating ERP not as legacy baggage but as a key enabler that makes intelligence scalable, safe, and valuable.”