Most Enterprise Agentic Projects Are Doomed, Here's Why — Jess Grogan-Avignon & Jack Wang, Accenture

2026-05-28 · Source: AI Engineer · Field: Business & Management — Corporate Strategy & Leadership, Operations & Process Management, Project & Product Management · Depth: Intermediate, long

Summary

Accenture's Jess Grogan-Avignon and Jack Wang identify five "enterprise tensions" that impede most agentic AI projects, where traditional human-paced organizational structures clash with machine-speed AI. Their research shows only 12% of companies achieve "AI achiever" status, with 80% stuck in costly piloting. Enterprise Speed, the first tension, highlights how legacy processes like security reviews bottleneck AI development; an agentic app built in two weeks took 12 months to deploy. Enterprise Finance struggles with AI's emergent nature, demanding upfront certainty incompatible with iterative development, despite AI achievers seeing 50% higher revenue growth. Agentic Delivery is often mismanaged by treating non-deterministic AI like traditional software, necessitating a shift to hypothesis-driven, statistical confidence building. Engineer for Trust emphasizes building confidence in AI outputs through progressive autonomy, moving from shadow mode to controlled actions based on outcome evidence. Lastly, Your Moat argues true competitive advantage lies not in static enterprise knowledge but in "living memory"—unique customer interaction signals driving rapid iteration and value creation.

Key takeaway

For CTOs and AI/ML Directors aiming to scale agentic AI, recognize that traditional enterprise processes are a critical drag. You must upgrade governance and engineering automation for machine speed, treating every human process as executable code. Shift finance to a VC-like portfolio approach for AI investments, embracing emergent value over upfront certainty. Engineer for trust using progressive autonomy, and cultivate your "living memory" from customer feedback as your unique competitive moat.

Key insights

Enterprise AI success requires adapting organizational speed, finance, delivery, and trust-building to AI's emergent, machine-speed nature.

Principles

• AI value is emergent, not predictable upfront.
• Trust in AI is built progressively with evidence.
• Feedback loops are the ultimate competitive moat.

Method

Implement progressive autonomy: start with shadow mode, then advisory, then controlled autonomy, each step gated by evidence in outcomes and building confidence in target behaviors.

In practice

• Automate every human process into executable code.
• Fund AI as a portfolio of bets, like a VC.
• Reshape programs around hypothesis-driven delivery.

Topics

• Agentic AI
• Enterprise AI Adoption
• AI Governance
• Progressive Autonomy
• AI Investment Strategy
• Feedback Loops

Best for: Executive, AI Product Manager, Director of AI/ML, VP of Engineering/Data, CTO

Related on AIssential

• #205: AI Labs Refocus on Agents and Enterprise, Trump’s New AI Framework, Meta’s Rogue Agent & What 81,000 People Want from AI — AI labs are intensely racing to develop autonomous agents and secure enterprise adoption, fundamentally reshaping the AI landscape.
• Your Company Passed Its Tipping Point. Leadership Didn't Feel A Thing. — Apathetic leadership and "overweight" legacy businesses are succumbing to agile, agentic "micro-unicorns" at an accelerating, non-linear pace.
• Agentic Explainability at Scale: Between Corporate Fears and XAI Needs — Agentic AI adoption at scale creates "Agent Sprawl" risks, necessitating robust explainability and governance.
• From AI promise to business impact: building future-ready enterprise AI — Scaling enterprise AI requires addressing data quality, architectural complexity, and realistic expectation management.
• What it takes to scale agentic AI in the enterprise — Scaling agentic AI demands robust governance, data infrastructure, and system integration, not just advanced technology.
• If more companies block bots and dial down bloat, AI adoption doesn’t collapse. It matures. — AI adoption is shifting from ubiquitous integration to constrained, trust-centric deployment, prioritizing governance and utility.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Engineer.