The sugar-high of the AI rush is fading. The sobering reality: you can’t AI your way to efficiency. Over the past two years, enterprises poured tens of billions into GenAI expecting instant breakthroughs. But when organizational dysfunction and technical debt go unaddressed, AI doesn’t create efficiency, it magnifies inefficiency. Without changing how you organize around value, streamline delivery, and embed technology into the flow of work, AI simply turbocharges broken workflows—accelerating friction, delays, and pain points.
AI doesn’t create efficiency, it magnifies inefficiency.
-Adam Mattis
Fact: Most AI Pilots Don’t Pay Off
A recent MIT study on enterprise GenAI initiatives found that the overwhelming majority of pilots deliver no measurable P&L impact. Only a small fraction scale to meaningful value. The models aren’t the issue—today’s systems are remarkable. The problem is us:
Processes aren’t redesigned.
Organizations don’t adapt.
Foundations—data, architecture, governance—are weak.
Even industry analysts underscore the trend: projects are being abandoned at alarming rates due to poor alignment with real business value. The core lesson: AI is not a strategy. AI is a tool. Unless you build the organizational and architectural muscle to wield it, pilots stall long before production.
Optimize Process. Organize Around Value.
If you build on a fractured foundation, you don’t accelerate progress, you amplify the pain that already exists in the system. To unlock real ROI from AI, do two things first:
Optimize process. Streamline flow of work, align data pipelines and feedback loops, and build continuous improvement into operations.
Organize around value. Structure teams for outcomes, not outputs. Make value visible, measurable, and accountable.
Tools like nVeris are emerging to help leaders visualize and optimize technical and operational value streams. But technology alone won’t close the gap. We need discipline and organizational and technical redesign to achieve what AI promises.
Why SAFe Matters in the AI Era
Organizing around value is not trivial. In the age of AI, SAFe (Scaled Agile Framework) is more relevant than ever. SAFe aligns teams to value streams, enables iterative delivery, reduces time-to-value, and sharpens customer focus while elevating product thinking.
SAFe provides:
Value stream alignment that makes complex programs (e.g., SAP S/4HANA) outcome-driven.
Agile delivery cadence that de-risks large initiatives and compresses timelines—from years to quarters, months to weeks, or even hours, calibrated to context.
Lean-Agile leadership that evolves governance and empowerment alongside complexity.
Example: SAP S/4HANA. Under traditional models, upgrades take years and deliver meager returns. With SAFe at the large-solution level, value is realized months or years faster. Combined AI for data cleansing, training development, and connector creation, and the upgrade becomes a value accelerator instead of a disruption.
Product Models Are Necessary, but Not Sufficient
Product Models are a powerful shift toward end-to-end value ownership. But philosophy alone won’t deliver outcomes. Without reinforced processes and enabling architecture, product models stall, just like most AI pilots.
To make product thinking work in practice:
Modular architecture. APIs, eventing, data enablement, and systems of record designed for scale and change.
Organizational architecture. Teams, governance, and metrics aligned to measurable outcomes.
SAFe as the operating system. Scale the product model, synchronize planning, and connect strategy to execution across the enterprise.
Becoming AI-Native
At SAFe Summit ’25, Andrew Sales laid out a research-backed blueprint for what it really means to become an AI-native organization.
The core message: AI-native is as much about rewiring how organizations think, decide, and deliver, as it is about tools. Right now, most enterprises are stuck in the “AI chasm”. As mentioned in the beginning of this article, MIT research shows that 95% of pilots never scale to meaningful business impact. Why? Because organizations confuse access with fluency, and outputs with outcomes.
Becoming AI-native bridges that gap. It’s both a diagnostic (“where are we today?”) and a roadmap (“what’s next?”). It starts with organizational catalysts that anchor intent and build capability, then moves into execution elements that operationalize AI safely and at scale. The destination is a human-centric culture that unleashes creativity, powered by AI-driven agility on SAFe.
Organizational Catalysts
Every AI initiative needs a spark. At SAFe Summit, Andrew called these organizational catalysts: the conditions that separate enterprises that dabble with pilots from those that build durable impact. These catalysts direct investment, focus effort, and create fluency across the enterprise.
1) AI Strategic Intent & Vision
Too many boards demand an “AI strategy” because competitors have one, not because outcomes are clear. Strategic intent flips this dynamic: define your own AI story with measurable business goals. Intent anchors investment. Without it, enterprises drift from shiny object to shiny object.
2) The AI Money Map
Costs don’t disappear just because tokens get cheaper. Complex workflows can still burn millions of tokens daily. The money map forces discipline: mapping use cases to costs, prioritizing based on ROI, and connecting consumption to outcomes. Without it, AI spend becomes the new shadow IT.
3) Org-Wide AI Fluency
Access is not fluency. Giving every employee ChatGPT doesn’t mean they know when, where, or how to use it responsibly. Fluency means a shared vocabulary, hands-on skills across roles, and the confidence to integrate AI into real workflows. That’s why the AI-Native curriculum launches globally this year, complete with skill badges and partner-led training.
These catalysts aren’t theory. They’re accelerators that close the value gap between experimentation and transformation.
Execution Elements
4) Responsible AI Governance & Ethics
Bias, hallucinations, and data leakage aren’t edge cases—they’re adoption killers. Responsible governance creates trust. Guardrails, audits, and clear education make AI safe to scale without creating brand, regulatory, or customer risk. Ignore this, and adoption collapses under fear.
5) Curated Data
Data lakes don’t make your AI smart. Discipline does. Curated data means context, lineage, labeling, and domain infusion so AI can “know” your business, not just repeat patterns. Without this, models hallucinate, insights mislead, and trust erodes.
6) Operationalizing AI Technology
Pilots are easy. Enterprise-scale AI is not. Moving from ad-hoc tools to integrated platforms means central provisioning, access controls, telemetry, and performance measurement. This is how AI becomes part of the daily flow of work instead of staying stuck in innovation labs.
Together, these elements operationalize intent. They’re how ideas become durable, scalable capability.
Cultural Destination and Agility Outcome
Human-Centric AI Culture
AI should not be about replacing humans, but about amplifying their creativity. An AI-native culture is one where creativity, decision-making, and problem-solving rise to new levels because rote work is automated. As we learned long ago from Kotter, culture change isn’t the starting point, it’s the result of delivering real wins with AI that free people to focus on higher-value work.
AI-Driven Agility with SAFe
Enhancing the operating model is a byproduct of having build the foundations of an AI-Native culture. SAFe becomes the operating system for AI-native agility; this is the overlap of the independent SAFe and AI-Native areas of business at SAI:
PI Planning accelerated with AI-powered insights.
Learning cycles compressed from months to weeks—or even hours.
Strategy modeling augmented by AI, allowing leaders to test options in real time.
Instead of waiting for the next “version” of SAFe, guidance, skills, and frameworks evolve continuously. Trainers and practitioners track capability through skills badges, not static course versions.
AI-native and the supporting IP are intended to serve as a a maturity model, moving organizations from scattered pilots to systemic utilization and deployment of AI as a tool throughout the enterprise. And, as Dr. Kersten discusses in his new book Output to Outcome, it ensures that AI accelerates outcomes, not just outputs.
Five Pillars: A Practical Approach to AI
Frameworks and models are useful, but leaders often ask a simpler question: “Where do I start?” That’s where the Five Pillars come in. They translate the high-level vision of becoming AI-native into a practical checklist, shared understanding, common tooling, optimized processes, value-centered organization, and agility through SAFe. Think of them as the scaffolding that turns intent into execution. Without all five, AI efforts collapse under their own weight.
Shared Understanding (Fluency). Build common language, GenAI, RAG, agents, so teams align faster and act with confidence.
Common Tooling & Guardrails (Governance + Ops). Standardize platforms, safe prompting, data usage, and compliance to prevent brand and regulatory risk while enabling scale.
Process Optimization (Ops + SAFe). AI amplifies whatever it touches; ensure it touches optimized flow and tight feedback loops.
Value‑Centered Organization (Strategic Intent + Money Map). Tie AI investment directly to value streams and measurable outcomes; prioritize with the money map.
AI + SAFe (Agility Outcome). Use SAFe as the operating system that binds strategy, architecture, and delivery—so AI accelerates outcomes at scale.
Without all five operational pillars, AI becomes just another expensive pilot.
-Adam Mattis
This is why SAI developed AI‑Native to sit alongside, and independent of, SAFe. The AI‑Native learning series and operating concepts provide a structured path to fluency, enablement, and leadership, helping enterprises move beyond hype and toward safe, responsible, measurable impact.
In Closing
AI wont fix your problems, but is a powerful lever if we build a system to support it. AI multiplies whatever is already true in your enterprise. Fragmented, siloed, inefficient systems? AI will expose and amplify the cracks. Aligned value streams, disciplined processes, and modular architecture? AI becomes an accelerator of durable advantage.
Don’t skip the fundamentals:
Process without agility is inertia.
Models without context are brittle novelties.
Architecture without alignment is fragmentation at scale.
When enterprises bring these elements together—optimized processes, value‑centered teams, modular systems—the result isn’t hype; it’s compounding advantage.
The future isn’t an “AI strategy.” It’s becoming AI‑native: fluent in agility, disciplined in execution, and relentless about value. SAFe provides the structure. AI‑Native provides the fluency. Together, they form the engine that turns hype into real, lasting advantage.