Sub-Headline: Why most AI pilots stall before delivering value, and how COOs can turn scattered experiments into an enterprise-scale advantage.
Introduction
AI has moved from buzzword to boardroom mandate. But as COOs and operational leaders rush to embed AI across planning, manufacturing, and distribution, many fall into the same trap: investing in hype without a clear strategy for impact. The result? Disconnected pilots, low ROI, and organizational fatigue. Recent surveys show that over 70% of AI initiatives in operations never progress beyond pilots, despite heavy investment (Source: “State of AI in Operations,” Deloitte, 2025).
This article lays out a grounded approach to building an AI operations strategy rooted in realism: one that drives tangible impact while navigating hype and risk.
The Hype Trap: What It Looks Like
Sattaris’s work with enterprises across industries reveals familiar patterns and symptoms that stall AI programs:
- Pilots launched without clear scope or business ownership.
- Technology teams driving the agenda based on legacy operational assumptions.
- Vendors selling “AI-in-a-box” solutions disconnected from business realities.
- Investment decisions framed by traditional technology and automation approaches.
- Disillusioned stakeholders after six months of “innovation” with no visible outcomes.
- A bottom-up culture where ideas surface organically but lack coherence, synergy, or alignment with enterprise priorities.
Our experience across regions shows the appetite and pathways for AI in operations vary widely, but the barriers remain strikingly similar: siloed thinking, tech-first mindsets, and a lack of shared ownership. In regulated industries like Pharma and MedTech, these barriers are amplified by stringent compliance requirements, while in consumer goods, rapid product cycles can make sustaining AI adoption equally challenging.
Often, COOs rely heavily on their teams to surface use cases organically. While well-intentioned, this approach frequently produces fragmented, non-complementary initiatives, sometimes even with opposing impacts. These efforts are rarely anchored to a cohesive operational strategy.
The Missing Cultural & Structural Foundations
Many large organizations lack the cultural foundation for frontline-led AI development in core operational areas. There’s limited space to ideate, test, and iterate in ways that energize frontline teams. For example, large pharma operations organizations are not geared to enable rapid innovation in quality, manufacturing, and supply chain compared to R&D teams.
Even when pilots are launched, they often enter what Sattaris calls “pilot purgatory”: significant investment in isolated initiatives with no line of sight to scale. While 80% of pharma executives see AI as critical to their future operations, fewer than 20% report having scaled AI projects beyond pilots (Deloitte, 2023). These challenges are compounded by siloed thinking and limited cross-functional collaboration.
Meanwhile, executives hesitate to invest further due to low confidence in ROI. Their mental models are shaped by traditional, deterministic technology investments, an ill fit for the exploratory and probabilistic nature of AI.
Reframing AI as a Business Strategy
To break this cycle, AI must be reframed through the lens of business and operational strategy:
- Start at the Top: A leadership-led approach ensures strategic alignment, overcomes functional silos, and broadens the aperture for investment.
- Anchor to Operational Priorities: Each AI use case should tie directly to enterprise and operations goals, such as service-level performance, supply assurance, or cost profile optimization. AI cannot be treated as a digital strategy; it must be business-led.
- Reimagine Delivery Models: AI offers the opportunity to fundamentally rethink how operations deliver value. Ask: if we could redesign a process from scratch, how would it look with AI embedded from the start?
- Design for Cross-Functionality: Use cases should cut across boundaries (from supply chain to commercial) to unlock compounding value. This can include threading together point use cases into end-to-end processes like Procure-to-Pay or Order-to-Cash.
- Manage AI as a Venture Portfolio: Treat AI investment like venture capital. Diversify across multiple use cases, expect a few high-conversion bets to drive outsized returns, and engage finance partners to support this mindset shift.
- Move Without Perfect Data: Data readiness is critical, but don’t let it stall momentum. One global MedTech leader started with imperfect quality data, piloted a predictive maintenance model, and within six months reduced unplanned downtime by 12%, proving value before investing in full-scale data overhauls.
Framework for a Realistic AI Roadmap
An effective AI roadmap rests on four readiness pillars:
1. Business Readiness
- Reassess the core operating model and business processes.
- Use AI as a lever to reimagine what’s possible and build sustainable advantage.
2. Technology Readiness
- Ensure foundational data infrastructure: availability, quality, and velocity.
- Evaluate platforms, integration models, and architectural scalability.
3. Capability Readiness
- Identify re-usable capabilities (e.g., LLMs, vision systems, pattern recognition).
- Create a strategy to transmit capabilities as use cases scale.
- Distinguish between differentiated vs. standard processes to decide in-house vs. outsourced delivery.
4. Change Readiness
- This shift requires not just capabilities, but the emergence of new leaders inside operations: those with the curiosity, conviction, and cross-functional credibility to own and grow AI solutions.
- Start shaping the workforce of the future, especially in operations functions where the average workforce age may be in the mid-40s.
- Leaders can accelerate this by creating cross-functional “AI guilds” to connect early adopters, share learnings, and sustain momentum.
Organizations should begin with focused use cases and expand horizontally into adjacent domains. For example, a learning platform built for one function can scale to others. Scale dimensions include products, sites, and cross-functional processes. Reusability of capabilities matters, but scaling should be use case-led, not platform-driven.
What Good Looks Like: A Field View
Sattaris’s AI strategy work is informed by our founder’s experience partnering with operations executive leadership to craft an AI strategy for a $50B global pharma enterprise. The approach grounded AI in operational realities, anchoring use cases to tangible priorities like service levels, cost profile, and supply assurance, not just digital ambition. The work involved building a multi-year portfolio of cross-functional initiatives spanning planning, manufacturing, quality, and distribution, tying it to a business-backed roadmap, and proposing changes to the operating model required for scale.
Our approach to AI strategy has evolved: we now design for both the technical blueprint and the organizational readiness required to execute it. That pharma engagement revealed an important insight: while the strategy created the blueprint for value creation, it also exposed the critical gaps in culture, readiness, and leadership conviction required to bring it to life. The foundational mindset shift (from incremental efficiency to transformative reinvention) was still emerging in the organization. Both insights (the path forward and the barriers to remove) proved equally important lessons in building AI programs that scale. That experience now shapes how Sattaris approaches AI strategy: addressing both what to build and how organizations must evolve to sustain it.
Closing Reflection
AI isn’t about automating everything. It’s about fundamentally rethinking how organizations deliver value to customers. It offers a rare opportunity to revisit business models, reshape operating constructs, and rewire how decisions get made and executed. The winners won’t be those with the most tech, but those who use AI to reimagine how their business runs, how their people operate, and how outcomes are delivered—with clarity, purpose, and conviction. If your AI strategy doesn’t start with a business problem, it’s probably a transformational experiment in disguise.
Call to Action
If you’re leading AI in operations today, what’s the biggest barrier you’ve faced moving from pilots to scale? We’d value your perspective—and if you’ve navigated this journey successfully, we’d welcome the opportunity to connect and trade notes on what worked and what didn’t.
Sources:
- “State of AI in Operations,” Deloitte, 2025
- Deloitte, 2023 (Pharma AI scaling statistics)