Scaling Generative AI With Confidence: Anna E. Molosky’s ROI-Driven Blueprint

 

Across global enterprises, I continue to see a consistent pattern: the barrier to Generative AI success isn’t vision — it’s sequencing.

Organisations are eager to deploy cutting-edge, high-visibility GenAI initiatives. But when those initiatives are launched before foundational workflows are stabilised, the results are predictable. The technology looks impressive. The financial statements, however, remain unchanged.

If enterprises want measurable, repeatable returns, they need a reset — one grounded in operational reality rather than experimentation hype.

1️⃣  Real ROI Starts in the Back Office

Executive teams often prioritise AI investments in customer-facing functions: digital experiences, sales acceleration, and engagement tools. These areas offer visibility — but visibility alone does not generate sustainable value.

The most reliable ROI consistently emerges from operational core functions, including:

• Financial close and reporting cycles
• HR case management and transactional workflows
• Procurement and sourcing operations
• Shared services functions
• Reconciliation, validation, and compliance routines

These processes may not be headline-grabbing — but they are economically meaningful.

Consider AT&T. By focusing on structured process automation before pursuing advanced GenAI experimentation, the company eliminated 16.9 million manual minutes annually, achieved a reported 20x ROI, and generated hundreds of millions in recurring value.

The lesson is straightforward: if an AI strategy doesn’t begin with operational fundamentals, it will struggle to scale or produce predictable returns.

2️⃣  Employees Are Already Mapping High-Value Use Cases

Many organisations still treat AI strategy as a purely top-down initiative. In doing so, they overlook a powerful source of insight: their own workforce.

The data tells a clear story:

• Roughly 90% of employees use personal AI tools to increase productivity
• Only about 40% have access to approved, enterprise-grade AI platforms

This “shadow AI” activity is often framed as a compliance risk. In reality, it’s strategic intelligence.

Employees are already identifying:

• The most time-consuming tasks
• Repetitive workflows ripe for automation
• Friction points across daily operations
• Practical scenarios where AI improves outcomes

Instead of guessing where ROI might exist, leaders should observe and formalise these organic behaviours. The highest-value use cases are often already being tested at the edges of the organisation. The opportunity is to secure, standardise, and scale them.

3️⃣  Enterprise-Scale AI Matures Over Years, Not Quarters

Another recurring miscalculation in GenAI adoption is compressed timelines. Six-month ROI expectations rarely align with enterprise complexity.

Large-scale AI transformation typically requires one to three years, shaped by:

• Systems integration across legacy platforms
• Data cleanup and harmonisation
• Process redesign and standardisation
• Governance and risk controls
• Workforce enablement and adoption

When initiatives are evaluated too early, normal maturation phases are misinterpreted as failure.

The often-cited 95% AI “failure rate” reflects this misalignment. The minority of organisations seeing returns aren’t necessarily more innovative — they are further along in sequencing, governance, and realistic expectation-setting.

A Practical Blueprint for Durable GenAI ROI

Enterprises that successfully scale Generative AI tend to follow a disciplined progression:

• Begin with structured, high-certainty back-office automation
• Leverage employee-driven experimentation to identify validated use cases
• Set timelines that reflect the complexity of enterprise transformation

When sequencing is intentional, GenAI shifts from speculative experimentation to a dependable driver of operational efficiency and financial performance.

With the right order of execution, AI doesn’t just look transformative — it becomes measurably so.