Stop Guessing. Start Scaling: A Realistic Blueprint for GenAI ROI| Anna E. Molosky

For all the noise surrounding GenAI, most enterprises still struggle with the same question: Where does the ROI actually come from?

In my work with global organizations, I have found that the issue isn’t a lack of ambition — it’s a lack of sequence. Companies jump straight into futuristic GenAI initiatives before they’ve optimized the workflows that drive measurable value.

Here’s a practical reset on how to convert GenAI investment into impact that shows up on the balance sheet.

1️⃣ ROI Doesn’t Start at the Front Door — It Starts in the Back Office

Executives often begin AI investment in the most visible places: digital marketing, customer interactions, sales enablement. But visibility doesn’t equal value.

Real, dependable ROI comes from operational heavy lifting:

• Finance close cycles

• HR transactions

• Procurement workflows

• Shared services

• Reconciliation, validation, and compliance routines

These aren’t glamorous projects — they’re profitable ones.

AT&T’s enterprise automation program, for example, saved 16.9 million manual minutes annually and generated 20x ROI by focusing on process-level efficiency, not glossy GenAI showcases.

Anna E. Molosky perspective:If your AI strategy doesn’t touch the back office first, it will never scale sustainably.

2️⃣ Your Employees Are Already Building Your AI Strategy — Quietly

Organizations often design AI roadmaps top-down. The problem? That ignores the thousands of bottom-up experiments already happening across the workforce.

The data is consistent:

• 90% of employees use personal AI tools

• Only 40% of organizations provide sanctioned AI access

This “shadow AI” usage is not a threat. It’s a diagnostic tool.Employees are telling you:

• Which tasks eat time

• Which workflows are repetitive

• Which processes are prime for automation

• Where AI genuinely accelerates productivity

Instead of guessing where value is created, collect and analyze this behavior. Your workforce has already pressure-tested the highest-value use cases. All you need to do is formalize and scale them.

3️⃣ Enterprise AI Success Is a Marathon, Not a Momentum Play

One of the biggest misconceptions I see is the assumption that AI ROI should appear on a six-month timeline.That’s not how enterprise transformation works.

Implementing global-scale technology — including AI — typically spans 1 to 3 years, influenced by:

• Integration depth

• Data hygiene

• Business process redesign

• Governance and risk controls

• Global training and adoption

Short evaluation windows mislabel “in-progress” as “ineffective.”

The often-cited “95% AI failure rate” isn’t a warning sign — it’s a maturity indicator. The 5% generating ROI today represent organizations that sequence investments correctly and manage expectations realistically.