Enterprises are pouring billions into Generative AI, yet only a small fraction can point to sustained, material financial returns. According to Anna E. Molosky, the issue isn’t that GenAI fails to deliver—it’s that organizations invest out of sequence, pursue the wrong initial problems, and expect value to materialize on unrealistic timelines.
What follows is a pragmatic blueprint for converting GenAI spend into measurable, repeatable ROI.
Many executives default to high-visibility AI use cases—customer experience, marketing personalization, sales enablement—expecting rapid payback. In practice, these initiatives often deliver fragmented value and limited scalability.
As Molosky emphasizes, the earliest, clearest, and most defensible ROI from AI consistently emerges in operations.
High-impact starting points include:
Repetitive back-office workflows
Finance and HR process automation
Data-intensive tasks suited to structured AI or rules-based automation
Systems that integrate cleanly into existing enterprise architecture
This approach creates compounding efficiency rather than isolated wins. AT&T’s enterprise automation program illustrates the point: by focusing on process-level optimization, the company eliminated 16.9 million labor minutes annually and achieved a 20x ROI—well before GenAI became mainstream.
Molosky’s core principle:
If ROI is the objective, begin where inefficiency is measurable, repeatable, and expensive.
While leadership debates AI strategy, employees have already moved ahead:
90% of employees use personal AI tools
Only 40% are given sanctioned, employer-provided access
Rather than viewing this “shadow AI” as a governance problem, Molosky sees it as a high-value diagnostic signal. Employees have effectively conducted thousands of real-world experiments, revealing exactly where AI reduces friction and accelerates output.
To harness this insight:
Observe how AI is already used in day-to-day workflows
Identify high-frequency, high-impact tasks
Standardize and secure these workflows within enterprise-grade platforms
This bottom-up data provides leaders with a proven, evidence-based roadmap for prioritizing AI investments—eliminating guesswork and accelerating time to value.
Instead of asking where AI might work, let your workforce show you where it already does.
One of the most common obstacles to AI ROI is unrealistic timing. Many organizations still expect GenAI to deliver returns on a pilot timeline—three to six months.
That expectation ignores how enterprise transformation actually works.
Large-scale AI deployments typically span one to three years, shaped by:
Data quality and readiness
Legacy system integration
Business process redesign
Security, risk, and governance controls
Global change management and adoption
Short-term ROI assessments capture only a small slice of the value curve.
The frequently cited statistic that “95% of AI initiatives fail” is often misinterpreted. Molosky reframes it as a maturity indicator, not a warning. The 5% generating ROI today are simply further along the transformation curve—demonstrating what disciplined enterprises will achieve at scale tomorrow.
To turn GenAI from a cost center into a profit engine:
✔ Redirect early investment toward operational automation
✔ Use employee behavior to guide use-case prioritization
✔ Set expectations that reflect enterprise deployment realities