Betsy Tong on AI Failures: Why ‘Doing More With Less’ Isn’t Always the Answer

In a recent LinkedIn post, Betsy Tong delves into the common pitfalls organizations encounter when implementing Artificial Intelligence, drawing from the findings of 17 postmortems on AI failures. Tong challenges conventional wisdom, suggesting that issues often lie not with budget, data, or ownership, but with organizational strategy and execution.

Tong highlights a disconnect between executive aspirations and on-the-ground realities. She points out a scenario where:

“The CEO promised big transformation at the all hands. The executives went to Cabo to celebrate AI cost cutting.”

This anecdote, according to Tong, illustrates a fundamental misalignment, leading to a situation where perceived gains are concentrated among a few individuals, while significant Return on Investment (ROI) remains elusive for the broader organization.

Understanding AI Productivity Breakdown

Betsy Tong presents a framework for understanding how AI productivity gains are distributed across different organizational levels. As Tong notes, the impact varies significantly:

• One person: Approximately 40–60%
• A team: Around 20–30%
• A process: Roughly 20–30%, depending on the process
• Cross-company: Highly variable and often unpredictable

Tong suggests that by 2026, as the initial AI hype fades, leaders will increasingly scrutinize the efficacy of their substantial AI investments. She argues that smart leaders are shifting their approach to ensure tangible returns.

Strategies for Effective AI Adoption in 2026

Tong outlines three key strategies that forward-thinking leaders are employing to achieve genuine AI success:

1. Identify and Leverage Heavy AI Adopters

A critical insight from Tong’s analysis is the unofficial adoption of AI tools. She states:

“Nearly 50% of them adopted AI tools IT banned. Find out what is actually working.”

This observation underscores the importance of identifying organic adoption patterns within the workforce. Tong advocates for understanding what tools are proving valuable in practice, even if they bypass official IT channels, to inform broader AI strategy.

2. Implement Parallel Work Streams

To foster innovation and proof-of-concept development, Tong proposes running parallel projects. She explains:

“Today a team of 12-15 gets assigned a project. Tomorrow, create another pod of 3-4. Let the bigger team maintain the existing project. The pod reimagines the result with AI.”

This approach allows for dedicated experimentation with AI without disrupting established operations, providing a clear pathway to demonstrate AI’s potential and build a case for wider adoption.

3. Consolidate AI Tools

Tong identifies tool sprawl as a significant impediment to effective AI utilization. She argues that managing an excessive number of disconnected AI tools leads to cognitive overload for employees.

In Tong’s view, the proliferation of AI tools, particularly in mid-market firms which may use between 100–200 active tools, creates significant challenges. She asserts, “No one can manage a stack of disconnected systems.” The focus, therefore, needs to shift from exploring numerous single-purpose tools to integrating solutions that offer broader capabilities on a unified platform. This consolidation, according to Tong, is essential for overcoming the ‘brain fry’ experienced by employees and streamlining AI integration.

In conclusion, Betsy Tong’s analysis suggests that the era of AI hype is giving way to a phase of practical implementation and demonstrable results. The challenge ahead, as highlighted by Tong, is not about adopting more AI, but about working smarter with the AI that truly delivers value.