Since the launch of ChatGPT by OpenAI just over three years ago, artificial intelligence (AI) has permeated various aspects of business operations but has yet to trigger a sweeping transformation across the economy. Despite fervent predictions of imminent disruption, companies are still grappling with challenges that complicate the widespread adoption and effective use of AI technologies.

Within many large organizations, AI tools are increasingly used to streamline routine tasks, such as summarizing meetings, drafting emails, and preparing initial presentation materials. Surveys of chief information officers and chief executives reveal a growing commitment to AI investments, with several studies indicating that firms are moving beyond pilot projects and incorporating AI into core business functions. For example, sectors like retail employ AI for dynamic pricing and personalized recommendations, while private equity firms utilize AI analysts to synthesize data for informed decisions. In manufacturing, computer vision is deployed to identify defects, and software development has seen significant efficiency gains from AI-assisted coding.

However, while adoption is advancing, the broader economic impact remains limited. Experts describe current AI capabilities as a “jagged frontier,” excelling at well-defined tasks such as legal document review, financial analysis, and programming but falling short in areas requiring nuanced judgment, contextual understanding, and tacit knowledge. This uneven performance, combined with the necessity for robust data infrastructures, privacy safeguards, and tailored organizational processes, presents substantial barriers to scaling AI’s potential.

Human factors pose arguably the greatest hurdles. Organizational inertia, risk aversion, and long planning horizons slow the pace of AI integration. Employees wary of displacement may resist adoption, while management often prefers to apply AI to existing workflows instead of undertaking the comprehensive redesigns that could unlock more substantial benefits. For instance, insurance companies might use AI to accelerate paperwork processing rather than reorganizing claims workflows to allow AI to assess damage and approve payments autonomously—changes that could disrupt entrenched routines and hierarchies.

This pattern echoes the historical experience of transformative technologies such as electricity and the internet, both of which took decades to generate widespread productivity gains and alter business practices fundamentally. Observers suggest AI is on a similar trajectory, with a realistic timeline of five to ten years before its full impact is realized. While artificial intelligence is poised to be as consequential as prior major innovations, the path to deep and lasting change will likely be gradual, marked by incremental advances rather than instantaneous upheaval.

As companies and policymakers navigate this evolving landscape, balancing optimism about AI’s promises with a sober understanding of its current limitations appears to be the prevailing approach.