McKinsey AI Report: Time to Obliterate Again?

Workforce Transformation via AI

Earlier this week, McKInsey published a comprehensive study on the potential impact of generative AI on the economy. It is worth taking time to read. If the analysis is even ½ right, we'll experience a transformative shock in coming years. Their analysis concludes that Generative AI could add up to $4.4 trillion annually across 60+ use cases. Most of the impact is concentrated in customer operations, marketing, sales, software engineering and R&D – knowledge worker domains that have historically evaded deep productivity gains from automation compared to other company functions. In short, McKinsey is forecasting a revolution in workforce transformation that will unfold over the next 20 years and longer.

While the report acknowledges that pilots are only getting started, with plenty of hard work in front of us before value is realized, this is the first study we’ve seen with bold analysis about the scope of productivity impact from Generative AI. We've seen the impact real-time through our work with OpenAI, Deep Mind and A121 Labs. But we have also observed that broad-based impact will require clever process integration with AI agents deployed in meaningful and trusted roles. What’s difficult about these integrations is the elliptical nature of knowledge work, where core activities like design and analysis can take unexpected routes to a meaningful outcome. The challenge of this next wave of process redesign reminds us of the first wave that hit U.S. companies in the early 1990’s, and the landmark article published by Dr Michael Hammer entitled Reengineering Work: Don’t Automate, Obliterate. It may be time to obliterate again.

Before taking any radical action, we recommend that every company get started with appropriate pilots and narrowly scoped deployments of Gen AI to learn about the opportunities and risks for their environment. Real value and insights can be immediately gained through selective pilots, such as adopting trusted models like Github Copilot for the IT shop, or Jasper for the marketing department. ‘Needle moving’ impact on workforce productivity on the scale cited by McKinsey will take broader actions and investments, including preparation of robust data infrastructures to support customized foundation models and development of consumable functions and workstreams for AI agents.

One thing we do learn from the McKinsey report, obliteration may be the only option.