McKinsey and Cognition are trying to prove AI coding pays off - Business Insider

McKinsey and Cognition are trying to prove AI coding pays off - Business Insider

McKinsey & Company and Cognition Partner to Showcase AI Coding Benefits

Under a partnership announced on Monday, McKinsey & Company will guide clients in identifying engineering projects suitable for AI integration and redesigning team structures and workflows accordingly. Cognition will provide its AI coding agent, Devin, capable of understanding, planning, writing, and testing code.

The launch of Devin in early 2024 sparked concerns among software engineers about AI tools potentially disrupting their workflows and even replacing them. Since then, development practices across industries have significantly evolved, with engineers shifting from writing code to directing AI output.

AI Coding Tools Evolve

According to Lari Hämäläinen, a senior partner at McKinsey & Company, the capabilities of AI coding tools have advanced considerably in the past year. Previously, these tools were primarily useful for specific tasks, such as providing technical guidance or syntax details. Now, more advanced AI models enable longer, more autonomous coding tasks.

However, this rapid technological progress presents a challenge for companies adopting AI tools. Hämäläinen notes that organizations might struggle to keep up with the pace of change, as the technology evolves faster than their ability to adapt their processes.

Scaling AI Implementation

One of the challenges in implementing AI tools at a larger scale is scaling them across extensive developer teams. While it's relatively simple to have a small team experiment with an AI tool, scaling it across thousands of developers requires careful management.

Measuring AI Payoff

The quest for AI's benefits and cost management has led several large companies, including Goldman Sachs, Cisco, and JPMorgan, to become more selective about AI tool access and usage.

Art Levy, VP of global partnerships at Cognition, explains that Devin leverages multiple AI models, allowing for the use of more expensive frontier tools for complex engineering tasks and cost-effective models for simpler tasks. He emphasizes the importance of choosing the right tool for the job, comparing it to not needing a luxury car for a trip to the grocery store.

Devin also aids in accelerating large software modernization projects that companies have previously postponed. Levy cites a client's example where a five-year enterprise resource planning software update is now planned for completion within 18 months. The focus has shifted from simply completing the project to doing so efficiently and cost-effectively.

Leave a Reply