Walmart Limits Use of Code Puppy to Avoid Redundancy

Walmart has recently decided to limit the use of its AI-based coding tool, Code Puppy, which was developed for employees at all levels. The growing popularity of this tool has prompted the company to impose restrictions on token usage, aiming to reduce repetitive requests. Walmart's CEO, John Furner, expressed optimism about the meritocratic potential of AI tools, stating that no matter where the best idea comes from, it should be adopted and scaled up. He specified that this could be in Bangalore, the Greater Toronto Area, Mexico City, or Wichita, Kansas.

Walmart's internal tool, Code Puppy, is experiencing increasing success. The company's Chief Technology Officer, Suresh Kumar, explained at the annual shareholder meeting in Bentonville, Arkansas, that large categories of problems are being addressed repetitively. Compared to traditional search or applications, asking the AI to perform the same task is significantly more costly. Companies are looking to balance the rising costs of AI as the adoption of agentic tools in the workplace increases.

Kumar added that users do not need to ask Code Puppy the same question over and over again. A Walmart employee using the tool told Business Insider that they had not yet reached the token limit, which resets after a certain period, although they knew colleagues who had. Executives indicated that AI usage caps are a way to encourage employees to find a solution already developed by someone else. Furner stated that this helps highlight what has already been built, and that adoption rates increase much more rapidly.

Kumar mentioned that Code Puppy is now used by as many non-engineers as engineers, including hourly associates. If enough people ask Code Puppy to create the same thing, it indicates that the idea deserves serious attention. He explained that if five associates or 100 associates request nearly similar things, it presents an opportunity for Walmart to elevate that capability to an enterprise level. Walmart's engineers use a different vibe coding platform.

However, Walmart's low-price selling philosophy has a downside: the Everyday Low Cost (EDLC), which Kumar says underpins its vision for using AI. He stated that Walmart wants to ensure it maximizes the benefits of AI, but to do so in a way that is specific to Walmart, which means avoiding the repetition of the same tasks over and over again.