Microsoft has introduced internal limits on employee AI spending, clarifying that increasing AI usage alone is not the company's objective. In an internal email reviewed by 404 Media, Jay Parikh, Executive Vice President at Microsoft, told employees the company is focused on getting better outcomes from AI rather than maximising token usage. "Tokenmaxxing is not what we are optimizing for. I want all of us focused on maximizing outcomes that move the needle for our customers and our business," Parikh wrote. The email also said Microsoft is making OpenAI's GPT-5.6 the default AI model for internal use, described as being less expensive than several alternatives. According to updated internal Copilot guidelines, every Microsoft division has had an AI token budget target since July 2026, and employees can view their individual AI token spending through an internal dashboard. The guidelines note that many engineers had been using AI tokens worth anywhere from a few hundred dollars to several thousand dollars each month. Microsoft said the policy changes are not linked to any weakness in its financial performance, pointing to its latest quarterly results showing year-on-year growth in revenue, operating income and net income, while exceeding market expectations. The changes aim to improve value generated from every AI token as Microsoft pursues its goal of becoming an AI-first organisation. The measures reflect a wider trend, with Amazon, Adobe, Atlassian and Citi also introducing steps to limit excessive AI usage, and Accenture discouraging employees from using AI for routine tasks.