Restaurant Dive published a Q&A with Oliver Ostertag, president of growth and AI at Par Technology, arguing that recent restaurant AI failures reflect immature technology rather than a poor fit for the sector. The interview cites Starbucks abandoning its computer-vision inventory counting system in May 2026 after less than nine months, McDonald's easing off voice AI testing in 2024, and a Pizza Hut franchisee's claim that an AI order management system cost $100 million in sales. Ostertag said results depend on "context equity" — platform-level data spanning POS, payments, back office, labor, inventory and loyalty — and described agent use cases such as analytics querying, promotions optimization and fraud detection. He said Par's Anthropic Claude token spend has risen month over month while coding efficiency metrics improved, and that Par builds primarily on AWS Bedrock plus OpenAI models, with Burger King using its POS and back-office products.