Amazon's cloud division is launching a new unit dedicated to developing software for managing artificial intelligence agents as the company strives to keep pace with its competitors in generative AI. Swami Sivasubramanian, an Amazon vice president with nearly 20 years at the company, shared these plans in a LinkedIn post on Wednesday. "Agentic systems present opportunities that surpass current chatbots and will enhance efficiency like never before, " Sivasubramanian stated, noting his recent role overseeing Amazon Web Services' database, analytics, and AI services. "They will manage intricate workflows and tackle problems with human-like reasoning, while maximizing efficiency and cost-effectiveness at scale. " On Tuesday, Reuters reported that AWS CEO Matt Garman described agentic AI as a potential multibillion-dollar opportunity for AWS in a memo regarding the new unit. Customers have begun utilizing agent software via AWS, while Amazon also employs it extensively in-house. For instance, the company's developers have started using AWS' Q developer service to write or update code. Sivasubramanian mentioned that Amazon has "saved 4, 500 developer years through Amazon Q Developer's capability in code transformation for upgrading Java applications. " In November, Microsoft introduced an Azure AI Agent Service, followed by Microsoft-backed OpenAI allowing paying users to trial its agent software called Operator in January.

In December, Google announced it would offer limited access to its Agentspace tool for select customers. Amazon leads the cloud infrastructure sector, generating nearly $29 billion from AWS in the fourth quarter, outpacing competitors Microsoft and Google. Additionally, Amazon has begun promoting an advanced version of its Alexa voice assistant, designed to leverage AI models from AWS as well as Anthropic, a company backed by Amazon. Although reports indicated that Anthropic is managing the most sophisticated features of the new Alexa+, an Amazon representative labeled this information as "false. " Several existing teams will transfer to the new organization, which will include vice presidents Asa Kalavade, Dilip Kumar, and Deepak Singh, as detailed by Sivasubramanian in a memo to his colleagues on Tuesday. "Our aim will be to create AI agents that are not only powerful and efficient but also trustworthy and responsible, " he wrote.