AI agents — the software systems that can take actions on a user's behalf rather than just answer questions — may soon be able to find and use their own tools while they work. According to ZDNet, a new open standard backed by Microsoft and Google would let AI agents search for and call the tools they need at runtime, rather than being limited to a fixed, pre-wired set chosen by their developers.
Think of it as a search engine built not for people, but for the AI agents themselves. Today, an agent can typically only use the specific tools and integrations a developer connected in advance. The approach ZDNet describes points toward agents discovering capabilities on the fly — locating the right tool for a task in the moment and putting it to use, much as a person searches the web to find a service they need.
The detail ZDNet emphasizes is that this is an open standard, and that two of the largest players in cloud and AI, Microsoft and Google, are behind it. Open standards matter because they let different companies' systems work together instead of locking users into a single vendor's ecosystem. Backing from rivals of that size signals an attempt to set a common, industry-wide way for agents to find and connect to tools.
ZDNet frames the development as something that "may soon" arrive, so it describes a direction of travel rather than a finished, widely available product.
Why it matters: if AI agents can locate and use tools at runtime through a shared standard, they could become far more flexible and capable without developers hand-wiring every connection — a shift that could shape how the next generation of AI assistants gets built and used.