Skip to content
NewEraAI

AI news

Gemini API Managed Agents add new capabilities, what business teams should do next

New Gemini API Managed Agents capabilities are aimed at helping teams build more reliable production agents. Here is what changed and how to translate it into practical rollout steps for business workflows in the UK.

28 July 2026

Diverse team of professionals collaborating in a modern office with tech gadgets and headsets.
Photograph by Mikhail Nilov · Pexels

Managed Agents in the Gemini API are getting additional capabilities designed to support production ready builds. For business teams, the key shift is not just that agents can do more, but that the platform is positioning them for reliability and operational use inside real workflows.

What changed in Managed Agents

The announcement highlights new Managed Agents capabilities in Gemini API, including 3.6 Flash and support for hooks. The stated goal is to help developers build reliable, production ready agents.

Why this matters for operations and customer workflows

When agents are built for reliability, business teams can plan for more consistent handling of operational tasks and customer facing work, such as answering routine questions, coordinating steps, and triggering downstream actions. Hooks also signal a more structured way to connect agent steps to surrounding systems and processes.

What to do next for business adoption

  • Run a short review of your current agent or automation workflow to identify where reliability and consistent execution matter most for customers and internal teams.
  • If you are developing agents, map your workflow to the new Managed Agents capabilities announced for Gemini API, focusing on areas where hooks and production readiness reduce operational risk.
  • Create a pilot that measures reliability and repeatability in your target workflow before expanding to higher impact tasks.
  • Document clear escalation paths for when an agent cannot complete a step as expected, and align stakeholders on what success looks like for the pilot.
Practical focus for the next two weeks, choose one workflow that touches customers or core operations, pilot with reliability in mind, and define how you will decide to scale or stop.

Next step

Start with the free AI Opportunity Assessment.

A short, no-obligation conversation about where enquiries, hours and revenue leak today. You do not have to pick a tier to have it, and what comes out of it feeds Discover, so the first paid day starts from evidence rather than a blank sheet.