01
Describe the job, not every step
Tell the chat what result you need — for example, summarise new enquiries and save a note. Your description becomes an agent draft; refine its rules and behaviour through conversation.
Managed Agents
Turn your task into an AI agent.
Summarise enquiries, update contacts, prepare content: describe the job you want your agent to do. Connect knowledge and suitable actions, test it, then publish it for use.
Here’s what you can do.
Three examples show how you can use Managed Agents.
02
Your knowledge. The right actions.
Connect your knowledge base and choose the tools the task needs. The agent can then read contact information and add an internal note, for example. Other apps require their own connection and appropriate access permissions.
03
See what your agent actually does
Before launch, test incomplete and difficult enquiries too. The activity feed then shows the trigger, status, tools used and results of individual runs — including failed ones.
What this changes for your day-to-day work.
The practical benefits of Managed Agents.
Classify enquiries more easily
Let the agent classify an enquiry. Your configured workflow can use its response in the next step.
Hand off recurring tasks
Start a published agent at a set time or at regular intervals. Choose the schedule duration and number of repeats.
Work with your connected apps
Supported MCP connections give the agent selected tools from other systems. Review the account, permissions and available actions first.
Prepare content drafts
With media capabilities enabled, an agent can generate images, audio or video. Review the results before any separately configured publishing action.
Connect your own software
A published agent can receive tasks through a configured API integration. This requires the matching trigger and an authorised connection.
Simplified examples, not live account data.
What you need to get started.
Confirm access and a usage budget. Choose knowledge, tools and triggers for the task, then test with sample contacts. Tests using connected tools can also make real changes. Publish only after review.
A form, contact event or workflow can start your agent when the matching trigger is configured. Workflow invocation requires a published agent configured for that purpose. Keep fixed sequences in the workflow and let AI handle interpretation.
Availability and any add-on or usage costs are confirmed for your ONE LOOP account before implementation.
Build on it.
All AI & Automation features
Your questions, answered.
What does “Managed” mean here?
It means an AI agent you create in chat and operate through its configuration, not an included human support team. Setup and ongoing management by ONE LOOP are agreed separately.
Is this the same as a chatbot or workflow?
No. A chatbot focuses on conversations; a workflow follows defined steps. A Managed Agent handles tasks using knowledge and tools. It can run inside a workflow or from other configured events, without automatically replacing existing bots or processes.
How do I reduce the risk of unwanted actions?
Enable only the tools and account access needed, define clear boundaries and test exceptions. A chat instruction is not a technical access restriction or a guarantee of error-free results. Sensitive changes need a suitable human review step.
Are Managed Agents included in my package?
We check this for your ONE LOOP account. Access, included runs and additional usage charges depend on your offer. Connected apps, media generation or messaging can add costs. Unlimited usage is not promised.
See it with your own use case.
Let’s explore your use case and what the setup requires.
Book a demo