- Getting started
- UiPath Agents in Studio Web
- UiPath Agents in Agent Builder
- UiPath Coded agents
Agents user guide
Escalations and Agent Memory
As a developer, you may want to give your agent a way to validate business case details with a human before running a tool, or to request manual assistance if they run into issues. In other cases, there may be business exceptions in a process when you want an agent to have a human make an approval or some other type of decision.
This is where escalations come in. Escalations are powered by Action apps in Action Center and give developers the tools to design and configure human-in-the-loop events in agent execution. Check out the Creating agent escalations with Action apps quickstart guide.
Agent Memory is a service embedded in each agent that allows it to retain facts and observations across runs. It enables agents to make more informed decisions by leveraging past escalations and their resolutions, thereby supporting long-term memory alongside Context Grounding indexes.
How it works
When an agent encounters a scenario that triggers an escalation, it first checks whether it has previously memorized a resolution for that situation. If a matching memory exists, the agent can resolve the issue automatically without escalating. If not, the escalation proceeds as usual. Once resolved, the agent stores this outcome as a memory for future use.
There is a distinction between design-time and runtime memories:
- Design-time memory: Stores escalations and their resolutions during agent authoring and testing in Studio.
- Runtime memory: Captures escalations resolved during live agent execution in production.
Each resolved escalation in Action Center stores a key-value pair of the question and supporting context:
-
The key is the triggering question or query (limited to 80 characters).
-
The value contains the resolution context provided during escalation handling.
How to enable
Configure Agent Memory during escalation setup:
- Open the Escalations panel in the agent definition.
- Select an escalation app.
- Enable Agent Memory. This links memory storage to the latest version of the published agent.
-
Once enabled:
- At runtime, the agent stores each Action Center task triggered by an escalation.
- Users resolving an escalation have the option to exclude it from memory if it should not be reused.
You can create multiple escalations for different situations where an agent may require human assistance.
Use the Properties panel to add and configure the escalation.
- Select an action app form your
available resources.
You must first deploy an app to your tenant.
- Add a Prompt to help the
agent know when to use the escalation.
The prompt acts as a triggering condition, guiding the agent on when this escalation is appropriate.
- Optionally, enable Agent
Memory.
Activate this option to store question and response pairs from escalations, and use text embeddings to automatically resolve escalations with content that is very similar to past escalations. Agent Memory stores design time and runtime escalations separately.
- In the Assignments
section, add the escalation Recipient.
You can assign an escalation to a specific user in your UiPath® organization.
- In the Inputs section,
define the fields and values the agent will pass to the Action App. This
corresponds to the Action Schema, the required fields in the selected app.
For each field, add a description to assist the agent in understanding what kind of value should be passed. These descriptions improve the agent’s ability to infer and pre-fill values based on context when raising the escalation.
- In the Outcome behavior
section, specify how each outcome should be handled by the agent after an
escalation was resolved. For every outcome defined in the Action App:
- Select Continue if the agent should resume its process using the inputs provided by the escalation assignee.
- Select End if the agent should terminate as soon as the outcome is selected. This setting allows you to control the agent’s autonomy following human intervention.
After you create an escalation, you can reference it from the system prompt. Use explicit instructions, such as: "Before using Tool X, be sure to raise the Human Confirmation escalation and include the details for the action you're proposing".
Agents do not currently support the In/Out argument type for Action Schema properties.
