You are here:
AI Actions for Problem Management
Accelerate problem resolution and improve accuracy by using problem actions. Problem fulfillers and managers use problem actions to streamline diagnostics and automate collaboration workflows to resolve issues efficiently throughout the problem lifecycle.
Required Editions
| Available in: Lightning Experience |
| Available in: Enterprise, and Unlimited Editions with Agentforce IT Service. |
Find and Update Problem Records
Problem fulfillers or managers can use the agent actions to quickly access and update record details, which streamlines their investigation and prepares them for the next step.
Here’s how a problem fulfiller quickly find and update the problem record for an ongoing investigation into slow system performance.
| Instructions | Example Utterance or User Input | Agent response | Standard Action Engaged |
|---|---|---|---|
| Update the Problem record's status and assign it to the responsible team. |
|
The agent updates the problem record with the specified changes and confirms that the status is now in progress and the owner is the infrastructure team. | Update Record Fields |
Investigate and Resolve a Problem
Problem fulfillers can use Agentforce actions to summarize problems, identify their scope, document root causes, propose permanent fixes, and record resolutions.
Here’s how a problem fulfiller can investigate a slow system performance problem from end-to-end using Agentforce.
| Instructions | Example Utterance or User Input | Agent response | Standard Action Engaged |
|---|---|---|---|
| Generate an overview of the slow system performance problem. |
|
The agent summarizes the problem record (for example, investigation: slow system performance, status: root cause analysis, impact: multiple users reporting application slowness and timeouts). | Summarize Problem |
| Verify attributes of the problem. |
|
The agent displays the key attributes of the problem (for example, category: hardware subcategory: laptop impact: high, urgency: high). | Check Problem Attributes |
| Find new and unlinked incidents related to the slow performance issue. |
|
The agent performs a semantic search and identifies 3 new incidents logged in the last four hours that contain the keywords slow and unresponsive. It then suggests linking these incidents to related records related to the problem. | Associate Related Records for Problems |
After generating summary and linking to related records, identify the root cause of the problem.  Note The problem fulfiller must copy the resolution summary from
Agentforce to the resolution summary field before marking the problem as
resolved or closed. This step ensures that the problem record is updated with a
clear summary of the resolution. |
|
The agent summarizes root cause of the problem. A misconfigured database query cache causes the primary application servers to use high levels of CPU and memory, which results in overloads during peak business hours. | Create Problem Root Cause Summary |
| Propose a permanent fix for overloaded servers based on best practices. |
|
The agent finds a solution in knowledge article KA-0003. Based on knowledge article KA-0003, the proposed fix is to change the database query cache and use a load balancer to distribute traffic across a new server. | Propose Resolution Summary For Problem |
| After the fix is implemented, document the resolution summary. |
|
The agent analyzes the change request notes and generates a draft summary. The problem was fixed by changing the database cache and installing a new load balancer. | Create Problem Resolution Summary |
Collaborate and Communicate on a Problem
Problem fulfillers can use Slack swarming to troubleshoot critical issues, and Agentforce summarizes the conversations, posting key findings directly to the problem record.
Here's how a problem fulfiller documents the swarming session focused on resolving the overloaded server issue and records its impact.
| Instructions | Example Utterance or User Input | Agent response | Standard Action Engaged |
|---|---|---|---|
| Generate a summary of the live troubleshooting discussion from the dedicated swarm channel. |
|
The agent summarizes the linked Slack conversation (for example, problem: 95% server CPU usage, cause: misconfigured query cache, contributors: Mark (DBA), Smith (infrastructure), actions: Mark to optimize cache, Smith to prep change request for load balancer, decision: optimize cache now, plan load balancer for long-term). | Summarize a Slack Channel |
| Post the generated summary to the problem record feed to document the findings and the action plan. |
|
The agent posts the generated swarming summary and problem resolution as a new Feed Item on the current record for visibility and historical tracking. | Post To Feed |
