--- sourceDocument: Australia IT Service Management sourceDocumentLink: https://www.servicenow.com/docs/r/australia/it-service-management Release : - australia ft:locale : - en-US ft:publication_title : - Australia IT Service Management ft:clusterId : - itsm bundleId : - itsm workflow : - Technology --- # Autonomous vs. supervised # Autonomous and supervised execution modes {#ariaid-title1} Release version: Australia Updated October 15, 2026 ![](https://www.servicenow.com/docs/portal-asset/ico-clock) 2 minutes to read Summarize ![AI sparkle icon](https://servicenow.com/docs/portal-asset/ai-sparkle-icon) Summarized using AI This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete. ## Summary of Autonomous and supervised execution modes Autonomous and supervised execution modes define how independently a Level 1 IT Service Desk AI Specialist manages incidents in ServiceNow. These modes determine whether the AI specialist acts on incidents automatically or requires human approval before taking action, helping you tailor AI behavior to your incident handling workflows. Show full answer Show less ## Execution Modes * **Autonomous Mode (Default):** The AI specialist independently triages incidents, investigates solutions via knowledge bases and past incidents, interacts with requesters through follow-up questions, and posts solutions without needing human approval. If the AI cannot resolve the incident or lacks requester responses after three attempts, it escalates to a human fulfiller. * **Supervised Mode:** While the AI specialist performs triage and research similarly to Autonomous mode, it must obtain explicit confirmation from a human fulfiller before posting any solution. If the human rejects the proposed solution, the AI continues investigating alternatives. Escalation and communication rules mirror those in Autonomous mode. ## Communication with Requesters The AI specialist communicates exclusively through the incident's activity feed and does not directly send emails. Existing email notifications configured for incident updates remain active and operate independently from the AI execution mode. To adjust email behavior, you must modify email notification settings separately. ## Configuration and Management You assign the execution mode individually for each AI specialist instance within the Investigate and Resolve task configuration. There is no global setting affecting all AI specialists. Execution modes can be switched at any time to adapt to changing operational requirements. Autonomous and supervised execution modes control how independently a L1 IT Service Desk AI Specialist acts when handling incidents. Use this information to choose the right mode for your use cases. AI specialists can help handle incidents from beginning to end. They can classify and triage incoming user requests, then investigate related incidents and other knowledge sources for resolutions. If the AI specialist requires clarification, it can send updates to the requester through the activity feed. If a resolution is found, the AI specialist can propose the solution to the requester. If it can't identify a solution with sufficient confidence, it can reroute the incident to a human fulfiller to continue the work. The execution mode of the AI specialist determines its autonomy when performing tasks on incidents. The two types of execution modes are Autonomous and Supervised. Each instance of an AI specialist can have its execution mode set in the Investigate and resolve task configuration. ## Autonomous execution mode {#auto-vs-supervised-l1-itsd-ai-spec__section_psv_xtk_xkc} In Autonomous mode, the AI specialist performs tasks on incidents without pausing to ask for permission before acting. This is the default configuration. The general flow of autonomous execution is as follows: 1. An incident is assigned to the AI specialist. 2. It triages the incident after analyzing the description, attachments, and relevant context. 3. It searches knowledge bases, past incidents, known errors, and catalog items to identify a solution. It can also ask the requester follow-up questions to get additional information. 4. If a solution is found, it posts a solution to the requester and moves the incident to Solution Proposed, which proceeds to Resolved. No requester confirmation is required. 5. If it needs additional information from the requester, but the requester doesn't respond after the configured number of actions, the incident is rerouted to a human fulfiller. Default number of actions is 3. {#auto-vs-supervised-l1-itsd-ai-spec__ol_qsv_xtk_xkc} ## Supervised execution mode {#auto-vs-supervised-l1-itsd-ai-spec__section_rsv_xtk_xkc} In Supervised mode, the AI specialist reasons and researches on its own, but it does not take any action without explicit permission by a human fulfiller. The general flow of supervised execution is as follows. The first three steps of the flow are the same as in Autonomous mode. 1. An incident is assigned to the AI specialist. 2. It triages the incident after analyzing the description, attachments, and relevant context. 3. It searches knowledge bases, past incidents, known errors, and catalog items to identify a solution. It can also ask the requester follow-up questions to get additional information. 4. If a solution is found, it presents the proposed solution and asks for confirmation by a human fulfiller. 5. The human fulfiller reviews the solution. 6. If the human fulfiller accepts the solution, the AI specialist posts the solution, and the incident moves to Solution Proposed. 7. If the human fulfiller doesn't accept the solution, the AI specialist returns to the investigation phase to find an alternate solution. {#auto-vs-supervised-l1-itsd-ai-spec__ol_ssv_xtk_xkc} The same communication and rerouting rules from Autonomous mode apply to Supervised mode. ## Communication with users {#auto-vs-supervised-l1-itsd-ai-spec__section_tsv_xtk_xkc} The AI specialist uses the incident's activity feed to communicate with the requester. It doesn't send emails on its own. If you have emails configured to notify requesters that their incident has changed states or received comments, those emails are still sent. To change that behavior, you must update the email notification settings. They are unaffected by the execution mode of the AI specialist. ## Configuring execution mode {#auto-vs-supervised-l1-itsd-ai-spec__section_usv_xtk_xkc} When configuring or activating any specific AI specialist, you choose which execution mode it uses. There is no global setting that affects the execution mode of all AI specialists. You can also change the execution mode of an AI specialist at any time.