I have already developed most of a workflow in n8n integrated with Microsoft Teams and GLPI, but I need expert support to complete it. What is already done ✅ App Registration (grant) in Microsoft is already created and functional (with proper permissions and working authentication in n8n). Initial integration with Microsoft Teams Trigger is working. Flow already sends a welcome message and presents an interactive menu (open ticket, reply ticket, exit). Partial implementation of variable storage (e.g., ticket title and description). Basic integration with GLPI REST API already set up to open tickets. The main challenge Correctly identifying the user’s reply in a 1:1 chat (not confusing bot messages with user messages). Avoiding bot loops, ensuring only user responses are processed and stored. Capturing and attaching files sent by the user in Teams to the ticket in GLPI. In the attached screenshot, you can see that the bot keeps sending the same question repeatedly every time the user replies. The correct behavior should be: Bot asks Question 1 → user answers → workflow saves the response. Bot proceeds to Question 2 → user answers → workflow saves the response. And so on, until the dialogue is completed. Currently, the bot fails to recognize that it’s part of an ongoing conversation after the first message, and therefore it does not properly continue the dialogue as expected. What I need Adjust the workflow to reliably capture user responses and attachments. Ensure that each interaction (title, description, attachments) is stored in variables and used to create/update GLPI tickets. Implement clear logic to differentiate between bot and user messages, avoiding loops. Test and deliver simple documentation on how to maintain and evolve the flow. Required Skills Hands-on experience with n8n (conditions, variables, state handling). Knowledge of Microsoft Teams API (1:1 chat, attachments, triggers/actions). Experience with GLPI REST API (ticket creation, file upload). Best practices in automation to avoid duplication/loop issues in bots.