AI Functions
2 min
chat ticket creation ticket triage ticket search email ticket triage email ticket triage (or manual ticket with triage task added when we get to integrations, allow them to flow through triage too) ticket enters triage task added as first task, assigned to triage ai agent analyize / summarize ticket ai set agreement determine if ticket belongs to existing larger topic problem (could be at sp level too ie m365 outage affecting all customers) project etc determine if there is already open tickets for this for the user or someone else in the org ensure all cis are added (devices, tickets) ai add ticket template ai review past tickets for similar ai determine if everything needed to work ticket is present ai check ticket template params are all filled and no further info is needed for ticket send email if clarification or additional information is required ai careful not to ask info we should know wait and receive info continue until clarification is adequate to complete the ticket, or user cannot provide clarification suggest resolution plan ai, optional if blank template, create a template resolution plan leverage documentation, tools/capabilitlies to make tasks set urgency and impact of issue if not defined by template set sla if not defined by template set board for ticket if not defined by template set category, type, subtype if not defined by template complete triage task copilot can be accessed from a number of screens and provides contextual ai assistance for those screens devices allows the user to ask questions about the device ranging from details about its config, health, security, logs etc to how to work with it, change it for scripts for it etc tickets ask questions about the ticket ask for recommended next steps for the ticket ask for a summary general tasks i'd like to request june 19 24th off i'll be out of the office tuesday afternoon for a dr appointment etc ticket notes ai can re write ticket notes for the tech based on their input and the context of the ticket, potentially adding detail and ensuring proper spelling, grammar and details customer updates ai can write/re write updates to customers (email, text, notifications), taking context from the ticket and input from the tech search boxes or function of general co pilot chat when using certain search boxes (such as devices) we give it all the tables/columns it can query by and it returns a custom query based on the users request (e g show me all devices with amd cpus, show me all devices missing patches ) to reduce prompt size, the recommendation is to have different ai agents/tools that perform queries around certain types of things (ie 1 for devices, 1 for tickets, 1 for companies & people etc ) generate scripts (tasks, in remote tools, as a tool for ai agents) device troubleshooting stage 1 review logs and other server side data stage 2 run diag script on device and analyze results stage 3 run diag script in monitor mode, and analyze results when ready remdiation ai / automated assess if tools exist to fix it determines a plan, runs tool or automation, waits for result, tests if problem resolved if not tries another idea logging each thing it tried along the way providing detailed ticket notes manual creates recommended next steps as internal ticket note adds a task for tech to perform recommended steps ticket summarization documentation creation review ticket to determine what solved the problem determine if a ticket template should be made, if so create one with appropriate tasks & automations create detailed documentation outlining the issue, resolution steps etc and the ticket template to use (this will be important for the msp to have documentation, and for the ai to reference to find the ticket template etc in the future) documentation repair techs can flag documentation as being inaccurate, incomplete or not working and provide some detail ai will attempt to rewrite/update the documentation based on their input