Raw LLM Responses

Same, work in identity security, dumping our code into copilot is a big no no. However, these AI companies have proven a track record of asking for forgiveness instead of permission, so it’s fun pretending like all our code hasn’t already been injected because it’s all in github. We aren’t touting a lot of useless AI vaporware features, yet. We have one product that has a custom query language. And dumping that into a small LLM has actually proven useful to end users that can just tell it “give me all users who logged in on X” and it will spit out the query and run it.