Fixing AI flags in my dissertation analytical essay?

I am a second-year PhD candidate in comparative literature and my supervisor just returned my chapter draft with a red flag for 'AI-generated patterns.' I am an ESL writer, so I often rely on structured templates to keep my syntax clean, but now I am terrified that my formal academic tone is being misread by the detection software. I have three weeks until my mid-term review and need to know how to manually revise the sentence variety to make the text sound more authentically human without losing my analytical precision.