Normalise a T-12 and find the loss-to-lease in it
Lesson 1 of 6 · Operate and hold
1 · Learn the move · Cognitive verifier
Your manager's trailing-twelve is about to go to your partners and your refinance lender, and once it leaves your desk it is your statement. Two things hide in it. On the income side, physical occupancy looks healthy while loss-to-lease, concessions and bad debt quietly take the difference, and only the rent roll shows how much. On the expense side, lines a buyer or lender expects are simply missing, and one-time credits flatter the year. The cognitive-verifier move is to make the model build the income stack from the rent roll line by line, then check that the stack ties to the net rental income the statement reports, and treat a failure to tie as a finding rather than something to smooth. The spine changes one habit from how this is usually done: a missing management fee or reserve is not filled with a typical percentage. It reads [NEEDS: the management agreement] or [YOUR NUMBER], because a range the model supplied looks exactly like a figure you verified.
Read this T-12 and rent roll for my own building. Derive loss-to-lease from the rent roll only, unit type by unit type, math shown. Build the income stack with source labels, then check it ties to the statement's net rental income and report any gap as a finding. Physical and economic occupancy, both shown. Missing expense lines read [NEEDS: document] or [YOUR NUMBER], never a typical range. [T-12] [rent roll] [GL detail]
2 · Your turn. You write the prompt
You built and hold The Commons at Harrow Crossing, the 168-unit apartment phase. Your third-party manager sent the trailing-twelve and a rent roll ahead of the refinance. The rent roll carries the manager's own market rents. The statement has no management fee line, and the insurance detail shows a credit. Your partners and the lender both see this package next week. Write the prompt that normalises it before it leaves your desk.
Write the prompt for this scenario as given, or swap in a real one from your own work. The score works the same either way.
Remember: the AI sees only your prompt, not this page. If the situation isn't in your prompt, it doesn't exist.
Optional. These shape the output when you run your prompt below, not your score.