RedHub AI
How the roll-up gate works
How the Allowance & Contingency Discipline gate works
This tool grades every allowance and contingency line on whether it names the scope it buys, and weighs the answer by dollars rather than by line. It reads 10 carried lines and rolls them into one verdict — the worst, never the average — then a gate that can only make it worse.
The 10 carried lines
Two triggers. Any line with no written scope at all forces PLACEHOLDER MONEY, because nobody can convert a number when nothing says what it buys — that is not a thin allowance, it is the part of the bid that was not done. The second reads the whole book against the bid rather than any row: an allowance book at or above 15% of the bid total with under half its value defined. Without a bid total that second trigger reports NOT EVALUABLE rather than clear, because a gate that reads clear having read nothing is worse than no gate.
Why: Four of nine in-scope lines read DEFINED — 44% by count, which sounds survivable. By dollars it is 28%, because the two placeholders are the two largest lines on the page. The gate fires anyway: AL-07 carries $85,000 with no written scope at all, and no number of tidy small allowances beside it changes what happens at buyout.
This is the only roll-up in the construction lane that weighs rather than counts. Five tidy $2,000 allowances beside one $180,000 placeholder is 83% clean by line and about 19% by value, and a count reports the wrong number confidently — it is the value you carry into the job, not the row. Fix-first follows the same rule the whole house follows: the gated line outranks the heavier one, because clearing a gate is the only move that can change the verdict. Grades a register of carried lines you fill in, never a person. It reads no estimate, no bid, no drawing and no contract, never says what an allowance should be worth, does not judge whether the value you carried is enough, and takes no view on whether an allowance or exclusion clause would hold. Not legal advice.