86 of 99 ordinance items have a cited finding. The rest are left open and named, each with who can answer it.
Prepared by Echo 1 Labs for 13|7 FrameWorks and Rawley McCoy & Associates. Inside: the eleven ordinance items for all nine sites, each with its citation; a takeoff beside what every sheet states; the follow-ups that close each open item; and a 3D scene of every site you can walk through. A working document, and corrections from the architect's team are welcome.
Grouped by the highway each plan fronts, because the corridor is what sets the TxDOT access rule and, on some sites, the zoning district or overlay itself. Select a site to open it in the matrix.
Arranged by corridor rather than by geography. Only six of the nine sites have a location verified to a parcel record, so plotting all nine on a map would imply a precision the drawings do not carry.
The jurisdiction gate is the one that matters. Zoning answers are worthless if the wrong authority is named, and on two of nine sites the authority was not the one the drawing implies. Nothing downstream runs until that is resolved from parcel and boundary data.
Two of the nine jurisdictions publish their codes in ways that defeat automated retrieval, so human review is a designed path through this process, not an error state.
The architect's checklist in the order it was written. Select any cell for the requirement, the citation, how the drawing compares, and where the finding was actually read. A cell marked not found means the governing text could not be retrieved, not that the item is unregulated.
Three axes, each computed from the screen rather than assigned. Certainty is the share of items that came back permitted, conditional, or outside any zoning authority. Evidence is the share that returned any finding at all. Clearance is the inverse of how many named approval steps still sit ahead of the site. Move the weights and the order changes with them.
Pick a site for its jurisdiction, zoning, approvals and findings. Every site has a 3D scene traced from the architect's own sheet, with a takeoff beside the numbers the sheet states. No sheet gives a building height, so height is marked assumed, and where a sheet states no scale or road edge the scene says what was derived. The scenes are for orientation, not a coordination model.
Ordered by how much each one changes the project if it goes the wrong way. The first two can each invalidate a site's program, and neither needs more research to resolve. Both need a phone call to a named office.
Finding what each jurisdiction requires means reading codes that were never written to be read, one site at a time. Echo 1 Labs built a workflow for the reading. A person reviews every item the code will not give up, and the decisions stay with your team.
Each site is tied to its adopting jurisdiction from parcel and boundary data. That jurisdiction's own published code is loaded, and the architect's checklist is the question set.
When a plan revises, the same screen can be run again. Every finding carries its citation. Every gap stays a gap, with the name of whoever closes it.
Where software cannot read a code, the item goes to a person. A check stops any uncited claim from appearing on this page.
The workflow that read these sheets for zoning can also read them for quantities. It runs in Titan, the Echo 1 Labs bid app, and a person stays on every step.
An AI takeoff reads the uploaded plan set into takeoff lines you can edit in place. On this engagement the same method read building footprints from the sheets' own vector layer and checked each one against the sheet's stated building area.
Each takeoff line is matched to the price book and flagged matched, fuzzy or unmatched. Unmatched items are researched and proposed as new entries with low, mid and high cost. A person approves each one before it enters the book, so the book grows with every job.
BOM templates per building type pull default quantities from the price book. The quote applies your waste, project management, overhead and fee markups, delivery, tax and deposit, with margin set by category.
The entitlement screen is one slice of a larger engine. Echo 1 Labs runs design-build work as an engineered pipeline: twenty specialist agents, each doing one discipline's job, joined by gates that stop the run instead of passing a bad answer forward. People decide at every gate.
Each step runs on the model sized to its job, retries once, and stops the whole run on a failure. Every step's full output is kept, so any number in a BOM or estimate traces back to the step that produced it.
Every open item already names who can answer it. Echo 1 Labs RevOps agents carry each question to that office by email and text, collect the replies, and bring every answer back with its source for a person to check. The screen updates, and the engine keeps what it learned for the next site in that jurisdiction.
The agents read codes, plans and replies in the jurisdiction's own words, draft each request around the exact section that is missing, route every item to the person who can answer it, and log what came back. They never decide a verdict on their own and never fill a gap by guessing.
This is a defined process, not a chat window. The work is broken into the same steps every time, with gates between them, run by agents and checked by people. Every answer, source and correction is recorded, so later cycles start with more jurisdictions known and fewer questions left to ask.
For each site, the first message drafted to the office that holds an open answer, with the other drafts for that site counted beneath it.
Drafts are written for the architect of record to send under his own signature.
This page is Level 1. Each level adds to the one before it, and every level keeps the same rule: a finding carries its source, or it stays a named gap.
Scope, turnaround and terms are set in the engagement agreement. Nothing here is a quote.
These are stated plainly because a screen that hides its gaps is worse than no screen. Most of what is missing is reachable. The single largest cause is that two jurisdictions publish their codes in ways that defeat automated retrieval, which is an access problem rather than a research problem.
Each finding names where it was read. Publisher means the jurisdiction's own published code. Unofficial mirror means a third-party copy to re-verify before any submittal. Sources that refused automated access are listed too, with what they affect, because a gap should be traceable as well.