Most field questions have an answer that already exists. It is sitting in a spec section or a detail somewhere in the set, and the only reason it turns into a phone call is that finding it takes longer than asking. A plan set and its specifications can run to hundreds of sheets and pages, organized for the people who drew them, not for the super standing at a wall who needs detail 5 at this exact condition. So the question goes out, and the schedule waits on a callback for something the project already settled.

This is not a small leak. Knowledge workers spend close to a fifth of the week just searching for and gathering information,[1] and construction has its own, heavier version of that number: field professionals lose roughly fourteen hours a week to non-optimal work, with about five and a half of those hours spent looking for project data.[2] A plan that takes ten minutes to locate, on a set you flip through fifty times a day, is the search tax in miniature. Brad's job is to collapse that ten minutes into one question.

The cost of search is not one dramatic event. It is a few minutes, fifty times a day, across everyone on the job, until it adds up to a day a week.

The needle is in the set. So is the wrong needle.

Finding information is only half the job. Finding the current information is the half that keeps you off the rework list. The superseded detail is always nearby: an older sheet still in the folder, a spec section a revision already rewrote, a dimension that an ASI changed three weeks ago and the field copy never caught up to. Keyword search makes this worse, not better. Search a PDF folder for a term and it returns every sheet that mentions it, current and dead alike, ranked by nothing that matters, and leaves you to figure out which one governs.

That is the gap that turns a quiet detail into a change order. A crew builds to the version they could lay hands on, the version on file was revised, and the difference surfaces at inspection or, worse, after the wall is closed. Bad data, the stale and the conflicting kind, drove an estimated $88.7 billion in rework in a single year, about fourteen cents of every rework dollar in global construction.[3] A search that cannot tell the current detail from the obsolete one is not saving you time. It is staging the next mistake.

How Brad searches: it reads the set, then connects it

Brad reads your drawings and specifications the way a construction administrator does, and then it does the part a folder of PDFs cannot: it connects them to the rest of the project. Revisions, change orders, RFIs, and clarifications all hang off the documents they touch. A spec section knows which detail it governs. A detail knows which revision supersedes it. That structure is what lets a question return an answer instead of a pile of candidates.

You ask in plain language and Brad answers from your documents. “What is the fire rating on the corridor walls at grid C?” “Show me the current curtain wall sill detail.” “Which spec section covers the gypsum board assembly here?” Brad surfaces the controlling detail or the governing CSI MasterFormat spec section, favors what is current because the revisions are connected, and flags the conflict where one document disagrees with a newer one rather than quietly picking a side. You get an answer you can act on, with the basis attached so you can check its work before you build to it.

Brad handles the finding and the cross-checking. The reviewer keeps the judgment. What gets faster is locating the current answer, not deciding what to do with it.
The useful result is not ten sheets that mention the term. It is the one detail that governs here, at its current revision, with the source shown so you can confirm it before the wall closes.
On what a good answer looks like

Who it helps, and how it compounds

Superintendents get the current answer at the wall without waiting on a callback from the office, which is where most of the day's lost float quietly goes. Architects and construction administrators stop re-answering the question the set already settled twice, because the field can self-serve the basics from the same connected record instead of routing every small thing through someone's afternoon. Builders and owners can check the picture themselves rather than opening a thread for a detail that was never really in dispute.

It compounds the way the search tax does, just in the other direction. Every question answered from the connected record is one less thread, one less re-measure, one less decision made on a stale assumption that becomes a change order three weeks out. The ability to influence cost is highest early and falls as the work proceeds, while the cost of fixing a mistake climbs the whole way,[4] so catching the wrong revision at the wall, before the crew builds to it, is worth far more than catching it at closeout.

Confirming the current detail before a crew builds to it is the cheap end of this curve. Discovering the wrong revision after the wall is closed is the expensive end.

Grounded in your project, with its work shown

Brad answers from your project's actual documents, not a generic model guessing at your job, and it shows what it drew from so you can verify it. That last part matters more than it sounds. An answer you cannot check is just a faster guess, and a guess at a fire rating or a structural detail is exactly the kind of thing that ends up in a dispute. The most common cause of construction disputes is a party failing to understand or comply with its contract obligations,[5] and the documents are where those obligations live. An answer that arrives with the governing spec section attached is one you can stand behind; one that arrives bare is one you have to go re-verify before you trust it, which puts you right back where you started.

So Brad shows its basis every time. The spec section, the detail, the revision it favored and why. When the set genuinely conflicts, it says so and points at both sources instead of papering over the disagreement with a confident answer. A discrepancy between the drawings and the specifications is a human's call, and Brad is built to surface it for that call, not make it.

Where Brad stops and the licensed reviewer starts

Brad is document intelligence pointed at your plans and specs. It reads them, connects them to the revisions and clarifications around them, and answers your question from your own set with the source attached. What it is not: a substitute for licensed professional judgment, or for your formal RFI and clarification processes. It makes those processes run from the current picture instead of a stale one. A person still owns the answer of record, and each workspace is walled off from every other one. If you have specific requirements about how your drawings and specifications are handled or retained, ask us and we will walk you through exactly how it works.

You will not stop the field from having questions. You can change whether the answer is ten minutes of flipping and a callback, or one question that returns the current detail with its source attached. The needle is already in the set. Brad is how you pull the right one out the first time.

Sources

  1. 1.McKinsey Global Institute. “The Social Economy: Unlocking Value and Productivity Through Social Technologies” (2012).
  2. 2.FMI & PlanGrid. “Construction Disconnected: Rethinking the Management of Project Data and Mobile Collaboration.” PlanGrid (Autodesk), 2018.
  3. 3.Autodesk & FMI. “Harnessing the Data Advantage in Construction.” 2021.
  4. 4.The American Institute of Architects. “Integrated Project Delivery: A Guide” (2007), p. 21. The MacLeamy Curve, after the Construction Users Roundtable white paper WP-1202 (2004).
  5. 5.Arcadis. “Global Construction Disputes Report 2021: The Road to Early Resolution” (11th Annual Edition, 2020 data). Arcadis, 2021.