Pre-Publish Insights
On the Review step before you publish, an AI read of your whole request and every BoQ line: scope gaps, missing requirements, budget, timeline, supplier coverage, and per-line findings routed to the person who can fix them. Advisory only, never blocking.
What you'll learn
- Catch gaps in a request, and in every BoQ line, before suppliers ever see it
- Read the request-level insights and the BoQ content findings, and route each finding
- Understand the Before you publish dialog and why it never blocks you
- Know what the analysis costs, when it runs in the background, and when it goes stale
You can publish a request that is technically complete and still get weak results. The required fields are filled, so nothing stops you, but the scope is vague, a key requirement is missing, the timeline is tight for the category, the budget does not match the work, or line 412 of the Bill of Quantities says "as per drawing" with no drawing attached. You only find out when the bids come back thin, off-target, or full of qualifications, and by then the request is already out in the market.
Pre-Publish Insights is the read you get before that happens. On the Review step, the last step before you publish, it reads the request the way an experienced colleague would on a final pass, and it reads every line of the Bill of Quantities the way an estimator would. It is advisory only. It never blocks publishing and never changes your request. It points out what it sees, tells you who should fix each line, links you to the step, and leaves the decision to you.
Why this matters
The required-field checks in the wizard tell you whether a request is complete. They cannot tell you whether it is good, and they cannot tell you that a BoQ line is unpriceable. On a construction bill of two thousand lines, the ambiguous ten are the ones that come back as exclusions, assumptions, and Q&A traffic. Reading them before publishing is cheaper than reading them in ten bids.
When to use it
Run the insights on the Review step, after you have built the request and before you publish. It is most valuable when:
- The request is detailed enough that a gap is easy to miss on your own read.
- The Bill of Quantities was imported from a client's workbook, so the wording is theirs, not yours, and you have not read every line.
- You are sourcing in a category or market you do not run often and want a sanity check.
- You want a second read on whether your timeline and budget are realistic for what you are asking.
Because it is advisory, there is no penalty for skipping it, and no requirement to act on what it says.
What it looks at
The analysis has two layers.
The request as a whole. Scope clarity, requirement gaps, budget feasibility, timeline tightness, and supplier coverage: the panel's own description says it "covers supplier coverage, scope clarity, BoQ quality, budget, and timeline." Each finding explains what it noticed and links to the step where you would fix it.
Every BoQ line. Each line is read for the things that make a line hard to price honestly: a description that depends on a document that is not attached, a quantity with no unit or a unit that does not fit the description, a line that contradicts another, an allowance typed as a measured item, scope that belongs to a specialist. There is no ceiling on how many lines it reads; the result says how many it covered (N BoQ lines reviewed). Findings are grouped under N BoQ content findings, and each one is routed:
- Clarify with client: the ambiguity comes from the source bill and only the client or consultant can resolve it.
- Fix internally: a correction your own team can make in the grid now.
- Ask a specialist: the line needs a trade or engineering judgement before it can be tendered.
Open BoQ on a finding jumps to the line. When the request looks solid, the panel says No major issues found rather than overclaiming that everything is perfect.
What it uses
Pre-Publish Insights uses AI credits. The Analyze button shows the cost before you run it, and a run that fails or that you cancel is never charged. The BoQ line review is included in that price: a large bill costs the same stated amount as a small one. It is available on every plan, gated by your credit balance rather than your tier. If your balance is too low, the panel shows Not enough AI credits to analyze with a Top up path for admins.
Wondering how credits work across the AI tools? AI credits covers your plan grant, top-ups, and your balance. This tool draws from the same balance as every other AI feature.
Running the insights, step by step
Reach the Review step and run the analysis
Work through the wizard to the Review step. The insights panel is marked advisory only. Select Analyze; the button shows the credit cost, so you know what the read costs before you start it.
Keep working while it runs
The analysis runs in the background: Reviewing your request and every BoQ line… with the note that you can keep editing while it works. On a large bill it takes a little while. Cancel stops it, and a cancelled run is not charged. When it finishes, the panel collapses to a one-line summary, N insights to review or No major issues found, that you expand when you are ready.
Read the findings and route each one
Request-level findings come first, each with a jump to the step that owns it. Then the N BoQ content findings block, each finding tagged Clarify with client, Fix internally, or Ask a specialist, with Open BoQ to go to the line. Long lists fold behind +N more lines. The footer records what the run cost and how many lines it covered.
Act, re-analyze, or publish as is
You decide what to do with each finding. Follow a link, make a change, and come back. Or judge that a flag does not apply to your situation and move on. If you edit the request after running the analysis, the panel shows Your request changed since the last analysis and offers Re-analyze, so an old read never sits over a changed request. Re-running with no meaningful change costs nothing; a real edit re-runs the read.
The Before you publish dialog
If open findings remain when you press Publish Request (or Submit for Approval), a Before you publish dialog steps in once. It says the advisor flagged things worth a second look, reminds you this is advisory only, and offers Review findings or Publish anyway. If the analysis is stale it says so here too. Nothing here blocks you; it makes sure you saw the list before the request goes live.
What good looks like
- Run the analysis once the request is genuinely finished, on the Review step, so the read reflects the real thing rather than a half-built draft.
- On an imported construction bill, treat the BoQ content findings as your first read of the client's wording. Route every Clarify with client finding to the consultant before you publish, not after the bids come back qualified.
- Read every flag, but apply judgement. A tight-timeline note may be exactly right for an urgent request you intend to push out fast.
- Re-analyze after a meaningful edit, so the fix is confirmed closed.
- Use the dialog as a checkpoint, not an obstacle: Publish anyway is a legitimate choice once you have read the list.
Catching the unpriceable lines before the bidders do
A main contractor imports a client's 1,800-line workbook, attaches the spec book, splits it into packages, and runs the insights on the first package. Beyond a thin-pool warning, the BoQ review finds nine lines: five reference a drawing that is not attached, two carry a unit that does not fit the description, and two are allowances typed as measured items. The five go to the client's consultant, the two units are fixed in the grid, the two allowances become provisional sums. The package publishes with zero "as per drawing" exclusions in the bids that follow.
What suppliers see
Suppliers never see the insights, the findings, or the fact that the request was reviewed by AI. They see only the request you publish. Everything the panel surfaces is private to you. What reaches the market is your reviewed, finished request, shaped by your decisions about which findings to act on.
What comes next
A reviewed request is ready to publish. If a flag pointed at a thin supplier pool, supplier matching can help you widen it before you go live; once the request is published, your next moves are managing who can bid, running the Q&A, and preparing to evaluate.
Next steps
- Publish your requestFinish the Review step and take the request live.
- Import your BoQ from a spreadsheetWhere the lines the review reads come from.
- Widen a thin supplier poolLet AI Supplier Matching rank suppliers that fit your request.
- See the full AI toolkitEvery AI tool on VEXORS and where each one lives.
Related guides