Submittals Portfolio
a program-wide view of every open submittal: progress metrics, review deadlines met or missed, and a breakdown by discipline, so nothing ages out of its review window unseen.
MEP / building-systems engineering
Catch the discrepancy before it’s built
Submittal and shop-drawing review for a mechanical and electrical engineering team, with AI running line-by-line cross-checks, ranking findings by severity, and citing the source for each one.
Anonymized client build . Mereline Review Hub
The challenge
When equipment arrives on a submittal, say an air handling unit, an engineer has to verify that the submitted product actually matches the specified one: airflow, static pressure, cooling and heating capacity, motor and electrical characteristics, dimensions, weight, accessories, control sequences. Dozens of parameters, against the spec section and the contract drawings, on every submittal, across every discipline.
Done by hand it’s hours of careful comparison per submittal, on a clock, with real liability if something slips through. Reviewer capacity, not engineering judgment, becomes the bottleneck.
How we solved it
We built a review platform that turns line-by-line comparison into a structured workflow, from a program-wide view down to a single parameter. The system runs across four connected screens.
Submittal Review Workspace
a split screen with the shop drawing on one side and the AI’s conformance summary and ranked findings on the other, so the engineer reviews without flipping between documents.
RFI Tracker
a Kanban board for requests for information, with a detail panel, so questions raised during review are tracked to resolution instead of lost in email.
Substitution Review
when a contractor proposes an alternate product, a side-by-side comparison of specified vs. proposed attributes with an AI compliance verdict.
How AI is integrated
AI does the comparison work and hands the engineer a ranked worklist:
- Automated cross-check the AI reads the submitted values against the specified values and the contract drawings, parameter by parameter, and identifies where they diverge
- Severity-ranked findings each discrepancy is classified critical, major, or minor, so a critical mismatch isn’t buried next to a cosmetic note
- Confidence scoring every finding carries an AI confidence level, so the engineer knows where to trust the machine and where to look closer
- Substitution analysis proposed alternates are scored against the specified product, with the deviations flagged and a recommendation attached
- Source-cited findings point back to the exact spec section or drawing, so the engineer verifies in one click instead of digging
The engineer stops hunting for discrepancies and starts adjudicating them. The system has already found, ranked, and cited them.
Why it matters
Review stops being gated by how many submittals a person can read in a day. The AI does the line-by-line cross-check and hands the engineer a ranked, source-cited worklist, so deviations get caught before they’re built, and reviewers spend their judgment where it counts.
Outcome metrics available on request, once the client’s numbers are cleared for publication.
A system that’s yours.
Build yours next.
Book a discovery call