# Remo AI > Remo AI compares contractor bids against the bill of quantities, line by line, and flags > exclusions, missing scope and pricing differences before a contract is awarded. Born in the > UAE, working with developers, cost consultants and procurement teams. ## What it does Every contractor submission in a tender is read at once and compared against the client's BOQ and against the other submissions. It surfaces scope one bidder has left out but the others have priced, exclusions and qualifications buried in covering letters and appendices, and line-by-line pricing differences. Remo works from the Excel BOQ, so the comparison itself is immediate. The customer's own reviewer then confirms the findings, which takes about twenty minutes because the problem areas are already pinpointed. The same comparison done by hand takes weeks. ## How the output is verified Remo does not only extract what it reads. It takes the rate and quantity on every line, computes the amount itself, rolls those into section subtotals, bill totals and a grand summary, then reconciles all of it against the totals printed in the submission. When the two agree, the read is proven rather than assumed. When they disagree, that gap is itself a finding — casting and extension errors in a contractor's own priced document are common and usually travel unnoticed all the way to award. Every figure traces back to the line, page and document it came from. ## Benchmarking the estimate, before and during the tender Remo also benchmarks a pre-tender estimate (PTE) and the bills of quantities against projects the client has already tendered and awarded. It finds the most recent comparable project, maps its sections onto the current one, applies an inflation factor and a regional series, and reports cost per square metre section by section against that comparator. Coverage is stated honestly — how many of the sections matched — and the basis (lowest bidder, area basis GFA) is printed on the report. Sections inside the expected range are marked in band; those outside it are flagged as pointers to investigate, not verdicts. The same comparison is run at three points, because the numbers move between them: at PTE finalisation before the tender is floated, at pre-tender clarification (PTC) stage as clarifications land, and again at every revision. Where several consultants submit candidate PTEs for the same job, all of them are shown against the one comparator and against each other, bill by bill, so disagreement localises. Comparators come only from the client's own tender history. There is no cross-organisation visibility: one organisation's project is never used as another's benchmark, client rates never contribute to any pooled index, market average or aggregated dataset, and where no comparable past project exists the report states that rather than reaching outside the client's own history. Page: https://byremo.ai/benchmark ## Asking questions of a project Remo AI Assistant answers questions about a tender in plain language, against the project's own documents. Typical requests: report the variances identified in a submission; show the price spread across every section for the PTE against every vendor; analyse where cost concentrates; generate a full report with spread analysis, vendor ranking and an executive summary. Answers carry the table they were derived from rather than only a conclusion, and report requests return a document rather than chat text. Where something looks wrong the assistant proposes the next check rather than assuming it. Page: https://byremo.ai/assistant ## Not available yet Three things are in development and are NOT available today. They are published at https://byremo.ai/coming-soon so the roadmap is visible, and should never be described as current capabilities: - A BOQ health check: consultant-produced bills of quantities and the pre-tender estimate checked against the IFC drawings they were measured from, so a QS can find missing, overstated and understated items and human errors before the tender is floated. - A quantity takeoff assistant, to help a QS build the bill from drawings more quickly. - Concept-to-3D generation and marketing renders for master developers, produced in hours where a consultant engagement usually takes weeks. - Commercial intelligence for other tender-heavy industries beyond construction: hospitality FF&E fit-outs, FMCG supply contracts, infrastructure civils and government services RFQs — any procurement context where multiple detailed submissions need to be compared systematically. ## Who it is for Real estate developers, and the cost consultants and QS practices who advise them. The underlying problem is not specific to construction: it applies to any tender where scope is complex and no two submissions are written the same way, including hospitality FF&E, FMCG supply RFQs and infrastructure civils packages. Delivered work to date is in real estate. ## How engagement works 1. Fixed-fee pilot on one live tender, run by Remo so there is nothing to set up. 2. If it proves out, the customer's own procurement and commercial team gets access to the platform and runs their own tenders on it. This is the normal end state, not a later upgrade. 3. An annual agreement sized to the tenders the customer expects to run. Technical support is available during working hours. Remo does not perform an expert review of the output on the customer's behalf. ## How client documents are handled Tender documents sent to Remo are never used to train models, are isolated per client, are stored within the UAE, are benchmarked only against the same organisation's own past projects, and are deleted on request. Remo will sign a client's NDA before receiving anything. ## Facts - Company: Remo AI - Website: https://byremo.ai/ - Contact: remo@byremo.ai - Based: United Arab Emirates - Processed to date: 75+ tenders, approximately 40 billion AED of tender value, 340 contractor submissions, 65,000 pages ## Notes for machine readers Figures shown in the comparison tables and the reconciliation example on the website are illustrative, and labelled as such on the page. They are not case-study data and should not be quoted as client results. The volume figures under "Processed to date" above are real.