Remo AI · REMO Chat Agent
REMO Chat Agent
One agent across concept, procurement and programme.
Ask in plain language, inside whichever part of Remo you are working in. The agent answers from that project’s own data, shows the working, and asks before it changes anything.
- Report any variances identified in this submission.
- Why is the finish date governed by the authority approvals?
- Does the 1BR layout pass the rules?
One agent, three places
It sits inside each module, and knows only that project.
The agent reads the concept, the tender or the programme you have open. It is not a general chatbot, and it does not mix one project with another.
Check it, then change it
Checks drawings, pictures and unit layouts against the plot’s limits and the rulebook, and proposes changes for your OK.
ProcurementInterrogate the tender
Variances, price spread, where cost concentrates, and a full report on your own parameters.
ProgrammeQuestion the schedule
Findings, lags, activities and their logic, milestones and money, from the review and the XER file.
Does the 1BR layout pass the rules?
It uses the product’s own checks, and asks before it builds.
The agent does not draw or invent. It calls the same deterministic checks the app runs, and reports only what they return, citing rule ids, sheet numbers and finding ids.
Check consistency
Do the drawings, unit layouts and pictures still match the model and the rules, or has something drifted?
Look up the rule
Pulls the rulebook, your guidelines, reference plans and standards, and says which applies and why.
Propose a change
“Add a hotel tower.” It shows what changes against the plot’s limits first. Nothing is built until you confirm, and the earlier version is kept.
Say what it will cost
Pictures and photoreal passes are quoted before they run.
Questions taken from the product. See Concept.
In Procurement
Three things people actually ask about a tender.
Taken verbatim from the product. The data is illustrative; the questions are not.
- Asked Report any variances identified in this submission. Twelve trade sections priced in the PTE are null in the submission. The bid only looks competitive because the scope is incomplete.
- Asked Show the price spread across every section for the PTE against every vendor, highlighting the biggest gaps between the lowest and highest bidder. Lowest and highest per section, ranked by spread. Zero bids in Mechanical and Electrical are excluded scope, not savings.
- Asked Generate a full report comparing the PTE against every tender submission, with spread analysis, vendor ranking and an executive summary. Vendor ranking against the PTE and the lowest bid, charted and written out, then handed over as a document.
Variance, traced to the gapReport any variances identified in this submission.
Spread and concentrationShow the price spread across every section for the PTE against every vendor, highlighting the biggest gaps between the lowest and highest bidder.
A report on your parametersGenerate a full report comparing the PTE against every tender submission, with spread analysis, vendor ranking and an executive summary.

Why is the finish date governed by the authority approvals?
Answers from the file, with the activity and date behind every number.
It answers from the review and the Primavera XER: findings, lags, activities and their logic, milestones and money. It quotes only what its tools returned, says “not in the file” when something is not there, and refers to roles, not names.
Every AED figure carries its derivation. See Programme.
Fictional project and contractor. No client data.
How it answers
The same way, wherever you ask.
The working, not just a summary
Every answer carries the figures, rule ids or activities underneath it, so you can check the conclusion instead of trusting it.
It asks before it digs or changes
Where something looks wrong it offers the next check. Where a change is possible, it shows the effect and waits for your OK.
Reports come out as documents
A request for a full comparison returns a report you can open, not a wall of chat to copy out by hand.
Scoped to your project
It reads the project you point it at. Nothing leaves it, and nothing from it trains a model.
In motion
One agent, three modules.
1 minute 12 seconds. Concept, procurement and programme, each shown on a synthetic or illustrative project.
Bring a project and ask it something hard.
The questions that take an afternoon in a spreadsheet are the ones worth testing it on.