AI Assistant
Your SoW, now you can talk to it
A construction-aware assistant lives inside every Xtimo project. Ask what a quantity means, where a rate came from, or which lines need a second look. Then tell it to draft a new section or rewrite a dozen items at once. It reads your full SoW, your blueprints and your spec, so the answers are about your job.
What it is
A surveyor's second set of hands
Powered by AI and wired into your project data. It sees the priced SoW, the take-off behind each quantity, the confidence score on every line, and the source drawings. That context is what separates it from a chat window bolted onto a spreadsheet.
Reads the whole project
Priced lines, quantities, confidence scores, flagged items, and the underlying blueprints.
Two modes, one window
Analysis answers and explains; Generation drafts and edits. Move between them in the same conversation.
Construction-literate
It understands SoW structure, NRM sections, take-off logic, and a provisional sum versus a measured item.
Always traceable
Every answer points back to the line, quantity or drawing it came from. Nothing without a place to check it.
Analysis mode
Reads, never rewrites
Use Analysis mode to understand the estimate without changing a thing. Ask why a quantity looks high, which items were flagged, or how the total breaks down. It pulls the answer from your live data and shows its working. It is read-only by design.
Trace a take-off back to the drawing, room and measurement logic.
Pull every flagged line, sorted by confidence, so your eyes go to the doubtful ones first.
Totals by trade, section or drawing reference, assembled from the priced SoW.
See the price library match and the comparable rates HAWK drew on.
Generation mode
Describe it, and it does the typing
When you are ready to change the estimate, switch to Generation mode. Describe what you want and the assistant drafts it: a new section, a missing trade, a correction across many lines. It proposes the edit, you review the diff, and nothing is committed until you confirm.
Describe a missing scope and it builds the items, structures them, and prices them through HAWK.
Change a spec note or restate a description across many lines in one instruction.
Point at work the drawings imply but the SoW never captured, flagged for your sign-off.
Every change arrives as a proposal you review and approve. Nothing is silently overwritten.
Verbal mode
Builds a SoW from a brief
Sometimes there is no Excel and no full drawing set, just a description of the job. Verbal Mode turns plain language into a structured Schedule of Work you can price, refine and export. The fastest way from an early conversation to something a quantity surveyor can work with.
Describe the building and scope in plain English. No spreadsheet or blueprint required to get a first SoW.
Measured items under sensible headings, slotting into the same workflow as an uploaded SoW.
Verbal output flows through HAWK pricing and the quantity engine, so every line carries a rate and a score.
Add detail, correct assumptions and tighten scope in the same conversation until it reflects the real job.
In practice
What it sounds like in practice
“Which items are flagged for review and why?”
Analysis lists every low-confidence line, groups them by the reason the engine flagged them, and links each to the quantity and drawing behind it.
“Add a section for external drainage and price it”
Generation drafts the measured items, structures them under a new heading, runs them through HAWK pricing, and presents the priced lines for you to confirm.
“Why is the blockwork quantity higher than expected?”
It traces the take-off to the walls it measured, shows the rooms and drawing references, and explains the logic so you can judge the number.
Capabilities
Built for real estimating work
Project-wide context
Works against your full SoW, take-off, pricing and source drawings, not a pasted snippet.
Streaming answers
Responses arrive as they are written, so a long breakdown starts reading immediately.
Drawing-aware
Reads the blueprint pages behind a project, so a quantity question is answered against its page.
Confidence on every line
Drafts and edits carry the same confidence scores and review flags as the rest of the estimate.
Bulk edits, one instruction
Restate, reprice or restructure many lines at once instead of dozens of manual changes.
Writes back to the SoW
Approved changes land in the live SoW, structured and priced, ready to export.
You stay in control
An assistant that proposes, never overwrites
An assistant that quietly rewrites your numbers is a liability, not a tool. Analysis mode cannot change anything. Generation and Verbal output is always proposed first and committed only when you say so. Every figure ties to a source you can open, with a confidence score that tells you how hard to look.
Confirm before it commits
Generated and edited lines arrive as proposals. You review and approve. The assistant never writes on its own.
Flagged, not hidden
When unsure, it says so and marks the line for review rather than presenting a guess as fact.
Traceable by default
Quantities link to drawings, rates link to the price library, every answer points to a source you can verify.
Put the assistant to work
Start a 14-day trial with 75 free credits. A card is required to begin and you are not charged until day 15, so you can run a real tender, talk to your SoW, and decide before anything is billed. Cancel any time before day 15.