Skip to content
AtheronLABS

You're visiting from the United States. Prices are shown in US dollars. Not right?

labs@atheron:~/industries/construction/ai$ train agent --grounded

AI for construction

A specs and RFI assistant that cites the section.

Most RFIs ask something the documents already answer, if you can find the page. An assistant grounded in the project's own specifications and addenda finds it in seconds. We run it on an open-source model because contract documents and pricing are commercially sensitive.

// what matters here

What is different in this industry.

01

Cites the clause, every time

Each answer names the section and any addendum that changed it, so a project engineer can check it before passing it to the field.

02

Knows what was superseded

Addenda and revisions are tracked, so the assistant answers from the current documents rather than the bid set.

03

Your documents stay yours

The model runs on GPUs you control. Tender documents, pricing and correspondence are never sent to a model provider.

04

The engineer of record decides

The assistant finds and cites. Design questions still go to the consultant through a formal RFI, which it can help you word.

// example projects

Priced examples for this service.

Each one is hypothetical, labelled as such, and priced live from our rate card. Open any of them in the estimator and make it yours.

Example projectWeb appAIOpen-source model

Specifications and RFI assistant

Hypothetical. Not a client, not a result.

// The problem

Project engineers spend hours finding the clause, addendum or past RFI that answers a question from the field, and the same questions come up on every job.

// What we would build

An assistant that answers from the project's specifications, addenda and past RFIs and cites the section, on an open-source model served on GPUs the contractor controls, so tender documents and pricing stay private.

// What is in it

  • Specifications, addenda and RFIs indexed per project, with superseded sections marked
  • Answers that cite the section and the addendum that changed it
  • Suggested RFI wording when the documents do not answer the question
  • An evaluation set of real field questions, run before every change
  • An open-source model on rented GPUs, with no documents sent to a model provider

// Stack

  • Llama or Qwen (open-source)
  • vLLM
  • PostgreSQL with pgvector
  • Next.js
  • DigitalOcean GPU

// estimate

Build
≈ US$96,100 to US$147,000, delivered within 22 weeksCAD 136,800 to 209,100
Hosting
≈ US$4,360 a monthCAD 6,210 a month
Support
≈ US$1,000 a monthCAD 1,425 a month

Prices in your currency are estimates from today's Bank of Canada rate. All invoicing is in CAD or USD.

AI route
Open-source model
Model running cost
≈ US$4,190 a monthCAD 5,960 a month

Timeline by milestone

Discovery
2.9 to 3.4 weeks
Specification and evaluation plan
0.4 weeks
Design approved
2.4 to 3.4 weeks
Core features
1.4 to 1.9 weeks
Full build
0.9 to 1.4 weeks
Working pilot
3.9 to 6.9 weeks
Testing and fixes
0.9 to 1.9 weeks
Launch
0.9 to 1.4 weeks
Production
0.9 to 1.4 weeks

How it is paid

Deposit 20%
CAD 27,300 to 41,760
Discovery 1.8%
CAD 2,457.00 to 3,758.40
Specification and evaluation plan 10.5%
CAD 14,332.50 to 21,924.00
Design approved 1.8%
CAD 2,457.00 to 3,758.40
Core features 5.6%
CAD 7,644.00 to 11,692.80
Full build 3.7%
CAD 5,050.50 to 7,725.60
Working pilot 31.7%
CAD 43,270.50 to 66,189.60
Testing and fixes 1.8%
CAD 2,457.00 to 3,758.40
Launch 1.8%
CAD 2,457.00 to 3,758.40
Production 11.3%
CAD 15,424.50 to 23,594.40
Holdback, 30 days after launch (10%)
CAD 13,650 to 20,880
GPU time for training and testing, at cost
CAD 300. Billed up front, at cost, outside the milestones

// questions

Questions we get about this.

Can the specs assistant read drawings too?

It reads text on drawings and title blocks, and links to the sheet. Interpreting the drawing itself remains a person's job; we say where the assistant's knowledge stops.

How is a new project loaded into the specs assistant?

Upload the specifications and addenda, or connect the document folder. The project is indexed separately from every other, and closed projects can be kept for future reference.

Why open-source for contract documents?

Contract documents and pricing are commercially sensitive and often confidential under the tender terms. An open-source model on GPUs you control keeps them away from third parties.

// next

Not quite your project?

Tell us what you have in mind. We will come back to you with a range and the questions that would narrow it. Or book a call and talk it through.

AI specifications and RFI assistant for contractors | Atheron Network Labs