AI-Powered Quote Assistant Development

Quote Assistant – Project Brief

is a commercial joinery manufacturer based in Queensland. We manufacture toilet partitions, lockers, whiteboards, pinboards, and other joinery items using compact laminate, melamine, and aluminium systems.

We receive RFQs (Requests for Quotation) from builders and architects, usually by email with attached PDF drawings and specifications. Preparing quotes is time-consuming and often repetitive.

We want a “Quote Assistant” tool to help our estimators create draft quotes faster, while keeping human review in the loop.

Objective

Build a Quote Assistant that:

1. Accepts incoming RFQ documents (emails + PDF attachments).


2. Extracts key details (scope, quantities, materials, compliance notes, exclusions).


3. Produces a draft quote in Word format, using our standard template and clauses.


4. Flags any missing or unclear information as “clarifications” for the estimator.



The estimator will then review, adjust numbers, and send the final version.

Deliverables

1. Input Handling

Simple way to feed RFQs into the system:

Option A: Dedicated folder (e.g., “/RFQ-IN/”) in OneDrive/SharePoint.

Option B: Email forwarding

2. Data Extraction (AI/LLM)

Identify customer name, project name, site location.

Extract items to be quoted (e.g., toilet partitions, lockers, materials specified).

Recognise standards (e.g., AS1428.1 ambulant compliance).

Produce a short summary of scope, inclusions, and exclusions.

List any missing details as clarifications.

3. Quote Draft Output

Generate a draft Word document (DOCX) using our branded template.

Auto-fill:

Job reference

Scope summary

Standard clauses (e.g., design freeze, variation rates)

Clarifications & exclusions section

Save to project folder (e.g., “/Quotes/”) with filename JOBID_ProjectName_DraftQuote.docx.

4. User Review

Estimator opens the Word draft, checks and edits before sending.

No direct sending to customer (human in loop).


Technical Notes

We currently use Microsoft 365 (Outlook, OneDrive/SharePoint, Word).

Drafting is done in Microvellum, optimisation in Ardis, ERP is Manager (Internet IT). This project does not need integration with those systems yet — just focus on the quote assistant.

AI engine can be Azure OpenAI (preferred, since we use Microsoft 365) or another suitable LLM.

The solution must be private — no customer data should be used for public AI model training.

Success Criteria

An estimator can drop an RFQ email or PDF into the system.

Within 5 minutes, a Word draft quote is generated in the correct folder.

Draft quote includes:

Customer/project details

Scope summary

Clarifications & exclusions

standard clauses (supplied by us)


Output requires only minor adjustments before sending.

What We Will Provide

Word template with branding and standard clauses.

Example RFQs (emails + PDFs) for testing.

List of common clarifications and exclusions.

What We Expect from the Freelancer

Advise on best technical approach (SharePoint automation, Azure Logic Apps, Power Automate, or custom Python script).

Build, test, and document the solution.

Train 1–2 staff (basic handover, recorded screen share is fine).

Estimate total cost (fixed price preferred) and timeline.

Budget & Timeline

Pilot project: 4–6 weeks.

Budget: Open to proposals — please provide a fixed or staged cost breakdown.

How to Apply

Please include:

1. Relevant experience with Microsoft 365, AI integrations, or document automation.


2. Example of a similar automation project you’ve delivered.


3. Suggested approach and high-level cost estimate.

This project is phase 1 of a broader AI roadmap (drawing compliance checks, factory QA, forecasting). Good performance on this pilot may lead to follow-on work.