Skip to content

Live demo · Packaging & manufacturing

NTS QuoteDesk

Turn a packaging RFQ email into a structured enquiry, a missing-spec list and a draft reply.

Runs in your browser. The validation, millimetre conversion, missing-spec checks, review status and draft reply follow the same rules as our n8n workflow; a small rule-based extractor stands in for the AI step.

Playground

Try it yourself.

Synthetic data only — please don't enter real patient, claim or personal details.

1. Pick or write an RFQ email

You're a customer emailing a fictional Australian packaging supplier.

Fictional details only. Nothing is sent or stored; the whole workflow runs in your browser.

2. Watch the workflow

Process an RFQ to see each step: untrusted-input check, extraction, validation, missing specs, draft reply and the sheet row.

The structured RFQ appears here: customer, product, quantity, size in millimetres, material, print and delivery.

About this demo

What it does and how it runs.

Our RFQ intake workflow for Australian packaging suppliers, running on a fictional supplier. Each quote request is read as untrusted input, turned into structured fields, checked with fixed rules, compared with the specs needed to quote, and given a draft reply for a person to approve.

  • TypeScript
  • Rule-based extraction
  • Validation rules
  • Google Sheets

Things to try

  • Process the “Vague enquiry” sample: the draft asks only for the specs that are missing.
  • Try “Inches and 5k”: the sizes are converted to millimetres and “5k” becomes 5,000.
  • Try “Prompt injection”: the instructions in the email are flagged and the reply stays a holding message.

In a live setup (not in this demo)

  • In a live setup an OpenAI model with a strict JSON schema does the extraction, so free-form emails are read far more reliably than by the rules used here. Everything after that step works the same way.
  • RFQs arrive from the supplier's Gmail or Outlook inbox, every result is logged to a Google Sheet, and approved replies are saved as email drafts. Nothing is sent automatically.
  • Required specs, product types and reply wording are set for each supplier. Prices, lead times and stock are always left to the estimating team.

An NTS AI Lab internal demo with a fictional supplier and customers, built to show how NTS QuoteDesk works. It is not client work.

More demos

Try another one.

All live demos
  • Clinics

    Missed-Enquiry Recovery

    Watch a clinic enquiry get an instant reply, a front-desk flag and a place in the daily summary.

  • Lead management

    LeadFlow Starter

    Send a test enquiry and watch it get recorded, answered and followed up.

  • Trades

    QuoteFlow for Tradies

    Talk through a job on site and get a job card, a three-option quote, a safety form draft and a quote chase list.

  • Customer support

    AI Support Assistant

    A website assistant that answers from approved FAQs and knows when to hand over.

  • Healthcare RCM

    Claim Denial Triage

    Mask PHI, classify a denial and draft the appeal — step by step.

  • Healthcare billing

    AI Claim Amount Calculator

    Turn messy claim status notes into pending, paid and open totals.

  • Careers & HR

    AI Job Match Agent

    Compare a résumé with a job ad — evidence-based, no invented skills.

  • Healthcare RPA

    Payer Portal Bot

    Watch an RPA bot log in, check claims and recover from errors.

Want a tool like this for your team?

We build automations like these around your own systems. Tell us what you'd like to automate and we'll reply by email.