Skip to main content
SolutionsApproachCase StudiesInsightsContact

Spartan Waterproofing

From overnight replies to under 20 seconds.

←Back to Case Studies

October 2025 · Brisbane, Queensland

At Spartan Waterproofing the bottleneck was not the work. It was the follow-up. A Nemotron 3 Super draft lands in under 20 seconds — open-weight model, Queensland-located infrastructure, no third-party API call. Mario reviews on his phone and sends before the lead cools.

Where the time was leaking

Spartan did not need more admin. It needed follow-up to move at the same speed as incoming work, without losing Mario's judgement.

Lead arrival

Enquiries land while Mario is on site, driving between jobs, or closing out paperwork.

Response drift

Replies had moved from minutes to hours, sometimes overnight.

Customer trust

Every response still had to sound like the operator customers trust.

Before

After

Response time

Hours to respond

Under 20 seconds

Starting point

Starting from scratch

Draft ready when the enquiry arrives

Coverage

Enquiries waiting in the queue

Every enquiry answered same day

Operator load

Writing replies after hours

Reviewing drafts in the flow of work

The system Mario sees

An AI response system that drafts replies in Mario's voice, a phone-ready review portal, and a cleaner enquiry path around the workflow.

AI Drafting

Contextual first drafts based on previous enquiries and Mario's review patterns, ready for fast human sign-off.

Review Portal

Mario reviews, edits, and sends from his phone while the original enquiry context stays in view.

Outcome Tracking

Outcome tracking captures which conversations become booked jobs, so Mario can see what works.

Technical approach

Vector search pulls the five most relevant past enquiries into each draft, so every response starts from relevant context.

Mario's approval stays in the release path. His edits are captured so future drafts can learn from them.

Processing stays on Queensland-located infrastructure. Mario's enquiry data does not leave Australia and does not pass through a third-party LLM API.

Results

Draft ready in

Under 20s

Draft ready for Mario to review before the lead goes cold.

Human judgement

Every message is reviewed before it reaches a customer.

Operator load

Reviewing quality drafts replaces blank-page writing.

Learning loop

Edits and real job outcomes sharpen the workflow.

“Every enquiry gets a quality response the same day. I can take on more work without bringing on more people.”

Mario from Spartan WaterproofingMario — Spartan Waterproofing, Brisbane. 25 years in the trade.

When this pattern earns its cost

Fast-response revenue

Leads lose value when replies wait hours instead of minutes.

Owner judgement

The operator still reviews the message before it reaches the customer.

Repeat enquiries

Past jobs, edits, and outcomes give each draft real context.

Outcome pressure

The business cares which communication patterns win work.

Related reading

  • →Human in the Loop AI: where the human approves before the message sends
  • →APRA CPS 234 applied to AI — the same architectural posture, scaled to a bank
  • →Nemotron 3 Super vs GPT-OSS 120B — the model Spartan runs, and why

Make response time a conversion advantage

AI-drafted replies, operator-approved before send

Scope a response system

Site footer

Company

Entity
Ryder AI Pty Ltd
ABN
24 681 083 983
Founded
2024
Base
Brisbane, Queensland
Data boundary
Australian data boundary

Site map

  • Solutions
  • Approach
  • Case Studies
  • Insights
  • Contact
  • About

Trust & transparency

  • Security
  • Accessibility
  • Press
  • Site Map
  • RSS
  • Privacy

Contact

[email protected]

© 2026 Ryder AI Pty Ltd

LinkedIn (opens in a new tab)Privacy Policy