TRACKED CALLS / 24H1,284+12.4%SMS RECOVERIES / WK287+8.1%DEMAND REACTIVATION RESPONSE RATE41.0%+3.2ppAU + NZ CLIENTS LIVE512+11AVG QUOTE RESPONSE47 MIN-9 minISO 27001PLATFORMTRACKED CALLS / 24H1,284+12.4%SMS RECOVERIES / WK287+8.1%DEMAND REACTIVATION RESPONSE RATE41.0%+3.2ppAU + NZ CLIENTS LIVE512+11AVG QUOTE RESPONSE47 MIN-9 minISO 27001PLATFORM
GibsonFamily-Owned Sydney Lead Engine · Since 2006
1800 950 347Get a quote
LIVESANDBOXAI BUILDSCASE STUDYSYDNEY · MELBOURNE · BRISBANE · PERTH · AUCKLANDMon to Fri · 9am to 5pm AEST

Sandbox · What we deliver for operators who need a custom AI build

// Education platform at scale · client build

Learning Sphere

By Albert Triolo, Gibson Promotions ·
// THE BASICS

What is a serverless education platform?

A serverless education platform is a training and assessment system built on cloud functions that scale to zero when idle, eliminating the fixed operating cost of a traditional server-based LMS. Gibson Promotions built Learning Sphere for an Australian education provider running recurring training cohorts that needed assessment logic and multi-tenant isolation the off-the-shelf LMS market could not deliver. The platform runs on 35 Google Cloud Run functions backed by Firestore, with Vertex AI generating and grading assessments at consistent standard. Idle cost approaches zero because of scale-to-zero architecture, restoring the margins that per-seat pricing models had eroded at scale. Assessment generation, which previously consumed an operator's morning, now runs in seconds. The working prototype was built in one week in Sydney and is in BETA.

What a workforce enrolment engine does and why the chase is the value›

“The LMS market has two modes: too rigid to support the assessment logic the business actually needs, or priced per-seat in a way that turns profitable cohorts into break-even ones at scale. Learning Sphere is neither. It is serverless, multi-tenant from the architecture up, and built to restore the margins per-seat pricing had been eroding.”

Albert Triolo, founder of Gibson Promotions

“Assessment generation used to consume an operator's morning. Now it runs in seconds on Vertex AI, at consistent standard, for any cohort size. The idle cost approaches zero on Cloud Run because the platform is not running when nobody is using it. That is the architecture that makes education at scale viable.”

Albert Triolo, founder of Gibson Promotions
Education platform at scale · client build

Learning Sphere

architecture

SYDNEY · 2026

SYDNEY, 2026. An Australian education provider running recurring training cohorts across multiple programs hit the wall every LMS shopper hits. The off-the-shelf options were either too rigid to support the assessment logic the business actually needed, or priced per-seat in a way that turned profitable cohorts into break-even ones at scale. The build-it-yourself path looked like two million dollars and 18 months of engineering. The market for serverless multi-tenant education platforms in Australia did not have a category leader. It still does not.

A serverless education platform on Google Cloud Run with Firestore as the data plane. Thirty-five functions handle cohort orchestration, assessment scoring, certification issuance, and real-time progress tracking. Multi-tenant from the architecture up with strict data isolation per cohort. Vertex AI generates the assessments and grades them at consistent standard. Integrated with the client's existing payment and CRM stack.

Thirty-five Cloud Run functions in BETA. Multi-cohort scale demonstrated. Idle cost approaches zero because of scale-to-zero serverless. Certification issued in real time. Assessment generation, which used to consume an operator's morning, now runs in seconds. Architected to restore margins at scale.

BETA. Working prototype built in 1 week.

‹ Back to all Sandbox builds

// THE ALTERNATIVES

How does a serverless education platform compare with the alternatives?

  • Off-the-shelf LMS (aXcelerate, Moodle, TalentLMS)

    Good at holding training records once a learner is enrolled. Not built for custom assessment logic, multi-tenant data isolation, or flexible per-cohort pricing. Per-seat cost compounds as cohort scale grows, turning profitable programs into break-even ones.

  • Build it yourself (traditional stack)

    The market estimate for a custom multi-tenant education platform in 2026 was two million dollars and 18 months of engineering. Serverless architecture on Cloud Run collapses both figures. The Learning Sphere prototype was built in one week.

  • SaaS course platform (Thinkific, Teachable)

    Designed for single-brand course delivery, not multi-tenant operator cohorts with custom assessment scoring and compliance requirements. Per-seat pricing, shared infrastructure, and no path to custom assessment or grading logic.

  • Serverless custom build (Cloud Run + Vertex AI)

    35 Cloud Run functions with scale-to-zero idle cost. Vertex AI generates and grades assessments at consistent standard. Multi-tenant from the architecture up with strict data isolation per cohort. Built and owned, not rented.

// FREQUENTLY ASKED

Frequently asked questions

What is a serverless education platform?

A serverless education platform is a training and assessment system built on cloud functions that scale to zero when idle, so the infrastructure cost approaches zero between cohorts. Gibson's Learning Sphere runs on 35 Google Cloud Run functions backed by Firestore, with Vertex AI generating and grading assessments at consistent standard per cohort.

Why build a custom LMS instead of using an off-the-shelf platform?

Off-the-shelf LMS platforms are either too rigid for custom assessment logic or priced per-seat in a way that erodes margins at scale. Learning Sphere was built because the market lacked a multi-tenant serverless education platform for the client's specific cohort model. The build cost and timeline beat the 2026 market estimate significantly.

How does Vertex AI generate assessments in Learning Sphere?

Vertex AI generates assessments tailored per cohort and grades submissions at the same standard each time. Assessment generation that previously consumed an operator's morning now runs in seconds. Every grade, every certificate, and every submission is audit-trailed in Firestore so compliance records are system-generated rather than maintained by hand.

How fast was the Learning Sphere prototype built?

The working prototype was built in one week. It runs on 35 Google Cloud Run functions and uses Firestore as the data plane. Multi-cohort scale was demonstrated at prototype stage. Gibson built Learning Sphere in Sydney as a bespoke client engagement; the platform is in active BETA.

What does scale-to-zero serverless mean for education platform costs?

Scale-to-zero means the platform's Cloud Run functions stop running when no cohort is active, so idle infrastructure cost approaches zero. Traditional LMS hosting carries fixed server costs regardless of usage. At cohort scale, the serverless model restores the margin that per-seat SaaS pricing or fixed hosting had been consuming.

// MORE ON THIS
// MORE SANDBOX BUILDS

Want one like this in your business?

Every Sandbox build started as one operator's bottleneck. Tell us yours.

A 30-minute diagnose call, no charge. We tell you whether it is a 48-hour build or something bigger, before you commit a dollar.