Case Studies | AI Systems We Built and Their Returns | ComplxAI
Case studies

Systems we've built and what they returned

ComplxAI designs and builds custom SaaS platforms, agentic workflows and production AI systems for Australian businesses. This page sets out what we built for each client, the capabilities behind each build and what it returned. Some clients are named. The rest are described by sector only.

  • Client-reported outcomes
  • Built and shipped by one small senior team
  • IP yours from day one of any paid stage

Work spanning EdTech, NDIS allied health, FinTech, PropTech, media, AgTech and security

ModeduLuxpipDisabilityAssessmentsBillyQAL VPN
Before you read

How to read these

Every return on this page is the figure the client reported to us, in the measure they used: revenue, ARR, share of operations automated, manual workload removed. We publish it as given rather than restating it in our own terms. Several engagements are under NDA; for those we give the sector, the shape of the system and the result, and withhold the name. Under each entry, the capabilities used link to the service page that covers them, and how we work explains how a similar build is scoped and priced.

What each entry shows

  • The return as the client reported it, in the client's own measure
  • Named clients with their name and logo; NDA engagements by sector only
  • The capabilities used, each linked to the service page that covers it
  • A link to the full case study where one is published
EdTech · Academic platform

Modedu

modedu.com.au

For Modedu we rebuilt the product as a custom SaaS platform with its own admin panel and ran agentic workflows through the operations behind it. On the product side, AI generates personalised academic papers, and a custom whiteboard was designed around agents from the outset.

  • Custom SaaS platform & admin panel
  • Agentic workflows across operations
  • AI for personalised academic papers
  • Agentic-first custom whiteboard software
Capabilities used
Read the Modedu case study
The return
2× revenue
within 6 months of going live
$400k → $1M ARR
Healthcare · NDIS allied health

DisabilityAssessments

disabilityassessments.com.au

DisabilityAssessments is an NDIS allied health provider. We built the administration around that service as one system: custom onboarding flows and a central knowledgebase at the front, automated matching and alerting in the middle, and end-to-end invoice processing for providers and plan managers with automated reconciliations at the back, all on secure storage for patient data.

  • Custom onboarding flows & central knowledgebase
  • Automated matching, alerting & invoice processing
  • Automated provider & plan-manager invoicing
  • Automated reconciliations
  • Secure patient-data storage
Capabilities used
Read the DisabilityAssessments case study
The return
90% automated
operations run on 10% human resource
$1–3M / year (client-reported range)
FinTech · Trading

Luxpip

luxpip.com

Luxpip was a complete build across the whole product: a custom design system, web and mobile apps, an administrator panel, agents handling internal operations, and a custom B2B API that gives Luxpip's own customers a direct way to integrate.

  • Custom design system
  • Web app, mobile apps & admin panel
  • Agentic solutions for internal operations
  • Custom B2B API for their customers
Capabilities used
Read the Luxpip case study
The return
−75%
manual workload removed
+40% scale-up with a growing customer base
More of our work

Five more systems, two of them named

Shorter entries for the platforms we can name and the engagements we cannot. The capabilities behind each are linked so you can read how that kind of work is scoped.

Billy

Custom platform · Operations

A custom platform built for Billy, with agentic workflows running through its core operations rather than sitting beside them.

Capabilities used
Routine operations carried by the system
Read the Billy case study

QAL VPN

Security · Quantum + AI

Quantum-encryption algorithms paired with AI, engineered into a VPN.

Capabilities used
Scaled to a global launch
Read the QAL VPN case study

Media company

Media · Sydney (NDA)

An agentic-first infrastructure designed and implemented for a Sydney media company, with agents at the centre of the architecture rather than added afterwards.

Capabilities used

Real-estate platform

PropTech · Brisbane & Sydney (NDA)

A fully agentic real-estate solution built for a PropTech platform operating across Brisbane and Sydney.

Capabilities usedRead the PropTech case study

Autonomous drone startup

AgTech · Autonomous drones (NDA)

Perception and decisioning for an autonomous agricultural drone startup: the models that let the aircraft see and decide in flight.

Capabilities used
Went global, expanding into the US
The pattern

The returns, side by side

All three are the clients' own figures in their own measures. Alongside them sit the global launches and the VC raises we have supported clients through. Different sectors, one small senior team, and software the client owns in every case.

revenue in 6 months (Modedu)
90%
of operations automated (NDIS health)
−75%
manual workload (Luxpip)
Questions

Questions people ask about this work

If yours is not here, bring it to the free intro call.

Book a free intro call
Who measured these outcomes?

The clients did. Each return shown against a client is the figure they gave us, in the measure they used, and we publish it without restating it as a claim of our own. Where a client gave a range, such as $1–3M a year, we show the range rather than a single number.

Why are some entries anonymised?

Those engagements are under NDA. We can describe the sector, the shape of the system and the result, but not the client's name or any detail that would identify them. Named clients carry their name and logo; NDA engagements carry a sector label instead.

Did ComplxAI do this work itself?

Yes. ComplxAI is a small senior team that does the work itself, remote-first across Australia and onsite where it helps. Strategy, design, engineering and delivery on each of these projects came from that team.

Who owns what was built?

The client, from day one of any paid stage. Documentation and handover are included in every engagement, there are no mandatory retainers, and every client has the option to take the software in-house.

How would a similar project start for us?

With two free conversations: a 30-minute intro call, then a one-hour pre-discovery. If it makes sense to continue, Discovery runs for 2–4 weeks from a fixed fee and produces a use-case map scored by impact, effort and risk. Every stage after that is priced in writing before it begins, and you can exit at any boundary. The full model is on our how we work page.

Do you only work in these industries?

No. The projects here span EdTech, NDIS allied health, FinTech, PropTech, media, AgTech and security, and we have dedicated pages for professional services and retail operators, for healthcare and real estate businesses, and for construction and hospitality firms. What carries between them is the engineering, not the sector.

Start with a conversation

Bring us the number you need to move

Tell us what you are trying to change and we will tell you whether we can build it. The first two conversations are free, and every stage after that is priced in writing before it begins.