A joint transformation program — TrustedClients × Fedix

Front of house.
Back of house.
One AI-first firm.

TrustedClients rebuilds the front office — the business model, the pricing, the pipeline. Fedix builds the AI back office that delivers the work — onboarding to lodgement — at a leverage that finally makes "a firm that runs without you" operationally real.

Not generic SaaS — installed & customised in your firm AI prepares. A human reviews & lodges. Built on real AU tax & compliance workflows

01 — The moment

Every firm feels the squeeze. Almost none have a plan.

The Australian profession is in a genuine bind — and it is not one you can hire your way out of. The firms that install a delivery system now don't just survive the squeeze; they turn it into margin.

55%
Tax-accountant vacancies that actually get filled. Enrolments in the Accounting Professional Year collapsed 7,122 → 340.
CA ANZ member survey, 2025
51-75%
Share of total compliance time that 39% of AU firms burn just creating & managing workpapers.
Xero, 147 AU practices, 2025
$76k+
Lost per firm per year to unbilled out-of-scope work. 41% write off invoices to avoid the conversation.
Ignition State of Client Engagement
78%
Of AU firms plan to raise fees in 2026. The market will pay more — if you can deliver more.
Ignition 2026 AU Pricing Benchmark

The AI paradox: 68% of AU firms have already piloted AI — but only 22% have any strategy for it, and 88% fear it will undermine professional judgement. The whole profession is in FOMO and stuck. They don't need another tool. They need a system, the guardrails, and someone to install it. (IDC/Caseware; Thomson Reuters; IDC 2025)

02 — The partnership

Two engines. One transformation.

One transformation, delivered by two engines that don't overlap. TrustedClients rebuilds the front office and the business model — how the firm wins clients, prices, positions as advisory, and runs without the owner. Fedix installs the AI back office that delivers the work. The Turnkey Model™ promise structurally requires both — the growth and the delivery engine to make it real.

Front of house

TrustedClients

The growth & business-model engine
  • Growth engine — managed marketing (LinkedIn, content, ads, website) + a sales pipeline that wins high-value clients on demand
  • Repricing & repackaging — off hourly onto value pricing and high-fee offers; earn per client, not per hour
  • Compliance → advisory shift — reposition as a value adviser / CFO: $1M in fees from fewer than 50 clients
  • Business model, operations & KPIs — the "System & Scale" layer, so the firm runs without the owner
  • Owner freedom & mindset — out of the bottleneck: from working in the firm to owning it
  • Done-with-you — "implementation over information": they install the engine, not hand over a plan
The handoff
Back of house

Fedix

The AI delivery engine
  • The production line — the actual tax, compliance & bookkeeping delivery TrustedClients never touches
  • End-to-end — onboarding → document chase → prep → review → lodgement → billing
  • One installed stack — CRM · MyLedger · FediPilot, customised to your firm
  • AI does the drudgery — OCR, coding, workpapers; the human keeps the judgement
  • Capacity per head — deliver more per accountant, without hiring or headcount cuts
  • Advisory, made deliverable — frees the hours that make the "fewer clients, higher fees" model real

TrustedClients grows the top line and repositions the firm; Fedix cuts the cost and the hours to deliver every return. Combined, that is the profitability jump — the firm sells more and delivers it for less, at the same time.


03 — The automation framework

One client lifecycle. Eight stages. Every one runs on a real engine.

This is the whole firm as a pipeline — booking to billing — where each stage is powered by a specific, working Fedix capability. Click any stage to see the pain it kills, what's automated, and where a human stays in charge.

Automated by AI AI does the work, human decides Human judgement, AI assists
04 — The three engines

Not a demo. A working stack.

The pipeline above isn't a slideware promise. It runs on three products already built and proven on real Australian client files.

Fedix CRM

Production
The practice operating system
  • AU entity model + AUSTRAC KYC/AML + ethical clearance
  • Quote → native e-sign → invoice → payment → dunning
  • AI document-chase portal with automated reminder cadence
  • Live ATO sync (prefill, client lists, lodgment status) via GovReports
  • Job Kanban, per-practice data isolation, ~690-tool AI Copilot
~690 AI tools, access-tiered & confirmation-gated on every write

MyLedger

Production
Shoebox → clean books
  • AutoRecon: AI bank-transaction categorisation + smart sort
  • Unified import, chart of accounts, opening balances as journals
  • BAS, ITR, Financial Statements, Trial Balance, Div 7A
  • Depreciation engine: Div 40/43, DV & prime cost, roll-forward
  • Immutable report snapshots for audit & history
Minutes to turn years of messy transactions into a reviewable ledger

FediPilot

Proven on real files
The preparation agent
  • Email → OCR → classify → workpapers → draft return
  • Real tax cascade: MLS, CGT, rental, depreciation, Div 7A
  • Proven end-to-end: Individual, SMSF, Trust, Company, 9-entity group
  • Judgment gates: hard-stops & asks a human on high-stakes calls
  • Fills LodgiT / Xero Tax as a draft — never auto-lodges
$5,600 loan-interest over-read caught by the review pass on a real client file

05 — The trust architecture

The AI never presents a guess as a fact.

We put this to three of the people who'd actually live with it — a $2-3M firm owner, an admin manager, and a hands-on tax accountant. Their number-one demand was the same: keep the human in control exactly where judgement lives. That demand is the product. Here's how we answer each.

For the firm owner

One accountable owner

No vendor seam to point across. A single accountable delivery owner spans front and back office — and a written data covenant: your data stays in-country, is exportable on demand, is deleted on exit, and is never used to train models.

"Sell me safety and cash, not intelligence. One throat to choke."
For the admin manager

The Chase Cockpit

One screen: every client with outstanding docs, what's missing, last contact, opened / uploaded — auto-stops the moment they upload or reply-email the docs. A scheduled-send queue you can preview and hold, a global pause, a VIP do-not-auto-contact flag, and dunning that always needs a human click.

"Never chase someone who already sent it. Never nag a VIP."
For the tax accountant

Provenance on every figure

Click any number → the exact source-doc page, highlighted → the calc that produced it → a tie-out to the ledger and prior year. Every value is tagged extracted / calculated / client-provided / assumed — and every assumption is gated, plus a "what I could not determine" list.

"Software that tells me what it doesn't know is software I trust."
06 — Radical honesty

Live today, and on the roadmap.

The owner told us plainly: he trusts a roadmap panel more than a demo. So here it is, without spin. This line is where credibility is won.

Live today

  • AI document-chase portal + automated reminder cadence + SharePoint sync
  • Quote → native e-signature → invoice → payment → dunning
  • AUSTRAC KYC/AML, entity setup, ethical clearance
  • AutoRecon bank reconciliation + BAS / ITR / FS / depreciation
  • FediPilot email→workpaper→draft-return, proven on real cases
  • Live ATO prefill, client-list sync & lodgment-status read (GovReports)
  • AI fills LodgiT / Xero Tax as a draft; judgment gates hard-stop

On the roadmap

  • Direct electronic lodgement to the ATO (today: prepared & filled as a draft; a human reviews and lodges)
  • Deeper two-way sync with existing practice-management systems (XPM / FYI / APS)
  • Expanded judgment-gate coverage (CGT concessions, trust streaming, GST mixed-supply)
  • Per-figure source-highlight drill-down across every workpaper type

07 — How it lands

Customised, installed, and certified — not shipped.

A firm-specific rollout over one to three months. We schedule the install outside peak lodgement season, and start with a limited pilot before any firm-wide switch is flipped.

Phase 0
Week 0

Discovery

  • Map clients, entities, current stack & TAN/software IDs
  • Pain & leakage audit → your baseline ROI
  • Pilot cohort selected (20-30 real returns, off-season)
Phase 1
Weeks 1–4

Install

  • Stand up CRM, MyLedger, FediPilot — isolated per practice
  • ATO / GovReports sync, client import from ATO relationships
  • Migrate clients; wire SharePoint; connect existing systems
Phase 2
Weeks 3–8

Train & certify

  • Role-based, task-based training on your real data
  • Each role runs live clients end-to-end before go-live
  • Client-facing one-page upload guide shipped with invites
Phase 3
Ongoing

Run & optimise

  • Monthly support + a real support line in peak season
  • New-feature rollout & quarterly ROI review
  • Continuous tuning to your client mix
08 — The curriculum

Trained by role. Trusted by proof.

Nobody is asked to trust the AI because a slide said so. Each role earns trust by watching it work on real files — including watching it deliberately fail a judgment gate.

TRACK A

Owner / Principal

Half a day · dashboard, economics, pricing
What you'll run
  • The firm dashboard & pipeline KPIs at a glance
  • Reading capacity, realisation & profit-per-partner
  • Pricing advisory with confidence (the fee conversation)
  • Interpreting the live ROI & leakage recovery
Why it matters
  • You see the leakage close in real dollars
  • You stop being the bottleneck on judgement work
  • You govern the AI's boundaries, not the keystrokes
Go-live gate: you sign off the pilot's measured hours-reclaimed before any firm-wide rollout.
TRACK B

Admin / Client Services

2–3 hands-on sessions · task-based, recorded, on your data
What you'll run
  • Booking, onboarding + KYC, quote → e-sign → invoice
  • The Chase Cockpit: nudge one, chase a batch, pause all
  • VIP whitelist, scheduled-send queue, sent-log lookup
  • Dunning with a human hand on every send
Built around your fears
  • Never double-chase; auto-stop on upload or reply-email
  • ≤3 clicks to chase one client; works on your phone
  • No double data entry — syncs your existing system
Go-live gate: you run 2–3 real, low-risk clients end-to-end before automation is switched on firm-wide.
TRACK C

Tax Accountant

3–4 sessions + a parallel-run shadow period
What you'll run
  • AutoRecon clean-up & the FediPilot prep pipeline
  • Workpaper review with source-doc drill-down & tie-outs
  • The judgment gates & review discipline
  • LodgiT / Xero Tax draft-fill, and the AI Copilot
How trust is earned
  • Run it against 5–10 of your own completed returns
  • Watch it hard-stop on a residency & a part-year MRE case
  • See the correction rate published — not a hero anecdote
Go-live gate: you run it in parallel with your manual prep on ~10 returns until every diff is explainable — then, and only then, unsupervised.

09 — The economics

What it gives back.

The program is customised, so your exact investment is set on a discovery call. But the return isn't abstract — move the sliders to your firm and read it off the ledger. Conservative, evidence-based assumptions; sources noted below.

Your firm

Drag to match your practice.
Estimated annual return — your firm
Preparation time reclaimed
$0
Document-chase hours recovered
$0
Out-of-scope & billing leakage recovered
$0
Value returned, year one
$0

Sell more. Deliver it for less. At the same time.

That is the whole thesis — and it's the only combination that actually moves a firm's profitability in the AI era. Let's map it to your firm.

Book a discovery session →