For growing Australian advice practices
Big-firm capex budgets not required — nor billions of tokens. Your own data, in Australia: freehold, not leasehold.
Arkie assembles the client picture — seven systems reconciled into one governed record, facts verified, strategy scored, ADW foundation laid. The adviser and paraplanner review and sign off. Nothing leaves without their approval.
When the meeting ends, Arkie keeps running — filing, implementation tracking, alignment checks. What was advised stays done.
Built on your Surity360 governed data — not free-wheeling, costly AI, and held in your own private database in Australia rather than pooled in anyone else’s data lake — so every answer traces back to a source you control.
The adviser makes the recommendation. Arkie handles everything around it.
8–12 hrs
returned to every adviser, every week.
When the brief is assembled before the meeting, the adviser walks in prepared. The judgement stays where it belongs.
Ask Arkie · the four-step adviser workflow
A 60-minute advice meeting carries 8–13 hours of work around it — before and after, across the whole advice journey. Arkie gets Discuss → Assemble → Prepare into pre-final shape before the meeting; then after it, files the outcome and prepares the implementation paperwork — so the paraplanner and adviser finalise and deliver the advice instead of assembling it from scratch. We prep. We don’t advise. Click any card.
The capacity problem, in numbers
Your stack stays where it is, and so does your workflow — the same process you run today, just a lot quicker and a lot less stressful. Arkie assembles and reconciles it into one record, and your governed-data foundation underneath makes every answer traceable to its source.
The connection labels show where Arkie takes each system as your integration matures — two-way sync where the platform allows it, a scheduled feed where it doesn’t. Most start as a read-only feed on day one, so nothing is written back until you turn it on.
Under the hood, that “one reconciled answer” runs through the same Bronze → Silver → Gold pipeline the world’s largest data platforms use. Each layer raises the quality of the data, and nothing from the layer before is thrown away:
The Silver layer is the reconciliation a data warehouse skips — a warehouse just copies your systems together and still leaves you three versions of the truth.
Speed over the wrong inputs isn’t a compliance strategy — it’s the same error, industrialised.
Seven systems can each tell you what they recorded. None of them can tell you whether BID was met, whether implementation matched the advice, or whether the client’s objectives have changed. That is an understanding problem — and it requires a different architecture. Most “AI for advice” bolts a clever model onto data nobody governs. Three ways to handle the same seven sources:
Copies your systems into one store and hands you dashboards. You now have three versions of the truth in one place — and you still have to pick one, by hand, with no record of why.
Points a language model at contradictory records and answers with confidence. A confident answer over data nobody reconciled is plausible, not provable — exactly at the margin that matters when the file is reviewed.
One declared value per field, every conflicting source kept and logged, every change dated and held for seven years. You can answer what was true on the date of advice — and prove it. Arkie never invents an answer: it operates on governed facts, deterministic rules, and an audit trail. Every answer traces to its source.
Processes don’t win audits — data does. And the same governed record is what lets an adviser advise with conviction instead of hedging against numbers they don’t trust: fewer reworks, faster reviews, and a book that is worth more at sale.
One piece of advice runs through more than one meeting. Arkie does the work at every step — accelerating it, keeping the systems in sync, and keeping your own governed data current, healthy and secure. The adviser keeps the judgement.
Arkie assembles the reconciled picture and the meeting brief before the adviser sits down — gaps flagged, drift quantified, compliance pre-flighted.
When a Record of Advice or SOA is needed, Arkie assembles the inputs — the ADW pre-populated with client facts, current position and contribution room, the strategy scored, the cost comparison built. The adviser and paraplanner complete the recommendation and its basis; with the foundation in place, that is the fast part. The SOA follows from that record, not from scratch.
The pack is ready for the presentation meeting; the client decides; Arkie files the outcome and opens the next step.
Arkie runs the implementation workflow — tracking the tasks and external waits, confirming the implemented position matches the advice and the agreed risk profile across HUB24, XPlan and Sharesight, and writing the results back so the whole stack stays in step. The adviser owns the instructions; Arkie tracks whether they landed.
The reconciled record stays current — every change logged for seven years, in your own bank-grade database in Australia. At the next anniversary, the journey reopens.
“Compliance by design” is the right instinct — but it tells you when to comply, not where the evidence lives twelve months later. You can have impeccable process and still fail the §961B(2) safe-harbour test, because the process ran across seven systems and none of them holds the whole chain. It only works when the data architecture is designed too. Most practices have the first. Almost none have the second — that is the layer Arkie is built on.
Not what the vendors built — what advisers say they want when they are honest about it, and where Arkie lands on each.
Point a language model at seven systems holding seven partial, contradictory views of the same client and it will produce something plausible and unreliable at exactly the margin that matters at review. Arkie is built the other way around.
Arkie never reads the raw stack. It reads one reconciled view of each client — every field carrying a source and a timestamp.
When systems disagree, the conflict is recorded, not silently resolved. Nothing is overwritten without a trail.
Deterministic questions get deterministic answers, each with the record it came from. Seven-year evidence chain by design.
That architecture is what changes where your data actually goes:
Ask Arkie. Built on Surity360 — your governed-data foundation for Australian advice.
You don’t replace XPlan, your CRM or your platforms — and if you want a new tool for sharper SOAs or file notes, add it. But every one of those tools is only as good as the data beneath it. Arkie gets that data right first.
Ask Arkie is not a CRM, not a platform, and not an advice document generator. It prepares the briefing — the licensed adviser gives the advice.
Arkie sits underneath the tools you already run and reconciles them. Nothing to rip out and replace — your stack stays exactly where it is.
A warehouse copies seven systems into one place and leaves you three versions of the truth. Arkie reconciles them — with provenance and Australian advice context — into one record where every value traces to its source.
Fact-find currency, fee and AAA validity, portfolio drift — answered by rules, not by burning billions of tokens guessing over contradictory documents. AI is kept for synthesis, on clean inputs.
Get the data right and everything downstream follows — the prep, the workflows and the implementation all run on a foundation you can trust. It is the layer the whole stack runs on, and the one that holds the value — so in advice it should belong to the adviser, not the product provider. That is AdviserOS.
Arkie’s prep runs on the same reconciled, governed data that powers a set of optional modules — turn on what your practice needs, leave the rest. Investment management and monitoring is where most practices start.
Run your own customised model portfolios on the wrap platforms you already use — HUB24, North — and monitor them against the indexes and peer funds you choose, across 1D, 4W, 3M, YTD, FY, 1Y and 3Y. Drift, drawdown and relative return surface on the adviser’s screen before the investment-committee call and the client review — evidence the BID file note can rest on.
When the model changes — one fund swapped for another — Arkie identifies every affected client, drafts a personalised ROA for each, captures consent through the secure portal, and tracks execution across HUB24 and North. A book-wide switch becomes a supervised batch, not a fortnight of paperwork.
Extend Arkie to SOA and ROA preparation (ADW, Strategy Explorer, ASIC Report 818 alignment); publish a white-labelled daily and weekly investment briefing for your clients; and give partners a firm-wide cockpit — action register, meetings log and metrics. Each is a flat per-practice add-on.
Every module reads from the same governed record — so a number on the monitoring screen and a number in the brief are the same number, from the same source.
The large groups are investing tens of millions in a governed data backbone — and saying, loudly, that you need deep capital, or their balance sheet, to scale and compete in the years ahead. They are right that the capability is real, and right that building it from scratch costs millions. They are wrong that you need millions, or a consolidator, to have it — or to scale.
Arkie is that backbone, productised: governed data, reconciliation, a seven-year evidence chain and the agent on top — delivered to a three-to-twenty-five adviser practice as a subscription. No capital raise. No equity deal. No licensee migration. Keep your ownership, keep your stack, keep your independence.
Preparing a single annual review is not one task — it is a chain of manual steps across every system, usually done by support staff, often offshore, for every client, every cycle. And when the assembled record is incomplete or out of sync, the cost does not show here — it surfaces later, in a reworked statement of advice.
Steps with the amber edge are the ones most prone to rework when systems are out of sync — redone uploads, caps and Supertracker reworked on corrected figures, every fact-find field re-keyed into XPlan by hand. This is where the time is lost, and the work Arkie removes by reconciling every system first.
Arkie runs the whole chain as one work plan — one reconciled record instead of a morning of logins.
So the statement of advice is built on facts that are right the first time — not corrected a week later, across seventy pages, when the fact find turns out to have been out of sync.
That rework — often a full day for a paraplanner and adviser — is the cost the manual chain never shows.
The assembly stops being a person’s morning. The judgement — has the client’s life changed, are they on track — stays exactly where it belongs: with the adviser.
A client review, the way it runs with Arkie — assembled, reconciled and evidenced before the meeting. Here is what happens after the question is asked.
Illustrative walkthrough · demo dataOne question, before the meeting — instead of a morning across seven systems.
Every figure traces back to the system it came from — with a timestamp. Not a guess from a retrieved document.
What you would have to prove if ASIC asked — surfaced before the meeting, not after.
Not a paragraph — the whole pack: household modelling, portfolio, compliance posture and talking points, every line traced to its source.
Proof
Arkie and the Surity360 foundation data-governance layer run every day inside a busy AFSL-licensed advice practice in Hobart — the pilot tenant since 2025 — against a real multi-platform client book of thousands of clients. It is designed by senior advisers with decades of frontline experience, so every workflow is shaped around how advice is actually delivered — not how a software team imagines it.
It is now onboarding a small number of like-minded independent practices — advisers who share the same belief, that better-governed data makes for better advice, and who want that capability on their own terms. The difference made in one Hobart practice is starting to reach others.
The preparation that used to run across seven systems and most of a morning is assembled before the meeting. The adviser reviews it and advises — the order of the day is back the right way around.
Pricing
Per-adviser pricing anchored to true operational value: A$150–200K per adviser (one to two onshore support staff). Standard pricing at A$900 / adviser captures about 5.4% of that value. Early adopters get 40% off, locked in for three years with CPI adjustments. All prices exclude GST.
A$900 / adviser / month
A$1,800 monthly practice minimum (2 advisers). Early-adopter pricing: 40% off (A$540 / adviser), locked in for three years with CPI adjustments.
Steward + any 2 add-ons
10% off the add-on prices. Pick the two that match the work you actually do.
Steward + every add-on
20% off the add-on bundle. The Insurance Module joins free at launch for any practice already on Sovereign.
The price does not scale with adviser count.
How we count advisers: relevant providers currently registered on ASIC’s Financial Advisers Register. Provisional advisers in their Professional Year, paraplanners and support staff are not counted.
Early-adopter rates, locked three years.
Steward (5 × A$540): A$2,700
All four live add-ons (A$2,900 × 0.8): A$2,320
Infrastructure pass-through (typical): ~A$1,000
All-in monthly (early adopter): ~A$6,020
Standard monthly (A$900 / adviser): ~A$9,020
Cumulative three-year early-adopter saving: A$72K
Infrastructure pass-through. Google Cloud and MongoDB Atlas costs are billed at our cost, no markup — typically around A$1,000 / month, depending on practice size and how much data changes each day. Itemised on the invoice. Each practice runs in its own dedicated Google Cloud project and MongoDB Atlas cluster.
No surprises. No usage meters, no overages on documents processed, no per-API-call surcharges, no automatic price rises mid-contract. Annual prepay earns one month free. Multi-year and licensee pricing is available on request.
First three practices per quarter (until further notice) — implementation fee waived, A$540 / adviser locked in for three years with CPI adjustments. Qualify by completing a discovery call, an initial assessment and an approved application.
Data sovereignty
This is the difference the AI-first tools can’t match: your data lives in your own private database, onshore, reconciled with provenance — not held inside someone else’s. The questions principals ask first.
Data sovereignty. The AI-first tools run on data they hold and you reach through them; Arkie runs on data you own — your own MongoDB cluster in Australia, reconciled with provenance, portable, and yours to leave with. They save the same admin; the difference is who holds your client record. We build the governed data layer first, and the AI sits on top of it.
In your own dedicated MongoDB Atlas cluster and Google Cloud project, in Sydney — provisioned for your practice alone and billed to you directly at cost, the same as the MongoDB cluster. It is never pooled into a shared data lake, never co-located with another firm, and no client data leaves Australia. Your own instance, your own bill, no margin on the infrastructure.
Neither. Arkie connects directly to your systems of record — XPlan, your CRM, HUB24, North, MyProsperity, SharePoint — through authorised connections (OAuth, API, SFTP). During the trial all connections are read-only: Arkie reads, reconciles and presents, but writes nothing back. If you decide to continue, write-back is gated: you first review how Arkie resolves conflicts when systems disagree on the same field (the merge policy) and which BID compliance rules are active for your practice (the Guardian settings). When you are satisfied with both, you enable write-back system by system — only for the systems you choose, and you can turn any of them off again at any time. Client information moves from the source system straight into your own governed database; it does not travel through email or sit on a laptop as an uploaded file. Where a client’s confirmation is needed, it is captured through a secure token-based portal, not free-form email — fewer copies of client data, fewer endpoints, and a cleaner audit trail.
Not until you say so, and not until two gates are passed. The pilot is read-only throughout — Arkie connects to XPlan, your CRM, HUB24, North, MyProsperity and SharePoint, reads from them, and builds the governed record. It surfaces the brief, the Guardian compliance posture and the reconciled data for your team to test. Nothing is written back. If you decide to continue, you review the merge policy (how Arkie resolves the conflict when XPlan and your CRM disagree on the same field) and the Guardian settings (which BID compliance rules are active for your practice). Once you are satisfied with both, write-back is enabled system by system — you choose which systems it applies to, and you can withdraw consent for any of them at any time. The connections you do not enable stay read-only permanently.
Yes. Because it lives in your own private database, your reconciled client record and its seven-year evidence chain are yours to export or have returned in full. Data sovereignty means no lock-in: you walk away with your data, not a partial copy held inside someone else’s platform.
No. A warehouse copies your systems into one store and hands you dashboards — you still have three versions of the truth and have to pick one. Arkie reconciles the sources with provenance and Australian advice context: one declared value per field, every conflict logged, every change kept for seven years. Reconciled, not copied.
No. Keep XPlan, your CRM and your platforms exactly as they are — Arkie sits underneath and unifies the data. If you later add a tool for SOAs or file notes, it works better, because the data beneath it is finally clean.
Keep it — a sharp advice document in minutes is real time back, and Arkie makes a tool like that better, not redundant. But a document is only as good as the data it is built on, and it is a snapshot of one moment: it does not check that what it recommended was implemented, or that the model portfolio on HUB24 still matches the risk profile in XPlan and Sharesight a year on. Arkie governs the data underneath the document, and stays on the job after the advice — so the recommendation is carried out and stays aligned, with fewer reworks and a file that holds up at review.
No. Your data stays in your own database. Most advice questions — fact-find currency, fee validity, allocation drift — are answered by deterministic rules, not a language model. Where AI is used it is for synthesis over your own governed data, reviewed by the adviser, and never to train a shared model.
Access is gated by two-factor sign-in, role-based permissions and a per-action audit log — every read and write attributable to a named person, onshore or offshore. Infrastructure is ISO 27001 and SOC 2 Type II certified, encrypted in transit and at rest, with secrets held in a managed secret store, never in code.
Yes. Arkie runs on Google Cloud — the same world-class infrastructure that scales to global workloads — with Cloudflare’s security edge in front of it. Capacity scales elastically with your book, so it stays fast under load and you never outgrow it: powerful enough for a practice of thousands of clients, secured to enterprise standard, and entirely onshore.
Ask them of every tool you trial — including this one. A clever model is easy to demo; what you are really buying is what happens to your client data underneath it.
Onshore or offshore? Held in your own private database, or pooled inside the vendor’s platform with everyone else’s? And if you leave, do you walk away with all of it, or a partial copy? If the answer is vague, the data is not really yours.
Reconciled with provenance — one declared value per field, every conflicting source kept and dated — or just copied into one store where you still pick the truth by hand? A confident answer over data nobody reconciled is plausible, not provable.
Connected straight to your systems of record through authorised read-only links, or reliant on someone uploading files and forwarding emails? Every manual upload is a copy of client data on a laptop, and every email is an endpoint to defend.
Our answers are on this page. Most AI-first tools cannot give the same ones without rebuilding what they are — onshore, your own private database, reconciled, connected. That is the difference the demo never shows.
Built by advisers, for advisers.
Australian advice, understood from the inside.
The prep, the compliance evidence and the workflows are shaped by the people who deliver the advice — for the people who deliver it.
About Ask Arkie
Ask Arkie is built by Australian advisers for Australian advice. Not adapted from a US RIA tool, not localised from a UK paraplanning platform — designed from day one against ASIC’s BID standard and the live Australian platform stack. It was specced, broken, and rebuilt inside a Tasmanian advice practice for two years before a single line of marketing copy was written. Every capability on this page was carved out of a real Tuesday afternoon — not sketched on a whiteboard.
Start with one screen
A 30-day pilot, scoped to one workflow — about four hours of your team’s time. All connections start read-only: Arkie reads from your systems, builds the governed record, and lets your team test the brief and Guardian compliance posture against real data. Nothing is written back during the pilot. If you decide to continue, you review how Arkie resolves conflicts between your systems and the compliance rules active for your practice — then enable write-back system by system, only for what you choose, only when you are ready. If your team does not use it, that is the answer, and you have lost nothing.
Book a discovery callArkie prepares, reconciles and drafts for adviser review. It does not give personal financial advice. All advice remains the responsibility of the licensed adviser. Data sourced from external systems reflects those systems at the last sync and must be verified before use in advice documents.
Get in touch
Tell us a little about your practice and we’ll be in touch to arrange a time. Your details go nowhere until you press send.