Frequently asked questions
The obvious questions.
Engagement size, timelines, data handling, offshore staffing, model choice, working with startups. Answered here so the discovery call can be about your problem, not our boilerplate.
Engagement & pricing
Engagement & pricing.
What size of engagement is typical?
Our engagements typically run A$25K to A$250K. The Audit is fixed-fee at the entry point; Pilots are fixed-fee for the highest-value use case from the Audit; retainers are scoped per quarter. If your problem is smaller than A$25K, we’ll tell you so and point you at a better-suited provider.
Do you publish prices?
Not on the site. Every engagement is scoped to your context — your data maturity, governance posture, and the specifics of the workflow we’re automating. We share the figure on the discovery call once we understand enough to scope it accurately.
Fixed-fee, day rate, or retainer?
Fixed-fee for productised work (the Audit, Pilots). Day-rate or monthly retainer for ongoing capability — fractional CAIO, governance reviews, advisory. We don’t do hourly billing.
Do you sub-contract or staff offshore?
No. Every project is delivered by senior AU-based practitioners. If we can’t staff something senior in Australia, we don’t take the engagement.
Timing & process
Timing & process.
How fast can we start?
Discovery call within 5 business days of you booking. Audit kicks off the week after. Pilot kicks off the week after sign-off on the Audit deliverables. We commit to dates, we hit them.
What’s your typical engagement timeline?
Audit: 2 weeks. Pilot: 6–10 weeks. Practice Retainer: 3–12 months. Full lifecycle from discovery to handover for a typical Pilot is 8–14 weeks.
Do you do remote delivery?
Yes. Most of our delivery is remote-first by a distributed senior team. We travel within the Sydney–Melbourne corridor for kickoffs, executive readouts and major milestones. Other locations on agreement.
What if the Audit shows we’re not ready for AI?
We’ll say so in the readout. The Audit is genuinely useful even if the recommendation is “fix your data first.” The R&DTI eligibility memo and governance gap analysis are valuable artefacts on their own.
Data, governance & security
Data, governance & security.
Where does our data live during an engagement?
By default, in your existing cloud tenancy (Azure, AWS, GCP). We deploy into your environment; we don’t move data into ours. Where AU-only data residency is contractually required, we configure inference regions accordingly.
Do you have ISO 27001 / SOC 2?
Not yet. Our compliance roadmap is on the Trust page. For engagements that require formal certification today, we partner with certified infrastructure providers (Microsoft, AWS, GCP) and pass through their controls where appropriate.
How do you handle audit trails?
Every deployment includes an audit log of model inputs, outputs, confidence scores and decision metadata. The schema is designed to survive a Privacy Act 2026 disclosure request. We don’t ship AI systems that can’t answer the question “why did the model decide that, on what input?”
Who owns the IP we build together?
You do. Standard clause in our SOW: you own the bespoke code, the prompts, the configurations and the trained artefacts we deliver. We retain the rights to the methodology and the reusable components we bring to the engagement.
Fit
Fit.
Do you work with startups?
Rarely. Our practice is built for mid-market firms (50–500 staff) with established operations and a board-level mandate. We occasionally take Series-B-and-later companies with mid-market characteristics. We don’t do pre-seed or seed-stage advisory.
Do you work with international clients?
Not currently. We’re an AU-only practice by design. The regulatory context we know best is AU. International work is on the roadmap but not the offering today.
What if we already have an in-house AI team?
Many of our clients do. We typically slot in alongside the in-house team for the parts they don’t have capacity or specific experience for — governance design, R&DTI scoping, senior peer review on architecture, or executive-facing translation.
Why not Big-4?
Big-4 is the right call for some engagements — large transformations, public-sector procurement, board-level coverage on a megaprogramme. We’re built for the mid-market band where Big-4 is over-staffed and SMB shops are under-skilled. We’ll tell you on the call if you’re actually a Big-4 fit.
Tooling & technology
Tooling & technology.
Do you have a preferred cloud?
We default to Microsoft Azure for AU mid-market because most of the segment already has it. We deploy on AWS and GCP regularly. We don’t take cloud-allegiance positions — we build where your business already is.
Which models do you use?
GPT-4o / GPT-4-class for most generative work, Azure OpenAI for AU-hosted inference, Claude where reasoning depth matters, open-weights (Llama / Mistral) for cost-sensitive or sovereignty-critical workloads. Model selection is a per-engagement decision, not a vendor allegiance.
Do you fine-tune?
Rarely. RAG and strong prompting beat fine-tuning for most use cases and don’t lock you to a model version. We fine-tune when the workload genuinely requires it — domain-specific summarisation, structured-output classification at scale, or where retrieval is structurally impossible.
What if a new model breaks our system?
Every system we deploy has versioned model references, eval suites, and a rollback path. Model deprecations are a routine part of operating AI in production — we treat them as expected events, not crises.
Couldn’t find your question?
Ask on the call. No deck.
Thirty minutes with a senior practitioner. We’ll cover anything not answered here.