Pricing · By line
How you pay follows the work,
not a menu.
We work three ways — Educate, Build, Run. Each line has its own commercial shapes: you choose the line that fits the work, and the format follows from it. The numbers below are illustrative ranges from real AU engagements; the final price is set at the scope-out call.
How you pay
By line, not by menu
The commercial formats aren’t a flat list you pick from. They distribute within the three lines we work in: pick the line that fits the work, and the payment shape follows from it. Every line works both ways on the model layer — already committed to a provider? We build on yours. Don’t want to manage one? We host and run it for you.
Educate
Teach your team to work AI-fluently.
How you pay
PAYG per engagement, or a fixed upfront engagement.
Best when the goal is capability in your people, not a system to run.
Explore Educate →Build
We build the tailored agent or workflow.
How you pay
An upfront paid build, or buy it outright and self-host — we hand over the codebase and IP. Built on the provider you’ve committed to, or one we host for you.
Best when you can name the workflow.
› De-risk on-ramp (for buyers burnt by a past pilot): we can scope and build the first thing at our risk, and you only pay if it works. It’s one way into Build, not a promise stapled to everything we do.
Run
We operate it in production.
How you pay
Managed, monthly.
Best when you want the outcome without owning the operations, monitoring and governance.
Explore Run →You pick the line, not the invoice. We don’t default-route you into managed monthly because that’s our preferred revenue. We start from the line that matches the work, your buyer DMU, and your internal AI capacity, and the commercial shape follows from there.
Where to start
Eight rungs, lowest commitment first, starting at Rung 0 (a free scope-out call). The middle rungs (4-6) are the three core commercial tiers. All numbers are illustrative ranges — final scoping done in your free call.
- 0
Free scope-out call
30 min
FreeTouchpointBook the call → - 1
Free workflow diligence document
1-page download
FreeTouchpointSee how we work → - 2
Workflow audit
5 business days
A$499TouchpointOrder the audit → - 3
Free 1-3 week build
1-3 weeks
FreeTouchpointHow it works → - 4
PAYG hosted
Per run
A$0.20-50 / taskCore tierSee tiers → - 5
Monthly hosted
Ongoing
A$200-15,000 / moCore tierSee tiers → - 6
Buy outright
One-time
A$8K-200K+Core tierSee tiers → - 7
Long-term partnership
12+ months
BespokeTouchpointTalk to us →
Hard rule: no cold entry above Rung 4. Engagements above A$20K need prior-rung evidence — for instance a prior paid workflow audit (Rung 2) or a completed free build (Rung 3).
The 3 Tiers
Augmentation, Automation, Agency — Anthropic’s three modes, productised. Pick the mode that matches your workflow shape. Each tier offers PAYG, monthly hosted, or buy-outright as commercial options.
Tier 1 — Augmentation
A/A/A mode: Augmentation
Best for
A single bottleneck workflow step. The work where one person, one prompt, one input source eats half the day.
Typical scope
1 agent, 1 prompt template, single integration
Examples
- Drafting first-pass email replies in your CRM voice
- Generating meeting summaries from Teams transcripts
- Coding-assistant skill tuned to your house style
We host. You pay only when the agent runs.
Fixed monthly. Includes monitoring + minor updates.
We hand over the codebase. You self-host on your own cloud.
Recommended
Tier 2 — Automation
A/A/A mode: Automation
Best for
A multi-step pipeline that today consumes a whole role's week. The work where 3-5 sequential agents pass structured data between systems.
Typical scope
Multi-agent, structured I/O, integration to ≥1 system (CRM / ERP / DMS / PMS)
Examples
- Insurance claims report pipeline (4 agents, claim → signed-off draft)
- Tax memo pipeline (intake → research → draft → cite-check)
- Real-estate listing pipeline (inspection notes → vendor report → listing copy)
We host. You pay per completed pipeline run.
Fixed monthly. Includes monitoring, MCP integrations, quarterly model upgrades.
We hand over codebase + MCP servers. You self-host.
Tier 3 — Agency
A/A/A mode: Agency
Best for
Full autonomous handover. The work where the agent operates on its own under policies you set — multi-agent + decisions + monitoring + escalation.
Typical scope
Multi-agent + decisions + monitoring + escalation + human-in-the-loop checkpoints
Examples
- Autonomous claims-triage agent across multiple insurance lines
- End-to-end contract-review agent with privilege controls + human review gates
- Autonomous compliance-monitoring agent with regulator-ready audit trail
We host + monitor + intervene. You pay per autonomous decision logged.
Fixed monthly. Full ops, observability dashboards, on-call escalation, quarterly governance reviews.
We hand over codebase + observability stack + governance docs. You self-host with our ongoing oversight option.
Pricing ranges are illustrative, not fixed. Final pricing set during your scope-out call — every engagement is custom-sized to your volume, complexity, and integration footprint. The framework is public; your numbers are tailored.
Other touchpoints
The rungs that bracket the three commercial tiers — entry points before committing to a build, and the long-term shape after one ships.
Free scope-out call
Free · 30 minutes
The conversion event. We talk through one workflow, sketch the A/A/A mode it fits, and tell you what a build would look like. No deck, no hard sell. Books straight to a real engineer.
Book the call →Free workflow diligence document
Free · 1-page PDF
One-page diligence brief on your workflow — workflow shape, agent pipeline, AU compliance footprint, illustrative ROI maths.
See how we work →Workflow audit
A$499 · 5 business days
Paid commitment-tier deliverable for buyers who want a written analysis before scoping a build. Documented workflow map, agent pipeline recommendation, ROI calculation, and a Go/No-go recommendation. Credited against any follow-on engagement.
Order the audit →Long-term partnership
Bespoke · 12+ months
What hosted Tier 3 engagements grow into — multi-agent fleets across multiple business units, governance review cadences, advisory or equity arrangements where it fits. Discussed once trust is earned through prior rungs.
Talk to us →Quick comparison
The three commercial options across all three tiers, side-by-side.
| Tier | PAYG | Monthly hosted | Buy outright |
|---|---|---|---|
| Augmentation | A$0.20–2 per task | A$200–800 / month | A$8K–20K |
| Automation | A$2–10 per task | A$1,500–4,000 / month | A$25K–60K |
| Agency | A$5–50 per task | A$5K–15K+ / month | A$75K–200K+ |
Pricing questions
Why is the build free?
Because the durable revenue is in hosting, not in one-off project margin. Claude Code compresses our build time so a free 1-3 week build is economically viable. We win when you stay hosted long-term — and you stay because the agent works.
What if I want to buy outright after going hosted?
Standard option. The hosted route lets you de-risk first and switch to buy-outright once you're sure of the ROI; the switch is available from the 12-month mark. Conversion price is negotiated against months of hosted spend.
What's included in monthly hosted?
Inference cost (Anthropic + cloud), monitoring, observability dashboards, quarterly model upgrades (when a new frontier model ships, you get it on day one), bug fixes, minor scope expansions. Ops escalation on Tier 3.
Are the tiers fixed or can we mix?
Mix. A full production setup runs a Tier 1 augmentation agent + a Tier 2 automation pipeline + Tier 3 governance/observability across both. The tiers describe build complexity, not a customer segmentation.
What's not covered?
Custom training on private data above ~10K examples (separate engagement, priced per project), specialised compliance certifications beyond Privacy Act + IRAP, and integrations to bespoke legacy systems that need their own MCP server (we can quote separately).
Pricing tailored to your workflow
30-minute scope-out call. Free. We talk through your real workflow, sketch the tier shape, and quote you the actual number. No deck, no hard sell.