Most asset finance brokers in Australia run a PPSR search at three points in a deal: pre-approval, settlement, and (if something looks weird) somewhere in between. At every point, the workflow is roughly:
- Open the AFSA portal in a new tab.
- Re-authenticate (the session timeout is short).
- Search by ACN.
- Download the certificate PDF.
- Drag it into the deal file.
- Maybe screenshot it for the credit team because the certificate's typography is from 2009 and gets lost in long deal folders.
- Forget which version is the latest if you've run the search more than once.
For a broker who runs ten to thirty searches a month, that's an hour to two hours of dead time. Not searching — just workflow. The actual AFSA call costs A$2 and takes 1.5 seconds.
The reseller math
The existing PPSR resellers (InfoTrack, Equifax PPSR Online, GlobalX, illion) all noticed this problem fifteen years ago. They wrap the A$2 AFSA call in workflow — integrations with practice management systems, CRM hooks, branded reports — and charge A$20 to A$30 per call. For the conveyancing market that pays for a 25-year suite of search products from the same vendor, this is fine.
For an asset finance broker who only needs PPSR + ABN, it isn't. You're paying enterprise-suite money to access a single A$2 register.
Three things we wanted to change
1. The PDF
AFSA's certificate is the legal artefact and you can't (and shouldn't) alter it. But the thing a broker drops in a deal file should be a one-page summary that names the deal, names the counterparty, hashes the certificate, and is verifiable later. We built that — call it the Due Diligence Record. It's deliberately not "the certificate" because that would be misleading. It's a wrapper.
2. The interface
A web form is exactly one of three reasonable interfaces. The other two are an API (for CRM webhooks and scripts) and an MCP server (for the broker who's started using their AI tool of choice to draft pre-approval emails and wants the assistant to run searches too). We are building all three surfaces: MCP first, REST/API and records at S2. Most products ship one and add the others "soon"; we want all three honest and well-shaped before any of them go live to paying customers.
3. The price
Our cost per search is A$2 (AFSA) + a few cents (infrastructure). The reasonable retail price is A$2.50 to A$3 — enough margin to fund the workflow layer, not enough to make a procurement officer wonder if they're being gouged. We landed on A$2.50 per call, included monthly allowances on the subscription tiers.
What we explicitly didn't do
Three things our competitors do that we don't:
- Individual-grantor searches. Our org-only scope doesn't cover them; we'd need a different reseller agreement. We don't, so we don't.
- Other registers. Title, court, ASIC company extracts. Adjacent but different problems. If you need them, InfoTrack and GlobalX are the right call.
- Credit reporting. CreditorWatch and illion do this well. We don't compete.
This is the honest scope. Saying no to features is half the work.
Why the MCP server matters
The most interesting thing about MCP isn't that LLMs can call your tools — it's that the same tool description gets read by every host. The string we write to describe ppsr_search_organisation is read by Claude Desktop, ChatGPT, Cursor, and Cline. Each of those hosts may handle confirmation differently, but the description is the same. So the description gets to be the canonical thing.
That changes how we think about tool design. The tool isn't an API endpoint with documentation alongside; the tool description is the documentation. It's marketing copy too — because if a broker's agent is choosing between three PPSR servers, the description decides which one gets called.
The first 50 design partners
We are pre-registering the first 50 design partners now, ahead of paid plans going live at S2-gate. Pre-registered partners get a planned lifetime price lock at A$199/mo (Pro) once paid plans open, in exchange for (anonymised, with your sign-off) case-study input on how the product fits your workflow. If you'd rather not be a case study, the planned price lock still applies.
Why? Because we're going to learn more from those 50 conversations than we'll learn from the next 500 marketing impressions. The product gets better the more we hear "I tried to do X and it didn't fit". Email us at [email protected] if you'd like to be one.
What's next
Webhooks (June). Salestrekker / Velocity / MyCRM native integrations (July). Per-customer encryption keys on Team (August). Status page (September). Localised Due Diligence Record PDFs in Mandarin, Vietnamese, Spanish (rolling, customer-driven).
Past the basics, our roadmap is decided by the design-partner conversations. Tell us what you'd build if you ran the product. We'll either ship it or tell you honestly why we won't.
By the Hoist Assets team. 2026-05-15. See the changelog · Get in touch.