What is Receipted AI?
Receipted AI is a new category of AI that doesn't just act — it proves it acted. Every booking, CRM write, triage decision, and system change produces an immutable, timestamped receipt that the business owner can review, export, and produce as evidence.
In a world where AI agents are taking real operational actions on behalf of SMBs, a receipt isn't a nice-to-have — it's the only way to know what your AI did.
The Difference: Call Logs vs Operation Receipts
Every AI voice agent in this market records conversations. That's a call log. What AliceHQ produces is different: a receipt for every downstream action the conversation caused. The distinction matters enormously when something goes wrong.
A call log tells you what was said. A receipt tells you what happened — the booking that now exists in your PMS, the CRM field that changed, the job card that was created in ServiceM8. You can't produce a call log in response to a TCIT audit request or a guest dispute. You can produce a receipt.
| Capability | Black-Box AI | Receipted AI (AliceHQ) |
|---|---|---|
| Handles the call | ||
| Records the call | Sometimes | |
| Produces booking receipt | ||
| Logs CRM write with action ID | ||
| Records triage decision | ||
| Evidence for TCIT audit | ||
| HIPC 2020 compliant | ||
| Can prove what AI said and did |
Why It Matters for NZ Service Businesses
The accountability gap in AI becomes real in three specific moments. Every one of these has happened to NZ businesses. The only defence is a receipt.
“When TCIT asks what happened”
The Tenancy Compliance and Investigations Team conducts 400–500 audits of property management companies each year. When they request your records for a maintenance interaction, “the AI handled it” is not an answer. A receipt showing the exact timestamp of the tenant call, the maintenance request created in Re-Leased, and the notification sent to the property manager is.
“A guest says your AI promised a price”
Air Canada's chatbot invented a bereavement discount policy that the company had never offered. The tribunal held Air Canada liable for what the AI said. NZ Consumer Law applies the same principle: your AI's representations bind your business. If a guest holds up their phone and says “your system quoted me this,” the only way to verify or contest it is with a receipt showing the exact interaction record.
“What did it book at what price?”
A trades business customer calls after hours, says the AI quoted $200, and the invoice arrives at $350. Without a receipt showing exactly what the AI communicated during that call — the job category, the quoted rate range, the booking confirmation — the business has no ground to stand on. With a receipt, the dispute resolves in minutes. That's what operational proof means in practice.
The Regulatory Tailwind
Several regulatory frameworks are actively creating mandatory receipt obligations for NZ and AU businesses using AI. These aren't future possibilities — they're current obligations that AliceHQ's architecture directly satisfies.
NZ Privacy Act 2020
Information Privacy Principles 2, 5, and 8 require businesses to limit data collection to its stated purpose, protect personal information from loss or unauthorised use, and ensure information is accurate. The Office of the Privacy Commissioner has explicitly stated that AI systems making or contributing to decisions affecting individuals should maintain an audit trail of those decisions. AliceHQ's receipts architecture provides that audit trail at the interaction level — not just the system level.
Australia ADM Transparency (December 2026)
Australia's Privacy and Other Legislation Amendment Act 2024 introduces new APP 1.7 requirements from December 2026: any regulated entity using automated decision-making that significantly affects individuals must disclose this in their privacy policy and maintain evidence of the decision-making process. For NZ businesses with Australian operations or customers, this creates a nine-month window to implement receipt-grade AI before compliance becomes mandatory. Full compliance breakdown →
RTA and HIPC 2020
The Residential Tenancies Act 1986 creates record-keeping obligations from the moment of tenant contact — including AI-handled contacts. The Health Information Privacy Code 2020 imposes prescriptive requirements on any system handling patient information. Both frameworks treat AI-generated records the same way they treat staff-generated records: if it happened, you need to be able to prove it happened, and prove it accurately. Vertical compliance guide →
Frequently Asked Questions
See Receipted AI in Your Business
A 30-day pilot gives you a full receipt log for every AI action. See exactly what your AI is doing — and prove it.