Evidence
Everything that can be checked, on one page.
UnityTrip publishes no customer names. The organisations that run it operate under confidentiality, and reference conversations are arranged under NDA at evaluation stage. What it publishes instead is listed here: the artefacts a reviewer, a procurement team or a machine can verify without asking anyone.
Each item links to the thing itself. Where a page states a limit, such as the status page covering the website only or the SOC 2 position, the limit is stated here too.
How to evaluate
Seven questions, and where each is answered.
Evaluate UnityTrip as an operational system rather than as a travel booking tool, and read in this order; no single page describes the whole. Public means it can be checked without asking anyone. Conversation means it is answered at evaluation stage, under NDA where the item says so.
| Question | What to inspect | Public or conversation |
|---|---|---|
| What problem does it solve, and in which category? | The category definition: the seat has no fare, so policy rations demand where spend cannot. What UnityTrip is, and is not. | Public |
| How are an organisation's own rules represented? | The policy engine, and the published policy schema, version 1.2, which a rule has to validate against. | Public |
| Is the policy model more than a description? | The free Policy Builder, the sample policy it produces, and the Policy Simulator. | Public |
| Can the allocation behaviour be examined? | The 19-seat allocation demonstration on the Policy Builder page, the same engine turned to face the traveller on the page for travellers, and the simulator. All three run in the browser: they show the rules, not the production platform. The booking video, with its transcript, shows the deterministic engine deciding a request made in conversation. | Public |
| How does it fit the systems already run? | Integrations, the S/4HANA boundary with its diagram, and the page for travel management companies. | Public |
| What is the technical thesis, and how is it built? | The platform thesis, the architecture with the production diagram, and the API reference, webhook reference and four OpenAPI specifications. The API reference is illustrative: the samples illustrate, the specifications are authoritative, and the production model is larger and evolves. | Public |
| What should a buyer challenge or verify? | The numbers and their basis with the FAA source, the security page with its own controls and its plain statement on SOC 2, and the status page, which covers the website only. The rest is a conversation, listed below. | Public, then conversation |
What needs a conversation
- Reference conversations with operators running the platform, under NDA.
- The validation behind the benchmark figures, reviewed under a confidentiality agreement during due diligence.
- Service levels for the platform itself, set per tenant in the private offer and available on request.
- Security questions beyond the published controls.
- Commercial terms per tenant, beyond what the Marketplace listing shows.
- The priced Review, which returns a ranked list from your own movement and claim records.
To start that conversation, talk to us; we reply within one working day.
The machine-readable summary points to the same seven questions and lists every page in one file.
The platform, documented
Reference
API reference
The operational ontology and the service surface: passenger types, trip reasons, transportation modes, priorities, quotas, penalties and booking events. An illustrative representation of a larger model that evolves continually.
Read →Reference
Webhook reference
Claim lifecycle events and their payloads, as a subscribed system receives them.
Read →Schema
Leased travel policy schema, version 1.2
The published JSON Schema a leased travel policy validates against. The two channels and the two chokepoints are encoded in it. The test that separates the channels is a constant, and a rule declared approved that carries a quota does not validate; the Builder refuses a rule whose trip reasons span both channels before export.
Read →Specifications
Four OpenAPI specifications
Booking, approvals, policy engine and user manager, published as OpenAPI files alongside the API reference.
Read →Architecture
How the platform is built
Event-sourced and natively distributed on Azure, described in enough depth to be checked: surge design, tenant isolation, per-identity observability, constrained-network performance.
Read →Status
Public status page
A public status page for the website, hosted by Better Stack.
Open →Things you can run
Tool
Leased Travel Policy Builder
Free and without an account: turns leased-travel rules into a machine-readable policy that validates against the published schema.
Open →Sample
A complete sample policy
The JSON document the builder produces, as a booking engine would consume it.
Open →Tool
Policy Simulator
Runs bookings through a policy and shows the rules deciding: the approval-or-quota gate, booking windows, sanctions, a priority-sorted waitlist, displacement and go-show at departure, with every verdict naming the rule that fired. It runs entirely in the browser.
Open →Demonstration
Allocation demonstration
A 19-seat aircraft, simulated demand and an allocation engine running in the browser: priority, quotas and first-come-first-served side by side, with the policy switched on and off. The same pattern of rules the platform executes; a demonstration, not the production system.
Run →Video
A booking, end to end, with transcript
A complete booking made through the AI travel assistant, with the full transcript, and the deterministic engine deciding underneath.
Read →The numbers and their basis
Methodology
Benchmarks: 40% to 90%, and how it is measured
How the utilisation and no-show figures are measured, averaged across deployments and years, what a passenger movement counts as, and why no client-attributable utilisation figure is published.
Read →Source
FAA, Economic Values, Section 3
The published load factors behind the 40% baseline: Part 135 commuter carriers 51%, on-demand turboprops and mid-size jets 37–43%.
Open →Security and the company
Security
Security and trust
UnityTrip's own controls; the certifications, which are held by Microsoft Azure; a plain statement that UnityTrip does not itself hold a SOC 2 report; and where each tenant's subscription, SLA and data-processing terms sit.
Read →Policy
Human rights policy
Version 1.0, approved 28 August 2026, with a PDF to download.
Read →Procurement
Azure Marketplace listing
The transactable offer: purchase and billing through an existing Microsoft agreement.
Open →Record
Company record on Wikidata
The public entity record. UnityTrip is a trading name of NZBN 9429033178714, registered in New Zealand.
Open →Document
Company overview
Ten pages: the platform, the policy engine, outcomes, architecture, integrations and how to buy.
Read →People
Who built it and why
The founders, the mission, the two channels, and the lesson the previous generation of this software taught.
Read →Going further
Engagement
Travel Operations Review
Read access only. The ten largest recoverable opportunities in your operation, ranked by money, with the rows that prove each one. The findings are yours whether or not you proceed.
Read →Finding
One review finding, anonymised
A refund request taken at face value, and what the record showed instead: refunds per passenger and per segment, a ledger drifted from the card gateway in three ways, and the treatment in order. Shape only: no client, figures or routes, so it cannot be checked against an operation.
Read →Machine-readable
The site in one text file
llms.txt: the definition, a reading order for evaluators and AI assistants, every page with a one-paragraph description, and the questions and answers. Served with open CORS so an assistant can fetch it directly.
Read →FAQ
What can be checked about UnityTrip without talking to anyone?
The platform's documentation (an API reference, a webhook reference, four OpenAPI specifications and a published policy schema), a free Policy Builder, the sample policy it produces and a 19-seat allocation demonstration, a complete booking on video with its transcript, the benchmark methodology and the FAA source behind the 40% baseline, the security page with its own controls and its plain statement on SOC 2, the human rights policy, the Azure Marketplace listing, the public company record and the company overview. UnityTrip publishes no customer names; what it publishes instead can be verified without asking. Reference conversations with operators running the platform can be arranged under NDA at evaluation stage. The page opens with seven evaluation questions, the artefact that answers each, and the list of what needs a conversation; the machine-readable summary at /llms.txt points to the same seven questions.