Training programs / For innovation & digital leads

For innovation and digital leads: taking practice from a pilot to a program

The questions a pilot meets on its way to scale, with the standard or document behind each answer.

Your questions

Pilot to scale, AI governance, data and agreements

A governance committee, a privacy officer and an IT reviewer each ask a different question. The answers point at the documents they already use.

A standup in front of a wall of notes at the start of a project, the discovery and consultation step. — A standup before a wall of notes.

A permits officer reviews a screen with colleagues. A VR training scenario for mining and resources. — Colleagues review a screen together.

People sort tokens beside an analytics display. A VR training scenario for immersive VR training questions answered. — People sort tokens beside an analytics display.

A vertical slice of a build reviewed on a laptop, the whole path working end to end. — A vertical slice reviewed on a laptop.

How does a pilot become a program?

Digital NSW’s innovation buying journey: “Approaching it in stages is recommended when buying innovation.” PTR states the same aim: “PTR builds toward a program the customer’s own people run, with a record their own assessors read.” The stages are in how a pilot runs.

AI governance

Which frameworks will your reviewers apply?

The Australian principles and standard, and the framework United States reviewers reach for.

Colleagues review blank tokens for AI safety on a glass table. An AI roleplay scenario for custom AI experiences. — Colleagues review tokens on a glass table.

Australia’s AI Ethics Principles and the Voluntary AI Safety Standard

The Department’s current Guidance for AI Adoption (October 2025) is “updated and simplified guidance for industry” that evolves the standard’s guardrails and the principles; reviewers may cite either. The principles include privacy protection and security: “AI systems should respect and uphold privacy rights and data protection, and ensure the security of data.” The standard’s guardrails ask a deployer to “Enable human control or intervention in an AI system to achieve meaningful human oversight across the life cycle” and to “Inform end-users regarding AI-enabled decisions, interactions with AI and AI-generated content.”

A professional stands beside a floating control panel of screens and a keyboard tray, setting the boundaries her digital twin will answer within — A professional sets boundaries beside a panel.

United States: the NIST AI Risk Management Framework

The NIST AI RMF “is intended for voluntary use”; its functions are govern, map, measure and manage. PTR’s Didymo names it as “The framework behind the approved-knowledge and human-oversight language”.

Data and the AI

Where is our data stored, and what does the twin draw on?

The answers a privacy officer asks for, as PTR states them.

Builders stand among desktop stations. A rapid prototyping scenario for fast builds, from idea to working product. — Builders work among desktop stations.

Where data is stored and processed

Storage, compute and hosting run in an Australian region by default. “Where a client needs its data kept in its own country or on its own systems, it is stored where the client chooses.” In a live twin call, what travels is “The consented reference photograph, live microphone audio and camera video from the browser, the persona prompt, the transcript and any recording.” The register records that category as processing outside Australia: see governance, security and data.

People rehearse a twin handoff to a real person. An AI roleplay scenario for digital twins. — People rehearse a handoff to a person.

Approved knowledge, human oversight, human handoff

“Didymo’s answers are drawn from the organisation’s own approved material, not general knowledge picked up from the open internet.” “When a question needs a person, the conversation moves to them, with the context already in place.” See knowledge twins and expert access.

Agreements and a first build

What is settled before a build, and how does a first version get made?

Ownership and consent sit in the agreement. A working first version is how most pilots start.

Designers compare abstract character shapes on screen. An AI roleplay scenario for digital twins. — Designers compare character shapes on screen.

Scenario content, likeness and the record

A custom scenario is built from your situations and your people’s words, so the agreement names what each side brings and keeps. “Consent, copyright and ethical governance of a real person’s likeness or voice are settled before the build”. API and SDK access “is arranged under an enterprise agreement”: see for developers.

A colleague hands a working phone build across the table to the person who asked for it. A rapid prototyping scenario for fast builds. — A working phone build handed across.

From idea to a working first version

Fast Builds take an idea to a working version across web, mobile, games, XR and AI. A rapid custom build is scoped around your situations; a custom AI experience answers from your own material. The first session with your people is the test.

Where practice fits

The programs a pilot usually starts with, and every way they are delivered

A conversation your organisation already trains for, rehearsed with a person who answers back.

Training programs & delivery

PTR’s published work

What is published, and who measured it

A peer-reviewed paper, a client’s own program evaluation, and published research on a program PTR built. Nothing published yet covers an AI practice partner or a twin deployment.

An office desk scenario in the Google DEI training, an overheard conversation involving tokenism. — An office desk scenario from the Google training.

Google, Mater Education and Mayo Clinic

A peer-reviewed paper evaluated VR inclusion training led by PTR co-founder Leonie Sanderson and built with Equal Reality, a subsidiary of People Tech Revolution: the 7 trainees who returned the optional survey gave 82% average agreement that it was better than Google’s standard DEI training, and 94% average satisfaction. In Mater’s own program evaluation, over 96% of participants wanted more training in VR. Mayo Clinic published research on a DEI module for nurse leaders that People Tech Revolution built: read the papers.

Download the Google research paper PDF Download the Mater case study PDF