--- title: "Applied AI for sensitive conversations, practised safely" description: "Digital twins that answer from approved material, AI roleplay for hard conversations and custom builds around your own knowledge, from People Tech Revolution." canonical: https://peopletechrevolution.com/ai --- AI # Applied AI for sensitive conversations. People are already taking their hardest questions to general-purpose AI. We build systems designed for that weight. THE BUILDING BLOCKS ## How does PTR build applied AI? The same parts appear in every build: a twin or companion that talks, the approved knowledge it draws on, the person who reviews it, and the workflow it hands into. A twin takes the first conversation. 01 ### Digital twins A version of a real professional that can talk to clients, answer questions and handle simple tasks around the clock. A companion is shaped around its purpose. 02 ### AI companions Domain-grounded conversational systems designed around a defined purpose, approved knowledge and clear boundaries. A hard conversation, held out loud. 03 ### Conversational agents Chat and voice assistants that hold real conversations, with configurable guardrails and escalation paths. Only what the organisation has approved. 04 ### Approved knowledge Retrieval over approved content so responses reflect the organisation's own policy, practice and expertise. A person stays over the answer. 05 ### Human oversight AI that extends practitioners' reach rather than replacing their judgement, with review and audit built in. The task carries on where it belongs. 06 ### Connected workflows Connect conversations to calendars, booking, follow-up and business tools. AI LANES ## What can PTR build with AI? A twin that answers as a named professional, a practice partner for hard conversations, and an experience built around one organisation's own knowledge. The same building blocks, arranged differently for each job. A professional faces her own digital twin. Digital twins: Didymo ### A digital twin for professionals with no time to spare Clinicians, advisers and educators answer the same questions every week. Didymo, PTR’s own product, answers them from the professional’s approved material, by chat, voice, a live avatar or a booked meeting, and hands over to the person when a conversation needs them. [Explore digital twins](https://peopletechrevolution.com/ai/digital-twins)[Learn About Didymo](https://peopletechrevolution.com/ai/didymo) Simon's Twin, ready to talk about PTR. ## Meet PTR’s CEO, Simon Lowe, as a digital twin. Talk about [VR training](https://peopletechrevolution.com/xr) or [AI](https://peopletechrevolution.com/ai) with Simon's Twin in chat, voice or a live avatar call. [Explore digital twins →](https://peopletechrevolution.com/ai/digital-twins)[Learn about Didymo →](https://peopletechrevolution.com/ai/didymo) Chat Voice Avatar Book Asking the twin from any screen. AI roleplay and practice ### Practise the conversation before it matters. A difficult conversation, rehearsed on a laptop or phone as many times as it takes, with the same briefing, structured review and human escalation PTR uses in its VR practice work. No headset needed. [Try AI roleplay now](https://labs.peopletechrevolution.com/) [More about AI roleplay](https://peopletechrevolution.com/ai/ai-roleplay-practice) > "AI plays a crucial role here as well, especially now, helping to script and gamify scenarios that provide a more realistic experience. It enables us to craft narratives that adapt in real-time, making the experience progressively more challenging and thus more effective." Simon Lowe CEO, People Tech Revolution, on AI inside PTR's scenarios. [The Cusp of a Revolution, Global Best Practice Group, October 2024](https://globalbestpracticegroup.com/airevolution/) A companion inside the boundaries set for it. Custom AI experiences ### Built around your knowledge, your people and your workflow. PTR combines the building blocks above, a twin or companion, approved knowledge, human oversight and connected workflows, with what an organisation already knows and how it already works. Every build moves through discovery and consultation, co-design workshops, prototyping and development, testing and deployment, and ongoing support. The full process is on [our approach](https://peopletechrevolution.com/about/approach). [Explore custom AI experiences](https://peopletechrevolution.com/ai/custom-ai-experiences) ## What does it look like in practice? The person the practising person faces, and the human adviser it hands over to. A colleague across the table, arms folded. The twin hands over to a human adviser. A colleague waits for your opening line. A manager across the boardroom table, mid-sentence. DIGITAL TWINS ## Meet the product by using it. Chat, a voice call, a live avatar on screen and a booked meeting are ways into the same conversation with the twin on this page. Each runs on the phone or browser you already have. [Chat Type a question on your phone or in the browser. The twin answers in text, drawing only on material the organisation has approved. Explore chat →](https://peopletechrevolution.com/ai/didymo/chat) [Voice Speak instead of typing. A voice call with the twin, on the phone or in the browser, in the twin's own voice. Explore voice →](https://peopletechrevolution.com/ai/didymo/voice) [Avatar A live face on your phone or laptop screen, for the conversations that need one. Explore the avatar →](https://peopletechrevolution.com/ai/didymo/avatar) [Book Book real time with the person the twin represents: a Google Meet appointment, made straight from the conversation. Explore booking →](https://peopletechrevolution.com/ai/didymo/book) Chat Voice Avatar Book Setting the limits of what it answers. ## Built for organisations where trust matters. Every build draws on approved knowledge, works inside boundaries the organisation sets, keeps a person over the answer and leaves a record that can be audited, with deployment and data-location options where available. [PTR's commitments on AI](https://peopletechrevolution.com/ai/custom-ai-experiences) ## Who is on the other end of a twin? A twin answers one person at a time. What changes is how many of those conversations the professional behind it can be part of. One adviser, a room full of clients. ### The same answer, given many times over A professional can hold one conversation at a time. A twin grounded in their approved material can hold the first part of many, then bring them the ones that need their judgement. Someone reading an answer as it arrives. ### And what it looks like from their side The person asking gets a reply at the hour they asked it, drawn from what the professional approved, and a way through to that professional when the question needs one. [How a twin is built](https://peopletechrevolution.com/ai/digital-twins) ## And at the hour they ask it Most questions do not arrive during office hours. The one that comes in late is answered from the same approved material as the one that comes in at eleven in the morning. A question asked with nobody at the desk. ### Nobody is at the desk, and the answer still comes A twin does not improve overnight or invent something to fill the gap. It answers from what the organisation approved, and the professional can read the whole exchange in the morning. AI IN XR AND MOBILE ## Can AI run inside XR and mobile apps? AI runs wherever your people are: a web page, a laptop, a phone, a VR headset or smart glasses. Only VR itself needs a headset; the twin and the practice partners above do not. [An expert twin inside a VR scenario. A staff member in a headset asks a specialist's twin from a regional office. Explore XR.](https://peopletechrevolution.com/xr) [A face that answers in the headset. The practice conversation, held inside the scenario instead of on a screen. Explore XR.](https://peopletechrevolution.com/xr) [A human adviser in the room. An older person in a headset beside a face, with an adviser sitting in. Explore XR.](https://peopletechrevolution.com/xr) [AR on a phone. A phone held over pipework shows the layer an app adds to the real thing. Explore mobile experiences.](https://peopletechrevolution.com/fast-builds/mobile-experiences) [An app on the warehouse floor. A picker checks a phone build between the racks, the kind of iOS or Android app a companion can live in. Explore mobile experiences.](https://peopletechrevolution.com/fast-builds/mobile-experiences) [A tablet in the waiting room. The check-in, tried in the room it is for. Explore mobile experiences.](https://peopletechrevolution.com/fast-builds/mobile-experiences) ## Practise with someone who answers back Put the headset on and the manager, the client or the upset customer is already in the room. Say what you would say at work. They respond to your words, not a script, so you can try the hard line a different way each time before it counts. Talking it through with a manager. A service conversation at the counter. Giving feedback at the meeting table. Facing a panel of interviewers. A first meeting with a new visitor. Raising it with a supervisor at the whiteboard. ### When more than one person is in the room Two colleagues in an open team area, and an adviser at a consultation table. The practice is the same: speak, and they answer. A team conversation in the open office. An adviser across the consultation table. ### After the conversation, the review The answer given is held up against the intended one, and a walk-through space shows how the parts of an experience fit together. Checking the answer against the intended one. Walking through the experience in the demo space. ## Where do the approved answers come from? Before a twin says a word, someone chose the material it may draw on and the lines it will not cross. Choosing the approved source it answers from. Setting what the twin may say. ## Written with the staff who will face it Your frontline staff bring the moments they dread. We turn them into scenarios, they test and critique each one, and it is rebuilt until it sounds like their workplace, before anyone else practises in it. Shaping the brief in a timber workshop. Testing the interview chair from the seat itself. A cafe service moment, rehearsed at the table. A designer checks the model at critique. Decide with these ## Which pages help you decide on AI? Start with how AI works, then the questions a review, a budget or a team will ask. ### Learn how it works Plain explanations of AI, digital twins, AI in XR and AI roleplay. [Sorting picture cards into trays at a table. What is AI? How AI learns patterns, where it fails, and when people must oversee it.](https://peopletechrevolution.com/learn/ai/what-is-ai) [A woman stands beside a floating panel. What is a digital twin? The engineering model and the AI that answers: what each term means.](https://peopletechrevolution.com/learn/ai/what-is-a-digital-twin) [A seated woman speaks, a panel beside her. AI and XR. XR builds the place and AI makes it respond.](https://peopletechrevolution.com/learn/ai/ai-and-xr) [An employee across the table, replies floating nearby. AI for training and roleplay. Repeatable practice for hard conversations, its limits, and where people stay essential.](https://peopletechrevolution.com/learn/ai/ai-training-roleplay) ### Decide and check The product pages, the program page and the answers a governance review needs. [A headset wearer faces an expert’s digital twin. Knowledge twins and expert access. Answers from approved material for associations, universities and scarce experts.](https://peopletechrevolution.com/training/knowledge-twins-and-expert-access) [A team maps a workflow around the table. For innovation and digital leads. AI governance frameworks, where data sits, and agreements, from pilot to program.](https://peopletechrevolution.com/training/for-innovation-and-digital-leads) [A team sorts tokens beside a coloured display. Governance and security. Where data is stored, how access is protected, and the document behind each.](https://peopletechrevolution.com/about/governance) [A wireframe twin stands beside a woman. Didymo. Chat, voice, an avatar or a booked meeting, answering from approved material.](https://peopletechrevolution.com/ai/didymo) [A customer mid-sentence, hand raised, at a desk. AI roleplay and practice. Rehearse a difficult conversation on a laptop or phone, then run it again.](https://peopletechrevolution.com/ai/ai-roleplay-practice) Questions buyers ask ## What do buyers ask before an AI build? Short answers, each with a link to the full answer. Rehearsing in the headset, a facilitator alongside. ### What does an AI roleplay partner do? It plays the other side of a hard conversation. A session runs as a briefing on the scenario and the objective, a live conversation in which the learner speaks or types and the AI answers, a structured review against set criteria, and the same scenario again. It can be replayed privately as many times as it takes. [AI roleplay and practice](https://peopletechrevolution.com/ai/ai-roleplay-practice) walks through the format. Knowledge passed to a wireframe figure. ### What does a knowledge twin answer from? From the organisation’s own approved material, not general knowledge picked up from the open internet. The professional decides what goes in and what stays out, reviews what the twin has said, corrects an answer when it needs correcting, and keeps the approved-source list as the boundary. [Didymo](https://peopletechrevolution.com/ai/didymo) shows how it works. Choosing between faces on the screen. ### How do people reach a twin? Through chat, a voice call, a live avatar on screen or a booked meeting, each on the phone or browser a visitor already has. Booking ends with a Google Meet in the diary of the person the twin represents, and the context of the conversation travels with it. [Didymo](https://peopletechrevolution.com/ai/didymo) walks through each way in. Reviewing controls at a glass table. ### What will our AI governance review ask? A governance committee, a privacy officer and an IT reviewer each ask a different question. Storage, compute and hosting run in an Australian region by default, or where a client chooses, and a subprocessors register lists each provider by category and location. [For innovation and digital leads](https://peopletechrevolution.com/training/for-innovation-and-digital-leads) maps Australia’s AI Ethics Principles and the NIST AI RMF to the documents behind each answer. A colleague talks it through in person. ### When does a conversation move to a person? When a question needs a person, the conversation moves to them with the context already in place, so it starts where the visitor left it. The twin is built to extend a professional’s reach, and the owner can read the whole exchange and correct an answer. [Knowledge twins and expert access](https://peopletechrevolution.com/training/knowledge-twins-and-expert-access) covers the handover. Practising in a headset against a green screen. ### Should practice be AI roleplay or a VR scenario? AI roleplay puts the counterpart on a laptop or phone, which suits staff spread across sites and shifts. A VR scenario places the learner in a convincing setting, with a sense of presence a flat screen cannot match, and AI can make the character in it respond. Each follows the same briefing, review and repeat pattern. [AI and XR](https://peopletechrevolution.com/learn/ai/ai-and-xr) covers the pairing. A coach talks the attempt through. ### Where does a human facilitator still lead? Conversational AI reads language, structure and pacing. It cannot read body language or cultural nuance the way a human facilitator can, so mentorship and peer discussion still do that work. PTR grounds each roleplay in the organisation’s own policy and practice, with human oversight and a clear path back to a person. [AI for training and roleplay](https://peopletechrevolution.com/learn/ai/ai-training-roleplay) covers where each fits. A blank head reviewed, clipboard in hand. ### What consent does a twin of a real person need? Consent, copyright and ethical governance of a real person’s likeness or voice are settled before the build. A digital twin then operates with consent, and with ongoing oversight of what it says on the person’s behalf. [Governance and security](https://peopletechrevolution.com/about/governance) lists what travels during a live twin call and how long information is kept. Bring us the AI problem this team is trying to solve. Open the contact form Ask our CEO Simon Lowe’s Digital Twin a question WHERE NEXT ## Where to go next in AI Each page goes deeper on one way PTR builds AI: the twin that answers for a professional, the practice partner for hard conversations, and the experience built around one organisation. [Digital twins. A professional's twin that answers when they cannot. Didymo answers from approved material by chat, voice or avatar, and hands over to the real person when it should.](https://peopletechrevolution.com/ai/digital-twins) [AI roleplay and practice. The hard conversation, rehearsed until it holds. Practise on a laptop or phone with a briefing, a practice partner and a structured review, no headset needed.](https://peopletechrevolution.com/ai/ai-roleplay-practice) [Custom AI experiences. Built around one organisation's own knowledge. Discovery, co-design, prototyping, testing and ongoing support, with a character and boundaries chosen with the people who will use it.](https://peopletechrevolution.com/ai/custom-ai-experiences)