What to Expect if You Take Part

1. Choosing how to work

At the beginning you choose a lead companion and basic settings. This is an entry point, not a permanent lock-in. You can change your mind later.

• Choose your language first (for example, English or Ukrainian).
• Choose a lead persona: Charlie (steady, direct, peer-like) or Mishka (warm, reflective).
• Add your name and context so the assistant can orient the conversation to your current situation.
• Your persona is a starting point only; over time the style can adapt to what helps you orient, regulate, and reflect.

You stay in control. The companion adapts to your preferences over time, but you can always change persona, language, or how "deep" you want conversations to go.

2. Contact rhythm, participant code, and continuity

The system can carry continuity over time through a participant code. This helps the companion remember enough context to keep support coherent across sessions, without relying on fixed stance options.

• The participant code links sessions so support can stay consistent over time.
• The companion can tune pace, tone, and prompts based on what has helped before.
• Reflection is invited when appropriate, but never forced or pre-set by a stance toggle.
• Contact schedule is intentional: more structure early on, then lighter pacing as self-direction grows.
• After a period of attunement, the assistant can become more familiar in style and timing: metaphor when useful, academic framing when useful, and method choice matched to the person.

The long-term aim is one relational core with transparent adaptive behaviour grounded in continuity, user choice, and clear boundaries.

3. What the companion will and will not do

The assistant is designed to support early contact, not to replace human care or crisis services.

• It will: check in regularly, reflect back patterns, suggest small grounding or planning steps, and help you notice change over time.
• It will not: make a diagnosis, pressure you into talking about trauma details, or give commands about what you must do.
• It will prompt you with options rather than demands, and respect "I do not want to talk about this now".
• If useful during stress, it may offer a brief alternative perspective later in the journey, once participant-code continuity is established, and only with your consent.

You can also use the companion for everyday topics (sleep, routines, food, training plans, future goals) when that feels more useful.

4. Boundaries and support signposting

The assistant is a non-clinical reflective support tool. It does not diagnose, classify, or provide crisis intervention. When language suggests significant strain, it slows down and signposts external human support in plain language.

• If you sound under strain, it may slow pace, simplify choices, and suggest short orienting steps.
• If your wording suggests you may need more than this tool can offer, it will acknowledge limits and encourage contact with appropriate human support.
• It does not perform clinical risk assessment or assume responsibility for user safety.
• The user remains in control: pause, stop, or change topic at any time.

You can always ask "why are you suggesting this now?" and the assistant will explain its reasoning in plain language.

5. Data, consent, and stepping away

This is a research prototype. Any use in a real study would follow a clear consent form and ethics approval.

• You would be told exactly what is stored, for how long, and who can see it.
• You would be able to stop taking part at any time without having to give a reason.
• Where possible, your data would be minimised, de-identified, and used for research purposes only.

For this prototype demo, no real personal data is collected or stored.

6. Reflective support (optional)

Later in the journey, if things feel stable enough, you may be offered the option to use short reflective prompts.

These later-stage prompts depend on participant-code continuity and are not an initial setup option.

• Containment: Spirit Guide style grounding and values reminders.
• Exploration: Philosopher style questions about meaning and dilemmas.
• Integration: Researcher style summarising of "what helped" and small next steps.

This is always optional, can be paused at any time, and is used only when language and psychoeducation suggest reflective work is likely to help.