PR Practice Lab

Table of Contents

• The problem it exists to solve
• How it works
• Phase A design
• AI is part of the situation, not the assessor
• Evidence passport
• Current status
• Design lineage
• Licence

An evidence-driven, time-aware public-relations and communications practice simulator for people who cannot get their first industry role because they do not yet have the experience that first roles ask them to demonstrate.

Important

PR Practice Lab is fictional professional-learning software. It is not an agency, an employer, a live client service, supervised employment, a placement, a professional qualification, a guarantee of competence or employment, or accredited CPD unless a competent external body separately approves it.

The problem it exists to solve

Entry-level applicants are often asked for evidence of work they can only obtain after somebody hires them. At the same time, AI is reducing the amount of low-risk production work through which new practitioners traditionally learned: first drafts, monitoring summaries, media-list maintenance, transcription, content variants and routine reporting.

The answer cannot be a faster copy generator. PR Practice Lab creates a safe, repeatable place to perform the work around the copy: interrogate an incomplete brief, research and prioritise stakeholders, define measurable objectives, choose channels, handle approvals, verify claims, disclose and check AI use, respond to changing circumstances, advise against unethical requests, and explain decisions afterwards.

The product’s core claim is deliberately narrow: a learner can produce traceable evidence of performance in realistic fictional situations. An employer, mentor, educator or professional body must decide what weight to give that evidence.

How it works

Each authored scenario has a hidden, immutable reality and a learner-visible record that begins incomplete. The learner works through an inbox and bounded time/resource budget, asks for information, records assumptions and uncertainty, creates and revises communication artefacts, seeks approval, communicates with fictional stakeholders, monitors response and adapts the plan.

The assessed unit is a defensible professional act:

available evidence -> stated interpretation and uncertainty -> objective
-> audience and option choice -> authorised action -> response and consequence
-> adaptation -> reflection

The event record allows feedback to cite what the learner knew at the time. A good process can be recognised even when a fictional campaign meets adverse conditions; a lucky outcome cannot conceal fabrication, undisclosed interests, misleading measurement or reckless publication.

Phase A design

The first authored scenario is northstar-mobility-001: a junior practitioner joins a small agency team preparing the public launch of a fictional accessible community-transport pilot. The learner begins with realistic assistant work— brief hygiene, research, monitoring, stakeholder and media records, drafting, version control and approvals—then encounters conflicting executive claims, an accessibility challenge, a journalist enquiry and a changing launch decision.

Several defensible routes exist. The scenario does not reward one secret phrase or click sequence. It assesses research, planning, audience judgement, writing and editorial control, relationship handling, delivery, measurement, ethics, AI accountability, collaboration and reflective practice.

See:

• docs/foundation/foundation-spec-v0.1.adoc — product contract and invariants.

• docs/foundation/experience-progression.adoc — the route from novice tasks to integrated judgement.

• docs/assessment/competency-framework.adoc — observable capabilities.

• docs/scenarios/northstar-mobility-design.adoc — first scenario design.

• docs/rules/authority-registry.adoc — first-party professional references and their review status.

AI is part of the situation, not the assessor

Scenarios may permit, require or prohibit particular AI assistance. Permitted use must be declared, source material and personal data must stay within the scenario policy, and the learner remains accountable for verification, editorial judgement and publication. The same task can be run with and without AI to reveal whether the learner can direct, check and improve automated work.

No external AI service is required for the reference product. Hidden scenario reality and authoritative assessment must never be sent to a learner-controlled model.

Evidence passport

A completed run can export a watermarked evidence pack containing:

• the fictional brief and disclosed source set;

• selected artefacts with revision and approval history;

• the learner’s decision log, assumptions, uncertainties and escalations;

• a timeline of fictional stakeholder responses and adaptations;

• rubric observations linked to events rather than unsupported scores; and

• the learner’s reflection on what they would repeat, change or escalate.

The pack records simulated practice; it must never relabel elapsed simulation time as employment, placement hours, client work or accredited experience.

Current status

The repository now contains the Phase A product foundation, competency and assessment model, authority registry, system architecture, first authored scenario, deterministic native kernel/CLI and scenario-boundary validation. The CLI is an executable replay and assessment slice, not yet an interactive learner interface. The authority pack and all scored interpretations remain PROVISIONAL/UNREVIEWED pending practitioner, educator, accessibility and early-career validation.

Run the executable repository checks with:

zig build test
zig build run -- golden
bash tools/validate-scenario.sh
bash tools/test-scenario-boundary.sh

Design lineage

The epistemic and assessment structure is adapted from the evidence-driven model in metadatastician/sim-insolvency (now presented as Insolvency Tycoon): immutable authored reality, monotonic disclosure, beliefs distinct from facts, append-only events, plural defensible routes, separate competency and outcome, and a post-submission reveal. The professional content and progression model in this repository are original to public relations and communications practice.

Licence

Code, configuration and scripts are MPL-2.0. Documentation and fictional scenario materials are CC-BY-SA-4.0. Per-file SPDX declarations are authoritative.