AI adoption does not move through org charts. It spreads through prompts, spreadsheets, shared folders, inboxes, shortcuts and deadline pressure.
Industrial Linguistics builds simulations and lightweight work probes that show how AI is actually changing work before policies, registers and governance committees catch up.
We find the AI-shaped work your governance register cannot see.
AI Work Radar finds the work your AI policy cannot see. We use simulations, bottom-up prompts, shared-work artefacts and debrief evidence to surface shadow AI, spreadsheet systems of record, AI-visible knowledge, unclear ownership and governance gaps.
Shared drives, spreadsheets, inboxes, transcripts and unofficial knowledge bases are becoming the raw material for assistants, copilots and agents.
Staff use agents, copilots, summaries, coding tools and private workflows in ways no central register can fully capture.
Ambiguous emails, stale files, hidden spreadsheet logic and missing context become operational risk when machines start reading and replying.
The useful AI conversation starts when people experience the system from the inside. Each simulation creates a small, playable version of a real organisational pressure: deadline shortcuts, hidden AI use, bot-readable communication, unclear ownership, incomplete telemetry or workforce shock.
A 90-day office-pressure simulation where official tools, peer shortcuts, outsourcing, partial telemetry and AI governance collide. Participants feel why safe paths must be usable, not merely approved.
Random calendar or SMS prompts interrupt real work and ask: “Could AI help with what you are doing now?” The captures become anonymous evidence for clustering, pilot design and governance conversations.
A tabletop forecasting game for AI-era decisions where teams must act with partial evidence, competing interpretations and live accountability.
A short simulation where teams try to meet new demand with today’s visible skills, then confront transition debt, cash burn and retraining tradeoffs.
Start with a short simulation or probe. Leave with a map of the work your AI policy cannot see and concrete decisions people can discuss.
We define the deadline, shortcut, shared artefact, missing context or governance gap that makes AI-mediated work hard to see from the top down.
The first version may be a simulation, prompt programme, browser screen, card deck, spreadsheet, script or mix of artefacts. The point is to make the work visible quickly.
Participants make ordinary choices under time, role, information and tooling constraints. The debrief captures what they notice, where they get stuck and what they trade off.
The replay becomes pilot design, policy discussion, capability planning, governance register input or a practical next step.
The same core design method can produce a short sales-room exercise, a field prompt programme, or a longer tabletop session.
For sales teams that need prospects to feel a problem before hearing the pitch.
For 60-120 minute sessions where teams need to practise a decision, policy or change.
For boards and leadership teams dealing with uncertainty, risk, AI governance or investment choices.
For AI adoption or discovery programmes where useful examples need to come from real work, not a brainstorm.
If you already run advisory, transformation or training work, Industrial Linguistics can help you add a simulation layer to it. That might mean adapting one of the existing games, designing a game around your method, or leaving you with facilitator notes and materials your team can run without me in the room.
Talk about consultant useBehind the simulations are practical instruments for seeing AI-mediated work: shared-folder mapping, spreadsheet system-of-record detection, bot-readable communication checks and local-first AI-use reflection. These are labs, prototypes and field experiments, not surveillance products.
Worker-visible instruments that help people see which shared files, folders, spreadsheets and messages are visible to assistants, copilots and agents.
Experiments for finding where spreadsheets have quietly become operational databases, where data is copied by hand, and where unofficial workflows carry real business risk.
Brand monitoring and persona panels can become market signals, stakeholder reactions, synthetic customers and debrief evidence for simulation rooms.
Board, cybersecurity, property and training research becomes source material for governance rooms, investment games and executive tabletop exercises.
Industrial Linguistics has worked with banks, technology companies, government, education, retail, professional services and high-change operating environments.
The client wall remains deliberately broad: the work has ranged from large institutions to specialist technology teams, government groups and smaller firms.
The simplest first project is a short AI Work Radar session: a simulation or probe that reveals bottom-up AI use, hidden work systems and practical governance decisions.