Level 1 · Research
Research-stage capability; substantial task-specific uncertainty remains.
Time horizon: [To be assessed; no forecast]
A practical, evidence-led method
Map the work. Generate the data. Integrate with care.
We start with the tasks and people in your operation, not a predetermined technology or robot.
Stage 1 · Workforce Task Audit
Map & Quantify
Map every relevant role, shift and location. Observe and break roles into timed tasks, including variation, exceptions and hand-offs. Build a cost-per-task business case and assess readiness without assuming a robot is the answer.
Methods
- Interviews and staff workshops
- On-site observation and task timing
- Shift, site and process variation mapping
- Cost assumptions and sensitivity checks
Outputs
- Role / shift / location task map
- Baseline timings and cost model
- Prioritised opportunity register
- Technology-readiness assessment and next steps
Illustrative task card — replace with measured customer data Task: Move a tray of clean linen from staging to a floor store Estimated cost: CHF [X.XX] per trip — placeholder Readiness: Level [X] — placeholder pending site assessment Recommendation: Observe route, load variation and lift access before a pilot
Physical AI readiness scale
Directional assessment only. Levels depend on the specific task and operating environment; horizons are placeholders, not promises.
Level 2–3 · Emerging / demonstrated
Parts of the task work in controlled or limited settings, with important conditions to validate.
Time horizon: [Indicative only; validate with vendors]
Level 4–5 · Operational / proven
Level 4: operational deployments under defined conditions. Level 5: commercially proven for comparable tasks and environments.
Time horizon: [Site-specific; evidence required]
Stage 2 · Training data
Generate Data
Plan and produce task-relevant data in the customer’s environment. Select capture methods that answer a defined technical question.
- Human demonstrations and teleoperation
- Video, depth and sensor capture where appropriate
- Simulation based on customer layouts and constraints
- Data quality checks, labelling and coverage review
Data governance comes first
Before collection, document purpose, roles, access, ownership, retention, permitted uses and deletion. Minimise personal data; anonymise or blur where appropriate. Complete the required privacy and workplace reviews before filming or sensor capture.
Customer ownership and any reuse rights are contract decisions, not assumed defaults. No collection starts without agreed safeguards.
Stage 3 · Operational integration
Integrate
Turn a promising task into a bounded pilot with safe working practices, clear responsibilities and an honest stop/go decision.
- Define task scope, baseline and measurable KPIs
- Assess hazards, safeguards and escalation paths
- Select partners and technology neutrally
- Onboard staff; monitor performance and incidents
- Scale only after agreed evidence and review
Useful KPIs
Task completion and quality · human intervention rate · uptime and recovery · staff experience · safety observations · total cost per task. Set thresholds with your team before the pilot.
Engagement shape
A timeline sized to the question.
Durations below are placeholders. Final scope, dependencies and timing are agreed before work starts.
Scope
Confirm sites, roles and data safeguards
[X weeks]
Audit
Observe, map and quantify tasks
[X weeks]
Prepare
Decide if data generation or pilot is justified
[X weeks]
Review
Agree evidence-based next step
[X weeks]
Frequently asked questions
Do we need to have a robot already?
No. The first stage assesses tasks and readiness without requiring a technology purchase.
What if a task is not ready to automate?
That is a useful conclusion. We may recommend process changes, another data step or no automation for now.
Who owns collected data?
Ownership and permitted use are agreed with the customer before data capture begins.
Can you guarantee savings or a successful pilot?
No. We make assumptions visible and define evaluation criteria, but outcomes depend on the real environment and technology.
Start with a no-obligation discovery call.
We aim to reply within two business days.