More content, fewer decisions
AI produces alternatives faster than the organisation can assess and use them.
Human–AI collaboration is the deliberate design of value, information flow, human judgement and accountability across work performed by people and AI. Human-AI Cognitive Lean Thinking helps teams reduce cognitive waste and build a safer, more effective way of working with AI.
Digitalisation and AI · LeanTrix × Trigger Mind
Human–AI Collaboration – Human-AI Cognitive Lean Thinking
Better collaboration between people and AI begins with the quality of thinking. We diagnose cognitive work, reveal waste and design how tasks, checks and accountability should be shared between people and artificial intelligence.
A real client processCurrent and Future StateHuman judgement at the centre
More than promptingWe design a work system in which AI supports understanding, decisions and learning without replacing human accountability.
Faster does not always mean better
AI can accelerate a task while weakening the quality of the decision
In many organisations, artificial intelligence is added to existing tasks without redesigning the full workflow. Teams generate more alternatives, reports and recommendations, yet have no clear verification rules, decision points or answer to a fundamental question: what should a person do, and what is worth delegating to AI?
This creates cognitive waste: effort, information and attention that do not improve the value of the result, understanding, decision quality or learning. It is difficult to see, but it appears as delays, rework, stress, context switching and dependence on a model's output.
“AI does not remove the need to think. It raises the standard of thinking.”
Unclear division of work
It is not clear where AI assistance ends and human judgement and accountability begin.
Loss of context
Information moves between tools, conversations and versions without one coherent flow.
Learning is bypassed
The team receives an answer without building the understanding required for the next decision.
WASTE 01Information overproductionMore versions and content than the work actually requires.
WASTE 02Information inventoryMaterial accumulated without a decision or practical use.
WASTE 03Decision waitingAnalysis paralysis and unclear decision criteria.
WASTE 04Information transportCopying information between tools and people.
WASTE 05OverprocessingRefining output without improving the value of the result.
WASTE 06Cognitive motionConstant switching of attention, tools and context.
WASTE 07Thinking defectsIncomplete sources, hidden assumptions and unverified conclusions.
Design the work system
Human-AI Cognitive Lean Thinking
This original approach, developed by Dr Iwona Burka and Trigger Mind, combines Lean and Kaizen with knowledge about cognitive work, human judgement and collaboration between people and AI.
Add AI to the task
The focus is on model functions, automation or prompts. The workflow, evaluation method and accountability often remain unchanged.
Design human–AI collaboration
First define value and understand the real work. Then map the flow, identify waste, protect valuable thinking, and set roles, quality controls and a learning loop.
From value to Cognitive Kaizen
Eight steps for improving human–AI work
One clear process leads from defining value to a future way of working and a small experiment. The compass, waste categories, diagnostic lenses and countermeasures are used at the relevant stage.
PURPOSE AI should support value, understanding, decisions and learning
01Define valuePurpose, recipient, quality criteria, constraints and the cost of error.
02See cognitive workCognitive Gemba: how an analysis, decision or recommendation is really created.
03Map the flowHuman-AI Cognitive Value Stream Map – Current State.
04Recognise waste and riskLosses of attention, information, quality, accountability and learning.
05Protect valuable thinkingRules and filters supporting critique, reflection and sound decisions.
06Allocate human–AI workRoles, human control points and clear ownership.
07Design the standardFuture State, flow, information, tools and working conditions.
08Improve through KaizenA small experiment, measurement, review and the next iteration.
WORK OUTCOME less cognitive waste, more conscious human judgement
AI Co-thinking Compass
Intent, exploration, critique, simplicity and reflection help teams select the right mode of thinking for the task.
Seven cognitive wastes
A practical structure for observing losses in attention, information, processing and decision-making.
Twelve diagnostic lenses
They expose a problem from Lean, cognitive, decision, ethical and organisational perspectives.
Countermeasures for AI work
Diagnosis becomes rules for flow, verification, closing the learning loop and protecting human judgement.
Delivery model
How we develop human–AI collaboration
We do not begin with a tour of every available tool. We select one important knowledge or decision process and take the team from diagnosis to a testable experiment.
01
Free consultation and process selection
We discuss the purpose, recipients of the output, current use of AI and the risk of error. We choose a process with business importance that can be observed.
Outcome: scope and guiding question.
02
Current State diagnosis
During Cognitive Gemba, we map information, tasks, interactions with AI, decisions, loops and control points. The team sees the whole flow rather than a single prompt.
Outcome: current workflow map.
03
Waste, risk and human judgement
We identify cognitive waste, points where quality is lost and situations in which an AI recommendation may be mistaken for a decision. We define what should not be delegated.
Outcome: waste and risk register.
04
Future State and work standard
We design the target flow, human and AI roles, verification criteria, quality gates and a minimum team standard.
Outcome: work allocation and Future State.
05
Small experiment and review
We introduce one change within a controlled scope, test predictions, capture learning and observe side effects. The next step follows evidence rather than enthusiasm for a tool.
Outcome: experiment and next Kaizen loop.
A practical way to begin
Human-AI Cognitive Lean Thinking Lab
The two-day laboratory can be a stand-alone workshop or the start of a broader implementation. Participants work on one of their real processes, not merely on a demonstration exercise.
1
Diagnose value and the quality of cognitive work
- Human-AI Cognitive Lean Thinking and five levels of analysis
- The AI Co-thinking Compass as a framework for work quality
- Cognitive Value: purpose, recipient, criteria and cost of error
- Cognitive Gemba and a Current State Value Stream Map
- Seven cognitive and information wastes and twelve diagnostic lenses
2
Design the human–AI work system
- Countermeasures that prevent cognitive waste from returning
- Human-AI Work Allocation: conscious allocation of tasks and judgement
- A minimum standard for team use of AI
- Future State with explicit quality control points
- A small experiment and the next Cognitive Kaizen loop
We tailor the scope after a conversation. Your organisation may begin with the LAB, a diagnosis of one process, or combine the workshop with a pilot and a structured review.
What remains after the engagement
Concrete working outputs, not general inspiration
The engagement produces agreed working artefacts that enable the team to continue improvement after the workshop.
Current State map
A view of the present information flow, human work and interaction with AI.
Waste and risk register
Cognitive losses, quality risks and points that require human judgement.
Human–AI work allocation
Clear roles, the scope of AI assistance, accountability and verification.
Minimum team standard
Shared rules for intent, sources, controls, decisions and closing the learning loop.
Future State map
A target flow with decision points, quality controls and reduced overload.
Improvement experiment
One hypothesis, a test boundary, expected outcome and review date.
Adoption and governance
Guardrails that make AI use sustainable
Responsible adoption is not a policy stored in a folder. It is visible in the daily workflow: who may use AI, for what purpose, with which information, how an answer is checked and who owns the final decision.
Human accountability
Name the person who reviews the output, makes the decision and remains responsible for the consequence.
Information discipline
Define permitted data, reliable sources, context requirements and what must not be entered into a tool.
Quality gates
Set verification criteria, escalation rules and control points in proportion to the cost of error.
Learning cadence
Review errors, surprises and useful patterns so the standard improves with experience.
Adoption by design
Begin with one meaningful process, involve its users and expand only after the workflow has been tested.
Connection to compliance
Where required, the work can be connected to a separate AI governance and regulatory readiness stream.
Who it is for
Teams performing knowledge-intensive work
Boards and leaders responsible for transformation, digitalisation and AI
Process owners, Lean and Operational Excellence teams
Managers producing analyses, recommendations and decisions
HR and L&D teams developing AI collaboration capability
Project, service and expert teams working across multiple sources
Example processes
Where to begin the diagnosis
Management recommendationsResearch, analysis, scenarios, alignment and decision.
Reporting and analyticsCollecting data, interpretation, conclusions and communication.
Planning and projectsPriorities, risks, alternatives and delivery reviews.
Customer serviceUnderstanding the need, response, escalation and team knowledge.
HR and people developmentMaterials, assessment, feedback and learning design.
Expert workDocuments, requirements, sources, evaluation and accountability.
Knowledge partnership
LeanTrix × Trigger Mind
We combine LeanTrix experience in process improvement and organisational transformation with Dr Iwona Burka's original approach to cognitive work quality and human–AI collaboration.
Lead expert and contact
Dr Iwona Burka
Academic lecturer, researcher, business trainer and innovator specialising in Lean Management, the quality of thinking, problem-solving and conscious collaboration between people and AI.
For more than 20 years she has combined research, education and business practice. She develops the Trigger Mind Project, the AI Co-thinking Compass, AI-Ready Mind and the Human-AI Cognitive Lean Thinking approach. Her work protects what should remain human: judgement, accountability, understanding and learning.
At LeanTrix, she leads diagnoses, workshops and the design of human–AI collaboration standards.
Trigger Mind ProjectThinking quality, cognitive waste, AI-Ready Mind and responsible human–AI collaboration.
Evidence and trust
Explore LeanTrix experience beyond this service
Human-AI Cognitive Lean Thinking is a focused service within the broader LeanTrix improvement ecosystem. Explore verified client feedback and documented improvement engagements.
Client voiceClient references
Read feedback and see organisations that have worked with LeanTrix.
View references PracticeCase studiesSee documented examples of process and organisational improvement.
Explore cases PeopleOur teamMeet the practitioners who connect Lean, leadership, technology and change.
Meet the team TransformationLeanTrix Transformation LoopPlace an AI pilot within a coherent transformation and capability-building system.
See the modelA clear boundary: client references and case studies demonstrate LeanTrix delivery experience. They are not presented as quantified results of this specific Human-AI service unless a case explicitly states that scope.
A coherent Digitalisation and AI ecosystem
Connect this service with other LeanTrix AI capabilities
Human-AI Cognitive Lean Thinking complements technical deployment. It helps design how people should work before automation locks in a weak flow or unclear accountability.
KnowledgeWhat is Lean AI?
Connect artificial intelligence with Lean thinking and continuous improvement.
Read the guide Technical capabilityPrompt EngineeringBuild better instructions, context and quality controls for model outputs.
Explore the service Operating modelAI OrchestrationCoordinate people, tools, agents, data and decisions across the workflow.
See the approach ScaleAI in the OrganisationMove from a controlled pilot to everyday team use with clear rules and risk assessment.
Plan the journeyFrequently asked questions
Human–AI collaboration in practice
Key answers about the service, the LAB format and selecting a process.
Is this a prompt engineering course?
No. Prompts may be one element, but the purpose is to design the full workflow: value, information, work allocation, human judgement, verification, decisions and learning.
What is Human-AI Cognitive Lean Thinking?
It is an original approach developed by Dr Iwona Burka and Trigger Mind. It extends Lean and Kaizen logic to cognitive work performed jointly by people and AI, helping teams identify waste and design a better collaboration system.
What can become a pilot process?
A good candidate is a repeatable, business-critical knowledge or decision process: preparing a recommendation, reporting, planning, analysing offers, customer service, learning design or expert work.
Is the two-day LAB mandatory?
No. The LAB is a practical starting format, but the scope may begin with a diagnosis of one process or expand into a pilot and review. We select the best option during a free consultation.
What outputs does the organisation receive?
Typical outputs include a Current State map, a waste and risk register, human–AI work allocation, a minimum team standard, a Future State map and the design of one improvement experiment.
Does the service include AI governance?
It includes operational guardrails such as accountability, information rules, verification criteria and escalation points. A broader governance or regulatory readiness stream can be connected where required.
Does AI make the final decision?
That depends on the process and risk, but the service makes the decision owner explicit. AI output is treated as input to a defined workflow, not automatically as an approved decision.
How do you support adoption after the workshop?
We define a small experiment, expected outcome and review cadence. The team learns from actual use, improves its standard and scales only after the process has been tested.
A first step without obligation
Find where AI work is losing attention, quality or accountability
During a short consultation with Dr Iwona Burka, we will select a process worth examining and recommend a suitable format: diagnosis, the Human-AI Cognitive Lean Thinking Lab, or a pilot with a review.