Capacity
Can the required cognitive operation or domain skill be performed reliably under this level of demand?
Cognitive Systems Intelligence · Organisations and institutions
Find what is constraining the work, learning or service.
Cognitive Systems Intelligence examines how cognition, AI, policies, incentives, authority and operating conditions interact. The aim is not to assume that a performance or participation problem belongs to the person, but to identify the leading explanations, give affected people an accountable voice and test what should change.
Leaders in organisations, education, services, research, coaching and performance can discuss fit or a possible pilot with Dr Mark Ashton Smith at mark@iqmindware.com.
One defined activity system
Competing hypotheses
Participant voice
Continue, adapt, stop or reopen
Constraint localisation
Do not assume the problem is inside the person.
The same delay, error pattern, disengagement or strain can arise from very different causes. CSI keeps four intervention loci separate until the evidence justifies narrowing.
Coupling
Does a useful policy exist but fail to be cued, recovered, enacted, supported or updated when the activity requires it?
Niche
Are workload, learning design, incentives, authority, information, access or AI configuration creating or maintaining the problem?
Mixed
Are several constraints interacting—or is the current evidence not yet strong enough to separate them?
Cognitive Systems Intelligence
A bounded learning loop for important activity systems.
CSI moves from observation to a small test, then uses the result to update the institutional model rather than jumping from a problem directly to a diagnosis or person-level attribution.
Define + map
Choose one important workflow, pathway, learning process or programme, with an outcome, affected participants and protected boundaries.
Localise
Build competing Capacity, Coupling, Niche and Mixed hypotheses. Record operational evidence and the experience of people inside the system.
Test
Select a low-risk, useful and preferably reversible intervention that can improve the outcome while teaching us something about the system.
Re-measure + update
Review functional outcomes, cognitive evidence, participant experience, cognitive agency and sustainability separately, then continue, adapt, stop or reopen.
Illustrative CSI Surface
From “what is wrong?” to “what should we test?”
The institutional Surface is designed to show a pressure point, the leading explanations, the evidence behind them and the next bounded test—not to rank workers, learners, clients, athletes or other participants.
Activity system: AI-assisted research and teaching
Pressure point: source verification after interruptions
Observed: rework increasing; participants report loss of context
Leading hypotheses:
Human–AI coupling — High
Interruption structure — Moderate
Binding demand — Moderate
Next test: source-linked output + resumption cue + protected review window
Review: 14 days, with participant and system outcomes separate
Two-sided cognitive systems improvement
Improve the system while strengthening the people within it.
CSI should create value for the organisation or institution and for the people whose cognition, judgement and participation make the system work.
Localise constraints and learn what improves the mission.
Use operational, cognitive and stakeholder evidence to select bounded interventions and build reusable knowledge without overgeneralising from one case.
- More accurate capacity, coupling, niche and mixed hypotheses
- Safer workflow, learning, service and human–AI design
- Evidence to continue, adapt, stop or scale
- Sustainable outcomes rather than hidden overload
Understand, influence and develop the cognitive conditions of participation.
Gain a clearer explanation of difficulty, an accountable route for raising issues, a role in bounded redesign and private support for portable mastery.
- See what evidence supports an interpretation and what remains uncertain
- Correct evidence, contest a material inference and request human review
- Build worker- or learner-controlled evidence of mastery
- Retain meaningful judgement, provenance and review rights when AI is used
Personal cognitive capital remains person-owned. Appropriate aggregate evidence may help the institution learn, but it does not create an unrestricted personal cognitive profile for managers or other decision-makers.
Seven forms of cognitive agency
Empowerment means more than access to an app.
Epistemic agency
Understand whether the current difficulty is better explained by capacity, coupling, environmental structure or interaction.
Voice and procedural agency
Raise a pressure point, choose an appropriate visibility level, see who owns the response and request review or escalation.
Work- or learning-design agency
Help define protected constraints, useful outcomes and a bounded intervention that can be tested without silencing lived context.
Mastery and cognitive capital
Build portable capacities, strategies, domain models and evidence that survives changed conditions and delay.
Progression and mobility
Use transparent role- or learning-demand maps and participant-controlled evidence—not cognitive-app rankings—to support development.
Human–AI cognitive agency
Understand what is delegated, preserve provenance and model ownership, challenge AI-supported conclusions and retain viable unaided capability.
Collective agency
Turn recurring private experiences into privacy-preserving evidence of a system problem, with representative routes where appropriate.
Cognitive Systems Intelligence · sector-general
One systems-intelligence logic across very different institutions.
Sector changes the roles, governance and outcomes—not the core question: is the constraint in capacity, how a useful policy is supported, the surrounding system, or a mixture?
Commercial & professional organisations
Examples: companies, professional services, regulated industries and distributed professional networks.
CSI focus: workload, human–AI allocation, decision rights, verification, rework and judgement.
Education & research institutions
Examples: schools, universities, training providers, laboratories and research networks.
CSI focus: learning load, source/context tracking, transfer, protocol execution and research decisions.
Public, emergency & mission-driven services
Examples: government, civic services, emergency/high-reliability settings, charities and NGOs.
CSI focus: queues, hand-offs, situational awareness, escalation, resource constraints and service delivery.
Health, care, coaching & performance
Examples: care and rehabilitation services, counselling/coaching, sport and human-performance settings.
CSI focus: governed support, readiness, follow-through, decision timing, feedback and transfer.
Recognisable entry points
Six ways into the same CSI Core.
These are focused routes into one underlying model, not separate theories of organisational or institutional performance.
CSI Human–AI
What should AI remove, what should people retain, and where are verification, orchestration, authority or governance creating new cognitive burdens?
CSI Cognitive Demand
Where are interruptions, information change, task switching, learning load or service conditions placing attention, memory and judgement under pressure?
CSI Decision
Are stopping rules, approval chains, error costs or decision rights producing overchecking, premature closure or blocked action?
CSI Engagement
How are activity design, resources and expected outcomes shaping participation, avoidance, strain, overcommitment or recovery?
CSI Incentives
Is the institution selecting the cognitive and behavioural policies it says it wants—or selecting something else in practice?
CSI Workflow
Where are queues, hand-offs, dependencies, information architecture, course structure or rework interacting with human cognition?
What can change
Develop the person. Improve the coupling. Redesign the conditions.
A CSI analysis does not assume that cognitive training is the answer. The intervention should match the leading constraint and preserve participant rights.
Develop
Cognitive training, strategies, domain learning, mentoring, external supports or targeted skills development where the evidence supports a capacity or mastery bottleneck.
Couple
Resumption cues, decision gates, learning scaffolds, AI-verification protocols, prompts and feedback routines where a useful policy is not reliably entering the activity.
Redesign
Workload, course or service structure, incentives, responsibilities, decision authority, access, feedback, information or human–AI role allocation where the niche is the main constraint.
Human–AI Activity Design
AI can remove cognitive work—and create new cognitive work.
CSI distinguishes immediate AI efficiency from human cognitive augmentation and wider system adaptation. Faster output is not automatically a better operating, learning or service model.
- Map what AI now retrieves, generates, analyses, recommends or executes
- Identify new verification, source-binding and supervision demands
- Clarify what people must still understand, learn and own
- Specify provenance, human review, escalation and accountability rules
- Check whether AI changes autonomy, intensity, engagement or access to mastery
CSI Pilot
Start with one important activity system and one bounded question.
The preferred starting product is a founder-led investigation that maps the problem, localises the leading constraint, makes participant and decision rights explicit and tests a practical intervention before any wider scale decision.
Define
Agree the workflow, pathway, learning process or programme, the outcome, participants, duration, protected constraints and privacy boundary.
Map + observe
Use operational evidence, participant input and optional cognitive measures to build the current system model and competing explanations.
Test
Run the most informative bounded intervention supported by the evidence: Develop, Couple, Redesign or a prespecified combination.
Review
Re-measure functional, experience, agency and cognitive outcomes separately and decide whether to continue, adapt, stop or reopen.
Important: G Track and Cognitive Control Coach may be used when cognitive measurement or training is relevant, but they are not mandatory components of every CSI pilot and their scores do not become institutional selection or ranking scores.
Current founder-led entry offers
Begin with a bounded design or evidence engagement.
The current pricing remains an early founder-led structure while CSI is being developed through real organisational and institutional cases. Scope is confirmed before work begins so the intervention matches the problem rather than forcing every setting into the same programme.
Invite only · maximum three
Founding Design Partner
Best for: a small organisation or institution willing to help shape the CSI workflow, ontology, participant surface and implementation model around one real use case.
$750fixed pilot fee
- Scope
- One activity system
- Format
- Founder-led
- Pressure-point and participant definition
- Capacity / Coupling / Niche / Mixed hypothesis map
- One bounded intervention design
- Voice, visibility and implementation requirements
Standard CSI starting point
CSI Workflow Pilot
Best for: one team, workflow, learning process, service pathway or programme with a defined friction point such as rework, interruption, AI verification, decision delay, disengagement or avoidable cognitive effort.
$1,250fixed pilot fee
- Scope
- One defined system
- Duration
- Typically 6–8 weeks
- Activity-system + cognitive-demand map
- Constraint hypothesis and confidence summary
- One bounded intervention
- Aggregate outcomes + participant and close-out review
Human–AI focus
Human–AI Work Design Review
Best for: a team, course, service or research process where AI has changed task allocation, learning opportunities, verification burden, source ownership, supervision or decision responsibility.
$2,250fixed review fee
- Scope
- One AI-assisted activity
- Output
- Design recommendations
- Pre-/post-AI cognitive-demand comparison
- Delegation, provenance and verification map
- Human judgement, learning and governance boundaries
- Recommended bounded redesign test
Methods-focused
CSI Evidence Pilot
Best for: an organisation or institution preparing for a research, procurement or larger-scale decision that needs explicit methods, data provenance, participant rights and evidence boundaries.
$3,500fixed pilot fee
- Scope
- One defined use case
- Output
- Methods + evidence pack
- Constraint graph and intervention rationale
- De-identified data export where appropriate
- Protocol/model versions and confidence limits
- Methods, governance, limitations and implementation appendix
Founder-capacity limit: no more than two managed engagements run at one time. Final scope, participant requirements, governance routes and any use of IQ Mindware cognitive apps are agreed before launch.
Evidence, agency and privacy boundary
Support the people. Improve the conditions. Keep the evidence honest.
CSI is designed as institutional decision support and participant development—not workforce, learner or client ranking and not automated management, selection or progression.
Aggregate first
Ordinary decision-makers should normally see activity-system and cohort patterns rather than individual cognitive profiles.
Participant authority
People should know what is measured and inferred, correct evidence, contest material interpretations and request accountable human review.
Separate evidence types
Cognitive scores, functional outcomes, participant experience, agency and intervention effects should not be collapsed into one result.
Consequential decisions stay human
CSI may organise evidence and propose tests, but employment, educational, clinical, safeguarding, selection and eligibility decisions remain governed human responsibilities.
No diagnostic or suitability inference. CSI does not diagnose cognitive deficits, burnout or mental-health conditions, and app performance does not establish job, academic, clinical or athletic suitability.
Have one activity system worth investigating?
Start with the pressure point and the people affected by it. We can decide what evidence is needed and whether the first test should develop capability, improve coupling, change the conditions or combine these approaches.