What FLAIMS rests on, and what it does not claim.
FLAIMS is a synthesis of research traditions and field practice, not a claim that every specific FLAIMS mechanism is independently validated as FLAIMS. This page shows which sources support which parts of the model, and what they do not prove.
Supports, does not prove.
Each source has two lines: what it supports and what it does not prove. The point is to avoid overclaiming. FLAIMS combines these sources into an architecture that practitioners use, test and correct in the field.
We do not claim every individual FLAIMS building block has been validated as FLAIMS in a randomized trial. We do claim that the underlying ideas are well documented across decades and hold up in practice when applied seriously.
Flow
Personal attention, information movement and delivery throughput. FLAIMS treats them as one continuous design problem.
- Csikszentmihalyi, M., 1990Flow: The Psychology of Optimal ExperienceHarper & Row
SupportsThe cognitive state where skill, challenge and attention align, and people perform at their best.
Does not proveDescribes individual experience. Does not prescribe how to design flow at team or organization level.
- Goldratt, E., 1984The GoalNorth River Press
SupportsThroughput of a system is constrained by its slowest step. Local optimization elsewhere is theatre.
Does not proveOriginally framed for manufacturing. Application to knowledge work requires translation.
- Newport, C., 2016Deep WorkGrand Central Publishing
SupportsFocused, distraction-free work produces disproportionate value in knowledge-heavy roles.
Does not proveTrade-book synthesis rather than a controlled study. Useful as a heuristic, not a measurement.
- Reinertsen, D., 2009The Principles of Product Development FlowCeleritas Publishing
SupportsQueue size, batch size and WIP limits drive lead time more than individual productivity.
Does not proveQuantitative models assume measurable flow. Many service settings need adapted proxies.
Leadership
Leadership separated from authority, kept visible and coachable. The pillar leans on motivation research and on humility-and-will leadership traditions.
- Deci, E. & Ryan, R., 1985Intrinsic Motivation and Self-Determination in Human BehaviorPlenum
SupportsAutonomy, competence and relatedness as the three durable drivers of motivation.
Does not proveDoes not say which organizational structure delivers them. FLAIMS proposes one of several.
- Collins, J., 2001Good to GreatHarperBusiness
SupportsSustained great companies are led with personal humility and intense professional will (Level 5).
Does not proveSelection-bias critique is well documented. Treat as orientation, not as a causal law.
- Edmondson, A., 1999Psychological Safety and Learning Behavior in Work TeamsAdministrative Science Quarterly
SupportsTeams learn faster when people can speak up about mistakes and half-formed ideas without losing status.
Does not provePsychological safety is necessary, not sufficient. It must be paired with accountability for results.
- Seligman, M., 2011Flourish (PERMA model)Free Press
SupportsWell-being at work has five components: positive emotion, engagement, relationships, meaning, accomplishment.
Does not provePERMA-Lead operationalization is one of several. FLAIMS adopts it pragmatically.
Accountability
Named ownership of outcomes combined with system-first responses to failure.
- Reason, J., 1990Human Error / Swiss-Cheese ModelCambridge University Press
SupportsSerious failures almost never have a single cause. They emerge when gaps in multiple layers align.
Does not proveOriginating in aviation and healthcare. Applying it culturally to commercial firms is still craftwork.
- Dekker, S., 2014The Field Guide to Understanding 'Human Error'Ashgate
SupportsNaming a person responsible is compatible with treating the system as the first thing to fix.
Does not proveCultural translation into commercial firms requires deliberate ritual, not just policy.
- Patterson, K. et al., 2013Crucial AccountabilityMcGraw-Hill
SupportsHow to hold high-stakes conversations about broken commitments without breaking the relationship.
Does not proveA skills toolkit. It assumes the surrounding structure already names owners and outcomes.
Intelligence and Human-Centered AI
AI as governed organizational infrastructure, always under human direction.
- Shneiderman, B., 2022Human-Centered AIOxford University Press
SupportsHigh automation and high human control are not a trade-off. Both can grow together with the right design.
Does not proveProvides design principles, not a turnkey operating model. Adoption decisions stay with the firm.
- Parasuraman, R. & Manzey, D., 2010Complacency and Bias in Human Use of AutomationHuman Factors
SupportsAutomation bias is a documented, repeatable tendency to over-trust automated outputs, especially under time pressure.
Does not proveCannot be eliminated by awareness alone. Structural verification rituals are required.
- Brynjolfsson, E. & McAfee, A., 2014The Second Machine AgeW. W. Norton
SupportsThe largest productivity gains come from complementarity: humans and machines doing what each does best in a shared workflow.
Does not proveMacro-level argument. Concrete role design at firm level is left open.
Mastery
A learning organization made of learning individuals. The pillar combines deliberate practice with strengths research.
- Ericsson, K. A., 1993The Role of Deliberate Practice in the Acquisition of Expert PerformancePsychological Review
SupportsExpertise comes from focused practice with fast feedback at the edge of current ability.
Does not proveTime-on-task is not enough. Quality of feedback and coaching matters as much as hours.
- Senge, P., 1990The Fifth DisciplineDoubleday
SupportsRecurring problems usually have structural causes, not personal ones. Learning organizations design for them.
Does not proveSynthesis-heavy and high-level. Translation into daily rituals is left to the practitioner.
- Buckingham, M. & Clifton, D., 2001Now, Discover Your StrengthsFree Press
SupportsSustained performance comes from leveraging strengths, not from grinding away at weaknesses.
Does not proveThe strengths instruments are proprietary. The principle is broader than any one tool.
Segmentation of Power
Decisions routed by weight through the Gravity Decision Model, with clearly separated forums.
- Conway, M., 1968How Do Committees Invent?Datamation
SupportsThe systems an organization builds mirror its communication structure. Decision forums shape outcomes.
Does not proveOriginally an observation about software. Generalization to organizations is mostly inductive.
- Dunbar, R., 1992Neocortex Size as a Constraint on Group Size in PrimatesJournal of Human Evolution
SupportsNatural group-size thresholds (≈5, 15, 50, 150) at which trust and coordination shift in kind.
Does not proveThresholds are approximations, not bright lines. They orient design, they do not dictate it.
- Laloux, F., 2014Reinventing OrganizationsNelson Parker
SupportsDocuments real-world cases of distributed decision-making and the price of the missing structures.
Does not proveCase-study evidence. FLAIMS keeps leadership visible, where many Teal cases dissolve it.
Cognitive Biases
Cognitive biases are how human cognition works. FLAIMS treats them as structural risks, not personal failings.
- Kahneman, D., 2011Thinking, Fast and SlowFarrar, Straus and Giroux
SupportsTwo systems of thought, fast and slow, and the predictable biases that emerge when fast takes over.
Does not proveSeveral specific findings have failed to replicate. Use the framework, not the individual effect sizes.
- Nickerson, R., 1998Confirmation Bias: A Ubiquitous Phenomenon in Many GuisesReview of General Psychology
SupportsPeople preferentially seek information that confirms what they already believe. Personalizing AI amplifies the effect.
Does not proveAwareness training reduces self-reported bias but rarely changes decisions. Structural friction is the lever.
- Samuelson, W. & Zeckhauser, R., 1988Status Quo Bias in Decision MakingJournal of Risk and Uncertainty
SupportsDefaults win. People stay with current tools and processes even when better ones exist.
Does not proveEffect sizes vary by stakes and time horizon. Exit rituals and review rhythms are the practical counter.
- Tversky, A. & Kahneman, D., 1974Judgment under Uncertainty: Heuristics and BiasesScience
SupportsAnchoring, availability and representativeness as the foundational heuristics that distort decisions.
Does not proveA foundational paper. Specific anchoring magnitudes vary across populations and tasks.
From evidence to architecture.
The sources above are the building blocks. How FLAIMS assembles them into an operating model is in the FLAIMS framework and in the field guide to biases.