The intellectual foundations of reflective intelligence lie in a question that has accompanied the study of human thought for centuries: what distinguishes the possession of knowledge from the ability to examine, question and improve the reasoning through which that knowledge is applied? An individual may possess extensive information, considerable technical expertise and a long record of professional experience without necessarily being able to recognise the limitations of established assumptions. Intelligence becomes more valuable when it includes the capacity to examine its own conclusions, reconsider its methods and determine whether the evidence justifies the decisions being made.
The word reflective derives from the Latin reflectere, meaning to bend or turn back. Its contemporary intellectual significance concerns the deliberate reconsideration of an idea, action or experience. The word intelligence, derived from Latin intelligentia, has long been associated with understanding, comprehension and discernment. Their combination expresses a particular conception of intellectual capability: the ability not only to reason about a subject, but also to examine the quality, assumptions and consequences of that reasoning.
The expression reflective intelligence does not have a single, universally acknowledged inventor or a definitive historical point of origin. Its conceptual foundations developed through several related traditions, including philosophy, reflective thought, cognitive psychology, professional practice, metacognition, cybernetics and organisational learning. These traditions differ in their methods and purposes, but share an interest in the ways that knowledge can be evaluated, revised and applied more intelligently.
A significant foundation was established by the American philosopher and educational theorist John Dewey in How We Think, published in 1910. Dewey distinguished reflective thought from the mere succession of ideas that passes through the mind. For him, reflection involved the deliberate examination of the grounds supporting a belief and the consequences to which that belief might lead. It required the suspension of premature judgement, the investigation of relevant evidence and the testing of possible explanations.¹
The significance of Dewey's contribution extends beyond education. His account established reflection as a disciplined intellectual activity rather than passive contemplation. A conclusion should not be accepted simply because it appears plausible or accords with established belief. Its foundations should be examined, competing explanations considered and further evidence sought where necessary. Reflection is therefore inseparable from the willingness to discover that an initial conclusion may be wrong.
This tradition became particularly relevant to professional practice through the work of Donald A. Schön. In The Reflective Practitioner: How Professionals Think in Action, published in 1983, Schön examined how professionals respond to complex situations that cannot be managed through the mechanical application of established rules alone. He distinguished reflection undertaken during action from reflection undertaken after an event, showing how professional understanding can develop through the continuing interpretation of experience. His analysis was especially concerned with the relationship between technical knowledge and the judgement required when practical circumstances exceed the reach of established methods.
Cognitive psychology developed a related perspective through the study of metacognition. In his 1979 paper, Metacognition and Cognitive Monitoring: A New Area of Cognitive-Developmental Inquiry, John H. Flavell examined how individuals acquire knowledge about their own cognitive processes and monitor the effectiveness of their thinking. The broader significance of metacognition lies in the ability to recognise not merely what one believes or knows, but also the limitations, uncertainties and conditions affecting the reliability of that knowledge.
The expression reflective intelligence itself has also received explicit treatment in psychological and educational literature. In Outsmarting IQ: The Emerging Science of Learnable Intelligence, published in 1995, David Perkins distinguished reflective intelligence from more conventional accounts of intellectual capacity and specialised knowledge. He associated it with the ability to recognise habitual patterns of thought and move beyond their limitations. This is an important documented use of the phrase, although it should not be taken as proof of its first appearance or as the origin of every subsequent framework employing it.
These intellectual traditions establish the background from which reflective intelligence can be understood as a broader concept. It concerns the relationship between information, experience, reasoning, evaluation and the willingness to revise a conclusion. Its purpose is not to encourage perpetual indecision, but to improve the grounds upon which decisions are made.
For GENERAL INTELLIGENCE PLC, this conception has particular relevance to artificial intelligence consultancy. The growing capability of machines makes it increasingly important to distinguish technical performance from genuine commercial contribution, confidence from reliability and information processing from sound judgement. REFLECTIVE INTELLIGENCE® applies the discipline of reflection to those decisions, providing a framework for determining where artificial intelligence can create enduring value and where its adoption may be unnecessary, premature or commercially unjustified.
From Reflective Thought to Reflective Artificial Intelligence
The development of artificial intelligence has introduced new possibilities for analysing information, identifying patterns, generating content, supporting reasoning and automating complex activities. These advances have also intensified an older intellectual problem: the distinction between producing an answer and establishing whether that answer should be trusted.
An artificial intelligence system may generate an apparently coherent conclusion while relying upon incomplete information, unsuitable assumptions or an incorrect interpretation of the question. A model may perform well in one environment but less reliably when circumstances change. A system may reduce the time required to complete a task without improving its underlying quality or the commercial value of its result.
Reflective intelligence offers a framework for examining these distinctions. Rather than evaluating artificial intelligence solely by its capacity to perform a task, it asks whether the proposed use improves the quality of judgement, whether the reasoning can be examined, whether uncertainty is adequately represented and whether the consequences of the decision remain understood.
This is particularly important in a professional environment where responsibility cannot be transferred to a machine merely because the machine supplied the analysis. The institution must retain the capacity to question the output, recognise its limitations and determine what action is appropriate.
The intellectual ambition of REFLECTIVE INTELLIGENCE® is consequently broader than technological assessment. It is concerned with the relationship between what an artificial intelligence system can do, what an organisation requires, what the evidence establishes and what should follow. The objective is not to constrain technological progress, but to ensure that increasingly capable technology serves a clearly defined institutional purpose.
The General Reinsurance Broker as the Subject of Reflection
General reinsurance broking is an information-intensive and judgement-intensive business. Its value lies in the ability to understand clients and their portfolios, interpret reinsurance markets, structure the transfer of risk, negotiate with reinsurers and apply specialist knowledge to circumstances in which the information is incomplete and the consequences may be substantial.
The broker's work involves more than the exchange of documents between insurers and reinsurers. It requires the interpretation of risk, the presentation of information, an understanding of market capacity and appetite and judgement concerning the structure and terms through which risk can be transferred. In the Lloyd's market, for example, brokers facilitate the transfer of risk between policyholders and underwriters, supported by specialist knowledge and market relationships. The general reinsurance broker performs a distinct role within the wider reinsurance market, advising and arranging placements according to the relevant client relationships and contractual circumstances.
Artificial intelligence may affect the economics and organisation of this work in several ways. It can assist with the analysis of exposure information, the preparation and examination of submissions, the comparison of treaty terms, market research, document processing, knowledge management and the organisation of information used in professional decisions.
Yet the existence of a possible application does not establish that it should be adopted. A system may be technically capable without being sufficiently reliable for a particular purpose. An apparent efficiency may be offset by implementation costs, ongoing supervision, integration difficulties or the need to verify its output. A new capability may also change the organisation's risk profile or create a dependency upon a supplier whose future products, prices or operating arrangements remain outside the broker's control.
REFLECTIVE INTELLIGENCE® begins with the brokerage's actual business objectives and asks whether artificial intelligence provides an appropriate means of achieving them. It considers the commercial proposition, the quality of the evidence, the suitability of the technology and the consequences for the institution over time.
The purpose is not to make the brokerage adopt more technology. It is to enable the brokerage to make better decisions about technology.
Research and Intelligence
REFLECTIVE INTELLIGENCE® is informed by the research into artificial intelligence undertaken by X from GENERAL INTELLIGENCE PLC. That research considers the development of machine intelligence from scientific, technical, organisational, economic and institutional perspectives, including computational architectures, reasoning systems, intelligent agents, learning methods, evaluation, safety, governance and regulation.
The purpose is to distinguish material developments from technological fashion and established capability from expectation. A new model or architecture may be scientifically significant without being commercially suitable for a reinsurance broker. A relatively modest development may, conversely, have substantial value when integrated into an appropriate information environment or professional process.
Research must therefore extend beyond the identification of available products. It should examine the conditions under which a capability works, the evidence supporting claims about its performance, the limitations that may affect its use and the implications of incorporating it into an existing business.
For a general reinsurance broker, this analysis must reflect the particular characteristics of the market: specialist knowledge, incomplete information, contractual and technical interpretation, uncertainty, professional judgement and long-established relationships. The same technology may have different implications for different brokers, depending upon their business models, systems, expertise, scale and strategic priorities.
The objective is an informed view of the technology in context. Where evidence is incomplete, uncertain or contested, that uncertainty should be identified rather than concealed. Reflective intelligence depends upon the willingness to distinguish what is known, what is inferred and what remains to be established.
Strategy and the Purpose of Adoption
The first strategic question is not which artificial intelligence product the brokerage should acquire. It is what the brokerage is seeking to achieve and whether artificial intelligence is the appropriate means.
The desired outcome may involve a stronger client proposition, improved analytical capability, more effective use of specialist knowledge, greater operational capacity, better decision support or a different approach to the organisation's business model. Each objective implies different technical requirements, costs, risks and measures of success.
REFLECTIVE INTELLIGENCE® helps brokers examine where artificial intelligence could create commercial advantage, identify and prioritise potential applications, evaluate organisational readiness and consider credible alternatives. It also examines where established systems, process redesign, additional expertise or a simpler technological solution may produce a better result.
The distinction between automation and contribution is central. A system that processes submissions more quickly may release professional time, but its value depends upon how that time is used and whether the overall service improves. A model that generates analysis may increase output without improving the quality of the decisions taken from it. A platform may appear strategically important while creating costs and dependencies disproportionate to its benefits.
The broker must therefore examine the connection between the proposed technology and the underlying economics of the enterprise. Does the application strengthen the business? Does it improve the quality of advice, the management of risk or the use of scarce professional expertise? Does it create an advantage that is likely to endure, or does its value depend upon assumptions that may change quickly?
A sound strategy considers not only where artificial intelligence should be used, but where it should not.
Independent Assessment
A central purpose of REFLECTIVE INTELLIGENCE® is to provide independent assessment of proposed artificial intelligence initiatives and the assumptions upon which they depend.
The assessment begins with the problem being addressed. Is the problem properly defined? Is it material to the business? What evidence establishes the current cost, limitation or weakness? What outcome is sought and how will the brokerage determine whether that outcome has been achieved?
The proposed solution can then be assessed against its intended use. Depending upon the circumstances, this may involve consideration of model capability, accuracy, reliability, robustness, information quality, data provenance, security, confidentiality, privacy, integration, intellectual property, supplier dependency, regulatory considerations and the total cost of ownership.
For a reinsurance broker, special attention may be necessary where an artificial intelligence system influences professional judgement, contractual interpretation, exposure assessment or the presentation of risk to reinsurers. The system's performance must be suitable for the intended task and the professionals using it must understand the limitations that could affect their conclusions.
The presence of an AI capability does not remove the need to verify important information or establish the basis of advice. Equally, requiring a human to approve an output does not guarantee meaningful oversight if the person cannot examine its evidence or challenge its reasoning.
The objective of independent assessment is to establish whether the proposed application can deliver the outcome sought and whether the investment, complexity and organisational commitment are justified. The appropriate conclusion may be to proceed, to proceed differently, to undertake further investigation, to defer or not to proceed.
Commercial Value and Long-Term Owner Value
Artificial intelligence can involve material expenditure, management attention, technical resources and organisational change. Its commercial value must therefore be assessed in relation to the contribution it is reasonably expected to make to the brokerage.
That contribution may arise through improvements in the quality or speed of analysis, more effective use of specialist knowledge, enhanced client service, greater operational capacity or the development of new capabilities. These potential benefits should be examined alongside acquisition costs, integration, training, monitoring, security, governance, maintenance and the opportunity cost of committing capital and management attention.
REFLECTIVE INTELLIGENCE® places long-term owner value at the centre of this assessment. A project may produce visible short-term efficiencies while creating a costly dependency, reducing professional control or becoming obsolete more quickly than expected. Conversely, a measured investment in information quality, system architecture or specialist capability may create value that compounds through subsequent applications.
The rate of technological change makes this evaluation particularly important. Decisions should consider not only the capabilities available today, but the likely durability of the assumptions upon which the investment rests. A significant commitment may become less attractive if the underlying technology is likely to change materially; a modest investment may prove strategically important if it establishes a sound foundation for future development.
The relevant question is not whether the proposed capability is sophisticated, fashionable or technically impressive. It is whether the investment can make a worthwhile and enduring contribution to the brokerage, with risks and costs proportionate to the value expected.
Technology and Application
REFLECTIVE INTELLIGENCE® enables general reinsurance brokers to evaluate artificial intelligence technologies against their actual operational, strategic and commercial requirements.
This may involve examining models, platforms and systems intended to support research, market intelligence, exposure analysis, placement preparation, contract and wording analysis, document processing, workflow, knowledge management or decision support. It may also involve considering how a capability fits within existing systems, what information it requires, how its performance can be evaluated and what dependencies would follow from adoption.
The assessment should reflect the intended purpose. A tool that summarises documents has different requirements from a system used to analyse complex exposures or influence a consequential professional decision. Accuracy, transparency, security and the ability to verify results should be assessed in relation to the task, the consequences of error and the methods available for human review.
The brokerage must also consider whether the task itself should be performed in its existing form. A process may contain unnecessary steps, duplicate information or depend upon responsibilities that are poorly defined. Automating it may preserve those weaknesses rather than resolve them. A better outcome may arise from redesigning the process before selecting technology.
The appropriate solution is therefore not necessarily the most advanced one. A conventional system, a better process, an adjustment to organisational responsibilities or no technological intervention may offer superior value.
REFLECTIVE INTELLIGENCE® treats this conclusion as a legitimate outcome of the consultancy. Its purpose is to establish which means are justified by the objective, not to create an argument for adoption regardless of the evidence.
Governance and Professional Responsibility
For general reinsurance brokers, artificial intelligence may create questions of confidentiality, security, accuracy, accountability, intellectual property and professional responsibility. These questions become especially important where the technology is used with client information, contractual documents, placement materials or information subject to commercial restrictions.
The Financial Conduct Authority's published approach to artificial intelligence, updated in February 2026, emphasises the application of existing regulatory frameworks to AI in financial services, including accountability, governance and consumer protection. It has stated that its approach is principles-based and focused on outcomes rather than the introduction of a separate set of AI-specific rules.⁶ The precise regulatory requirements depend upon the firm's activities, the intended use of the technology and the relevant circumstances.
REFLECTIVE INTELLIGENCE® considers the governance arrangements necessary for the responsible use of artificial intelligence. Depending upon the application, this may include defined decision rights, risk assessment, testing, monitoring, documentation, access controls, incident management and criteria for reviewing, modifying or discontinuing a system.
The governance process must also address the relationship between machine output and professional judgement. The broker should know which activities are supported by AI, what limitations affect those activities and who remains responsible for the resulting decision, advice or outcome. Material information should be capable of verification and the use of external systems must not compromise confidentiality or contractual obligations.
Governance should be proportionate to the significance of the application. It should provide effective control while allowing legitimate commercial development. A sound framework is not an obstacle to innovation; it is the means through which an institution can adopt new capabilities while maintaining responsibility for their consequences.
Independent Challenge
Reflective intelligence requires the willingness to question a preferred conclusion. This is particularly important when a proposed technology has attracted substantial investment, persuasive supplier claims or strong internal enthusiasm.
REFLECTIVE INTELLIGENCE® examines whether the business problem has been defined clearly, whether the expected benefits are supported by evidence, whether relevant alternatives have been considered and whether the proposed solution is more complex than the objective requires. It also considers the assumptions that would have to remain true for the investment to succeed and the circumstances in which those assumptions might fail.
Independent challenge is not opposition for its own sake. A proposal should not be rejected merely because it introduces a new technology, just as it should not be accepted because the technology is innovative. The purpose is to establish the strongest available basis for the decision.
For a general reinsurance broker, this may require examining whether a proposed AI initiative genuinely strengthens client service, improves the economics of the brokerage, enhances professional capability or creates a defensible advantage. It may also require asking whether a more modest investment, a different operating model or an established solution would achieve the same objective with lower cost and less risk.
Independence requires the willingness to reach any of these conclusions. The appropriate outcome may be to proceed, modify the proposition, investigate further, defer or not proceed.
The Five Principles
REFLECTIVE INTELLIGENCE® is founded upon five principles: long-term owner value, contribution, judgement, simplicity and ordinary decency.
Long-term owner value is the governing commercial objective. Artificial intelligence should strengthen the enterprise over time rather than merely produce an attractive immediate proposition. This requires a realistic assessment of benefits, costs, dependencies and the durability of the assumptions underlying an investment.
Contribution concerns the proper relationship between technological capability and business purpose. A system should make a substantive contribution to the brokerage's performance, resilience, client proposition or long-term prospects. Activity, novelty and processing volume are not substitutes for value.
Judgement concerns the distinction between information and understanding. Artificial intelligence may assist analysis, but professional responsibility remains essential where conclusions depend upon context, uncertainty, commercial priorities or contractual interpretation. The brokerage must retain the capacity to challenge evidence and decide what should follow from it.
Simplicity is a commercial discipline. The simplest arrangement capable of achieving the legitimate objective with acceptable reliability and risk is ordinarily preferable. Additional technical complexity should be justified by a corresponding contribution rather than accepted as an inherent sign of sophistication.
Ordinary decency establishes a boundary to commercial calculation. Honesty, proper conduct, confidentiality, responsibility and fair dealing are fundamental to durable institutions. Artificial intelligence should be applied in ways that respect the obligations owed to clients, markets, employees and other parties affected by its use.
Together, these principles provide a basis for making decisions that are commercially rigorous without becoming technologically acquisitive and intellectually ambitious without losing sight of institutional responsibility.
Proportionate Engagement
The scope of consultancy should reflect the significance and complexity of the decision under consideration. Not every question requires a substantial research programme and not every proposed system warrants the same level of technical investigation.
REFLECTIVE INTELLIGENCE® may be applied through a focused strategic assignment, an independent assessment, an investment evaluation, a technology review, an application assessment, a governance review or continuing strategic advice.
The approach favours the smallest engagement capable of answering the question properly. The scope, objectives, responsibilities, outputs, timetable and commercial terms should be established in proportion to the matter under consideration. The consultancy should provide sufficient depth to support a sound decision without imposing unnecessary complexity upon the client.
The initial requirement is to understand the brokerage, the question being considered, the decision that needs to be made and any material commercial, technical, regulatory or institutional constraints. Only then can the appropriate scope of research and assessment be established.
Proportionate engagement is consistent with the wider philosophy of the framework. The object is to provide the intelligence required for the decision, not to create a larger exercise than the circumstances justify.
The REFLECTIVE INTELLIGENCE® Test
The framework ultimately asks a set of fundamental questions.
- What creates long-term value?
- What contribution is genuinely expected from the proposed capability?
- What requires professional judgement?
- Is artificial intelligence the appropriate means of achieving the objective?
- Can the same result be obtained more simply?
- What are the enduring costs, dependencies and risks?
- Does the proposal rest upon evidence or expectation?
- Can its assumptions be tested?
- Are its limitations understood?
- Can the resulting capability be governed and controlled?
- Will professional responsibility remain clear?
- Would the decision remain sound if technological fashion were removed from consideration?
These questions are intended to expose the reasoning underlying a decision before the commitment is made. They also establish a basis for review after implementation. The expected benefits should be compared with actual outcomes and material differences should inform subsequent decisions.
The test is not designed to prevent change. It is designed to ensure that change is justified. A decision to decline an unsuitable investment can represent a successful application of reflective intelligence, just as a decision to adopt a technology can be justified when the evidence supports it.
REFLECTIVE INTELLIGENCE® is therefore not an exercise in finding reasons to deploy artificial intelligence. It is a framework for determining where artificial intelligence can create genuine and durable advantage within a general reinsurance brokerage.
General Reinsurance Broking
The application of reflective intelligence to general reinsurance broking is particularly important because the profession combines specialist knowledge, market relationships, contractual understanding and decisions made under conditions of uncertainty. Artificial intelligence may change the economics of these activities, but the appropriate response requires judgement about the business as a whole.
A brokerage considering an AI investment must evaluate its intended contribution to research, market intelligence, analysis, placement, knowledge management, client service or another defined activity. It must also consider whether the proposed capability will remain suitable as technology develops, what dependencies it may create and how its performance will be evaluated.
These questions cannot be answered through a technical assessment alone. They require an understanding of the brokerage's strategy, commercial position, operating model, professional responsibilities and long-term interests. A capability that is attractive in isolation may be unsuitable for the institution. A modest application may be valuable if it addresses a specific weakness or improves the use of scarce expertise. A decision not to proceed may protect capital and management attention for more important objectives.
REFLECTIVE INTELLIGENCE® combines independent research into artificial intelligence with strategic, technical, commercial and institutional analysis. It seeks to establish what matters, what is justified by the evidence and what, if anything, should follow. The emphasis is on the relationship between technology, judgement, commercial value and institutional responsibility.
The conclusion may be to adopt a technology, change the proposed approach, undertake further investigation, defer the decision or not proceed. The purpose of the consultancy is not to promote artificial intelligence, but to help general reinsurance brokers make better decisions about when and how it should be used.
The Objective
The intellectual history of reflective intelligence is not the history of a single invention. It is the convergence of several traditions concerned with the quality of reasoning: Dewey's reflective inquiry, Schön's account of professional judgement, the study of metacognition and subsequent work on the capacity to recognise and revise habitual patterns of thought. These traditions share an important principle: intelligence is strengthened when conclusions remain open to examination and experience is used to improve future judgement.
REFLECTIVE INTELLIGENCE® is an artificial intelligence consultancy framework provided by X from GENERAL INTELLIGENCE PLC to general reinsurance brokers. It applies that principle to the strategic, technical, commercial and institutional decisions associated with increasingly capable artificial intelligence.
Its purpose is not to encourage technological adoption for its own sake or to treat professional judgement as an obstacle to automation. It is to establish where artificial intelligence can make a genuine contribution, whether the proposed investment is commercially justified, how the relevant technology should be assessed and what governance is required to use it responsibly.
The objective is a brokerage that understands both the potential and the limitations of artificial intelligence; that can distinguish capability from value, information from understanding and technological novelty from enduring advantage; and that remains willing to challenge its own assumptions before committing capital, resources and institutional attention.
The most intelligent decision will not always be to adopt a new technology. It may be to modify a proposal, choose a simpler means, defer investment until the evidence improves or decline to proceed. The measure of success is not the quantity of artificial intelligence deployed, but the quality of the judgement through which it is selected, governed and applied.
REFLECTIVE INTELLIGENCE® therefore represents a disciplined approach to the relationship between human judgement and machine capability: research to establish what is changing; reflection to determine what it means; independent assessment to test what is proposed; and commercial judgement to decide what should follow.
The ambition is not merely to make the brokerage more technologically capable. It is to make the institution more thoughtful about capability itself, more rigorous in the allocation of resources and more certain that the technologies it adopts serve a clearly defined purpose.