The Brain
Brings together knowledge, memory, cognitive playbooks and decision methods. We are extending it with causal models, procedural learning and mechanisms for evaluating and revising reasoning.
Science of Cognition & AI Engineering
Emergent Labs
Advancing machine reasoning, judgment and taste.
Emergent Labs brings together research, engineering and product development to advance the quality, independence and sophistication of machine cognition.
Our ambition is to build autonomous thinking systems whose reasoning, judgment and taste meet the highest standards of human thinking and extend the boundaries of what intelligence can achieve.
We are developing the cognitive architectures, learning systems and decision methods through which machines can direct their own investigations, improve how they think and produce original work.
FOUNDATIONAL RESEARCH & DEVELOPMENT
The cognitive capabilities we are building.
Our research combines cognitive science, decision science, information systems and AI engineering. Its thrust is to develop systems that can determine what to learn, acquire the evidence they need and improve the methods by which they reason.
Explore each programme
01Autonomous Learning & Curriculum Generation
We are developing systems that identify gaps in their own competence, generate learning tasks and select the next challenge or experiment. This work gives systems a role in designing their own intellectual development, including when to seek specialist knowledge or human judgment.
02Decision-Centred Retrieval & Active Information Acquisition
We are building hybrid information systems that work backwards from a decision to the evidence that could change it. They combine lexical and semantic retrieval, indexing and vector methods with structured data, source provenance and temporal knowledge.
Our research develops policies for choosing what to investigate next, guided by the expected value of information and the cost of acquiring it.
03Metacognitive Control & Verified Self-Correction
We are developing mechanisms through which systems monitor the quality of their own reasoning, recognise uncertainty and initiate further investigation or correction. These mechanisms combine competing hypotheses, independent evidence checks, numerical verification and deliberate challenge.
The objective is a system that can detect when an explanation has failed and rebuild the reasoning behind its decision.
04Continual Learning & Procedural Memory
We are building learning mechanisms that turn verified experience into reusable cognitive capability. Decision histories, expert feedback and observed outcomes inform changes to memory, cognitive playbooks, tool use and reasoning procedures.
Our research includes improving learning procedures themselves, alongside targeted model adaptation. Changes must demonstrate gains on unfamiliar problems while preserving existing capabilities.
05Causal Modelling & Counterfactual Decision Systems
We are developing systems that construct and test explanations of why outcomes occur, model the consequences of alternative actions and evaluate choices under uncertainty.
This work connects causal inference, quantitative models and decision theory. It includes learning which mechanisms can transfer across settings and which depend on local conditions.
06Computational Creativity & Evaluative Judgment
We are developing methods for generating original problem frames, concepts and strategic possibilities, together with the judgment required to select and develop them.
Our research connects generative search with expert evaluation of originality, explanatory power, coherence and value. Taste is part of this work: recognising what deserves to be pursued, refined or rejected.
EXCEPTIONAL HUMAN CAPABILITY, MADE SCALABLE
Turn a lifetime of expertise into a cognitive product.
Through The Collective, we work with Master Craftsmen to make their distinctive knowledge, methods and judgment available through AI-native products that can operate, evolve and scale far beyond them.
We are developing the platform for this work, combining expert knowledge elicitation, cognitive playbook engineering and the learning systems developed in Labs.
A specialist playbook can become a product in its own right, powering AI-native services, enterprise systems and specialist applications. We are building products that retain the expert’s distinctive craft and can test, refine and extend its application.
COGNITIVE ARCHITECTURE & ENGINEERING
Engineering the research into working intelligence.
Our research is being developed into three interconnected capabilities.
The Agent Studio
Assembles and coordinates specialist agents. We are developing investigation, generation and verification agents that select tools, challenge one another’s conclusions and revise their work as evidence changes.
The Autonomous Engineering Platform
Provides the computational infrastructure for these capabilities. It connects hybrid retrieval, model adaptation and routing, analytical tools, evaluation systems, enterprise interfaces and deployment architecture.
THE COGNITIVE PRODUCT FOUNDRY
Embedding cognitive advances into the products we build.
Our product development brings these research programmes into demanding professional applications. Marketing, consumer intelligence, business strategy and growth provide our first development environments. The underlying cognitive ambition extends beyond them.
Agency Brain
We are embedding decision-centred retrieval, verified self-correction and expert cognitive playbooks into connected strategy and planning workflows.
Development focus
Development focuses on a system that can interrogate a brief, determine what to investigate, challenge competing growth hypotheses and construct a strategy whose evidence and assumptions remain available for review.
Client Brain
We are developing enterprise cognitive systems around organisational knowledge, proprietary methods and consequential decisions.
Development focus
Active information acquisition directs investigation towards decision-critical gaps. Procedural memory and continual learning are being developed to carry verified lessons into subsequent work, within each enterprise’s knowledge and intellectual property boundaries.
Consumer & Category Growth Systems
We are bringing together hybrid evidence systems, causal modelling and counterfactual analysis to build a consumer-goods strategy decision engine.
Development focus
Development connects category dynamics, consumer behaviour and competitive evidence to choices about growth, portfolio, innovation, pricing, distribution and investment. The learning programme examines which growth mechanisms transfer across markets and where a different explanation is required.
Specialist Cognitive Products
We are turning exceptional professional methodologies into products with expert cognitive playbooks at their foundation.
Development focus
Development combines The Collective’s domain knowledge and standards of taste with evaluative judgment, verified learning and complementary AI capabilities from Emergent Partners. Each product is designed to preserve distinctive expertise and develop its application to new problems.
COGNITIVE EVALUATION & EXPERIMENTAL VALIDATION
The work must prove the intelligence.
Decision quality
Effort required
Independence achieved
We are developing evaluation methods that test understanding, reasoning, originality and judgment, including whether a system recognises its own limits.
Expert comparison, reproducible experiments and real-world outcomes test whether a proposed advance holds up on unfamiliar problems. We measure decision quality, the effort required and the independence achieved, and check whether improvement in one capability damages another.
Evidence from product use informs subsequent research. Successful methods become reusable capabilities within agreed intellectual property and confidentiality boundaries. Client-specific knowledge remains protected.