Applied Research
Intelligence close to life.
Flink Labs explores the near future of intelligence as it moves into everyday objects, environments, relationships, routines, decisions, and small systems that live close to people.
Our work is practical, exploratory, and artefact-led. We build small systems, simulations, interfaces, agents and objects to observe behaviour, expose assumptions and make emerging possibilities concrete.
Research territory, not a service catalogue.
The three threads describe the territory Flink Labs returns to across independent research and client work.
A commercial engagement begins with an organisation’s real question, then draws on whichever research ideas, methods and technologies make that question testable.
Tangible Ambient Interfaces
Intelligence situated in the everyday world.
AI and data do not need to live only inside dashboards, documents, or chat windows. They can appear through objects, spaces, cards, lights, small screens, terminals, gestures, projections, and quiet ambient signals.
Tangible Ambient Interfaces explores how computation can become present in daily life without becoming intrusive. This includes physical devices, microcontrollers, NFC cards, projected interfaces, ambient displays, smart objects, and everyday artefacts that help people notice, understand, and act.
Questions this thread helps us explore:
- Where should an intelligent capability live if a chat window is not the right form?
- How might data become present in a room, object, tool or routine without becoming intrusive?
- What changes when an interface can be touched, carried, encountered or noticed peripherally?
- How can ambient intelligence remain understandable and under human control?
Playable Computation
Models and data as explorable worlds.
Most dashboards are static surfaces. They show results, but they rarely let people explore how a system behaves.
Playable Computation treats data, models, agents, and scenarios as interactive worlds. Instead of only reading charts, people can test choices, follow paths, change assumptions, inspect state, and see consequences unfold.
This thread draws from games, simulations, interactive fiction, procedural systems, cellular automata, artificial life, and complex systems visualisation. It is not about gamification. It is about making computation navigable, responsive, and rich enough to think with.
Questions this thread helps us explore:
- How might a complex model become something leaders, experts or communities can actively investigate?
- What would allow someone to change an assumption and immediately encounter its consequences?
- How can uncertainty, feedback and system behaviour become more visible than they are in a dashboard?
- When does interaction reveal something that a report cannot?
Distributed Intelligence
Computation spread across agents, devices, signals, and networks.
Not every intelligent system needs to be a single central brain. Some forms of intelligence emerge through many smaller parts communicating locally: agents, devices, sensors, protocols, messages, memories, signals, and people.
Distributed Intelligence explores systems where computation happens across the network. Intelligence lives in the relationships between parts, not only inside each individual node.
This includes agent-to-agent communication, local broadcast, peer-to-peer structures, gossip-like protocols, artificial pheromones, swarms, ad hoc networks, and small agents that coordinate over time.
The focus is on bounded, legible, non-monolithic systems. Less global command centre. More field radio, ant colony, neighbourhood network, or living ecology.
Questions this thread helps us explore:
- How might many small agents, devices or people coordinate without a single central controller?
- What forms of memory, signalling and local communication support useful collective behaviour?
- How can emergent behaviour remain bounded, observable and shapeable?
- When is a distributed system more resilient, legible or appropriate than a monolithic model?
Research in practice
The research becomes commercially useful when it changes what an organisation can see or decide.
Flink Labs often applies these threads within three connected organisational territories.
High-stakes knowledge work
Exploring how intelligent systems can assist expert judgement, evidence navigation and complex information work while preserving provenance, discretion and human responsibility.
Complex systems made playable
Turning models, scenarios, datasets and organisational dynamics into interactive environments that people can enter, manipulate and reason with.
Distributed and situated systems
Investigating how agents, devices, interfaces and environments might coordinate, adapt and provide intelligence outside a single central application.
These are not packaged solutions. They are places where Flink’s research methods can help a real organisational question become testable.
Intelligence should not only work. It should be understandable enough to reason with, design around, and trust.
See how research engagements work
Some questions need to be made before they can be answered.
Bring the emerging capability, difficult workflow, unusual system or technical possibility that your organisation cannot yet see clearly.
Flink Labs can help turn it into a bounded investigation, a working artefact and a more defensible decision about what should happen next.