• Introduction

Intent & Interaction

All interaction has a substrate: the actor perceives a situation, interprets what it means, decides what to do, and acts. Intent — the relationship between what an actor needs and what a system offers — forms and sharpens across many such loops.

Two dimensions follow. How engagement unfolds over time traces the arc of work across six stages, from loose recognition to action to reflection. Navigation describes how those stages produce different traversal patterns.

Interaction cycle

Every interaction shares the same basic shape:

  • See – perceive a situation
  • Think – interpret what it means and decide what to do
  • Do – act on that interpretation Action produces a new situation to perceive, and the loop repeats.

What varies across frames is how each phase can be assisted and who runs the loop at any given moment:

  • When the human stays in the loop and the system helps at each phase, the relevant frame is assistance — the kinds of support a system can offer at each stage of the actor’s cognitive cycle (perceiving, knowing, planning, performing, reflecting).
  • When the system runs the loop in the background and the human returns at discrete moments, the relevant frame is delegation — the lifecycle and touchpoints of agentive systems.

How engagement unfolds over time

Each stage represents a different kind of work, from first recognition through enactment and reflection. Each stage typically contains many turns around the interaction cycle.

The framework connects navigation and action patterns into one narrative.

Trigger & need recognition

Intent: recognizing gaps and opportunities.

Resolution state: intent emerges — neither party is yet committed. See formality on early-stage looseness.

Patterns here surface possibilities without requiring commitment — they make latent needs visible, lower the threshold for engagement, and create openings for intent to form.

Seeking & access

Intent: retrieving and gathering relevant information.

Resolution state: intent decomposes — the goal breaks into sub-intents; system structure begins shaping what is possible.

Patterns here help actors break goals apart and retrieve what’s relevant — they narrow scope, expose structure, and let the system’s organisation do some of the thinking.

Evaluation & relevance construction

Intent: assessing relevance and value.

Resolution state: intent sharpens — dimensions acquire values, irrelevant paths are pruned, criteria become explicit.

Patterns here support judgement — they let actors compare, contrast, and inspect at varying levels of detail, revealing what matters and discarding what doesn’t.

Sense-making & integration

Intent: synthesizing and organizing insights.

Resolution state: intent integrates — dimensions connect as the actor builds a coherent picture. Resolution consolidates.

Patterns here help actors organise, connect, and externalise emerging understanding — they turn scattered findings into structure the actor owns.

Commitment & enactment

Intent: taking decisions, completing tasks.

Resolution state: intent commits — resolved intent is enacted. The system executes what was agreed.

Patterns here operationalise decisions — they confirm, execute, and provide feedback proportional to the commitment involved. Friction should match reversibility.

Reflection & feedback

Intent: communicating outcomes, adapting future behaviours.

Resolution state: intent reflects — outcomes are evaluated against the original need. Feedback loops into future intent formation. Adaptability matters most here: systems that learn from this stage shape better relationships next time.

Patterns here make outcomes visible, shareable, and reusable — they close the loop between what was intended and what happened, feeding forward into the next cycle.

Activity types and agency

Each stage of engagement involves different kinds of activity with different agency distributions. The two axes cross freely.

The intent relationship expressed as movement. Two dimensions shape navigation behaviours: intent resolution (how clear is the goal?) and structural mapping (how well does the information architecture match the actor’s mental model?).

The combination produces three categories (following Dan Ramsden):

  • Motivated movement — resolved intent + supportive structure → directional progress
  • Delightful discovery — unresolved or absent intent → the system can help it form
  • Foggy finding — resolved intent + poor structural mapping → friction (a relationship failure, not an actor failure)

Motivated movement

Purposeful progress through a known or emerging structure. Intent is sufficiently resolved for directional movement, though precision varies. Feedback from the system reinforces progress, and transitions emerge as clarity increases.

Movement with precision and high intent. The actor knows both what they’re looking for and how the structure is likely organised. Success is driven by targeting (choosing a specific goal) and predictive understanding of the structure.

Transitions

Can emerge from browsing when confidence increases, or from transactional search when results are explored.

Browsing

Less precise but still purposeful movement. The actor works within a known part of the system, refining both their goal and their understanding as they go. They engage in sweeping (broad scanning) to reveal patterns or next steps. Browsing supports learning-in-action, often serving as a bridge between low and high precision behaviours.

Transitions

Can sharpen into navigating, or loosen into exploring if uncertainty grows.

A focused, tool-based method for retrieving specific results. The actor constructs inputs (queries, filters, prompts) with care and intention, relying on the system to evaluate, discriminate, and deliver. This is not exploratory—interrupting or failing to deliver damages trust. Transitions out of this mode are usually unwanted, unless the actor gives up and switches strategies.

Transitions

Navigational transitions are discouraged – switching into browsing or navigating can feel like system failure unless clearly supported. Successful search results in doing the job.

Delightful discovery

Intent is unresolved or absent — the system can help it form. Discovery happens when actors encounter information they didn’t explicitly seek but find valuable, relevant, or exciting. This category rewards designs that support curiosity, enable surprise, and create psychological safety for getting a little lost. Transitions often depend on motivation and momentary clarity of purpose.

Exploring

Movement through loosely bounded content or ideas, often triggered by a vague goal or simply curiosity. The actor’s understanding of both the content and their own need sharpens as they go. Sweeping dominates early stages; conditioning emerges as patterns form. Exploration can either settle into more directed behaviour or remain open-ended.

Transitions

May mature into browsing or navigating once a goal crystallises

Monitoring

A strategy for staying updated about known areas of interest. Can be active (manual checks) or passive (automated updates). Monitoring creates continuity over time and can either lead to deeper engagement or enable light, ongoing awareness. It’s a low-friction way to maintain relevance and context.

Transitions

Can lead into exploring, transactional search, or navigating when something changes or catches attention.

Passive discovery

Serendipitous encounters with unexpected relevance. The actor may not have been looking for anything, but good design sparks interest, prompting targeting and movement almost simultaneously. This is where systems surprise and delight—turning idle moments into value.

Transitions

Often initiates movement or exploration; can also feed directly into the “job”.

Foggy finding

Intent is resolved but can’t map to the system’s structure — a relationship failure, not an actor failure. The actor knows what they want but can’t access it easily. Friction stems from memory gaps, unclear structures, or unhelpful tools. The risk is high frustration, especially when actors believe the system is “hiding” something from them. Transitions here are about either clarifying the goal or improving system feedback to allow escape into more fluent modes.

Re-finding

The actor wants to return to something they’ve seen before, but lacks the exact pathway. They rely on cues, memory, or system support (like history or recognisable patterns). This mode rewards good differentiation, meaningful labels, and predictable paths.

Transitions

May evolve into navigating or transactional search if memory strengthens

  • Any model with clear structure and good wayfinding helps
  • Multilevel tree — hierarchy provides memory cues

Uncovering

Systematic trial-and-error. The actor is motivated but blocked by structural opacity or unclear interaction rules. This is often a sign of poor affordance or overly strict syntax (e.g. voice commands, search logic). Good design reduces the opacity of systems to make uncovering unnecessary.

Transitions

May improve into transactional search or navigating when actor figures out “how the system thinks”

Hunting

Similar to uncovering but more adaptable and determined. The actor doesn’t care where the solution comes from—only that they find it. They may use multiple strategies simultaneously (search, scanning, filtering, backtracking). This is an intense, high-motivation behaviour that thrives when systems reward skill and persistence.

Transitions

Can resolve into navigating or browsing; if prolonged, risks frustration or abandonment

Cross-cutting qualities

Several qualities cut across everything this page describes. They apply at every level — within a single turn around the interaction cycle, across any stage of the engagement arc, and across any delegation touchpoint. They are orthogonal to both the loop and the arc: any loop phase and any arc stage can involve any distribution of any of them.

  • Agency — who is driving at any given moment, and what the system contributes at each stage
  • Temporality — the rhythm and pacing of turn-taking
  • Formality — parallels the arc: as engagement matures, so does the structure around it, loose to precise

Activity types (agency across activity types and intent resolution) describe what kind of work is happening. The arc tracks how committed the relationship is; activity types track what kind of work is happening within it. The two axes cross freely — an actor whose relationship with the system is still forming may be expressing a query, perceiving results, and thinking about criteria within a single stage.

Resources & references

Related patterns

Enacts

  • Formality — parallels the arc: as engagement matures, so does the structure around it, loose to precise
  • Adaptability — the engagement arc reshapes itself as intent sharpens; what the system offers shifts stage by stage
  • Agency — who is driving at any given moment, and what the system contributes at each stage
  • Temporality — the rhythm and pacing of turn-taking

Related