Definition

Intent is the user's pursuit of their sought state — energy directed from their now-state toward their sought state, with continuous regulatory tracking and adjustment.

The three components

Intent decomposes into three parts that all have to be present for the pursuit to function. Energy is what the user brings — metabolic, attentional, motivational capacity for the pursuit. The offering doesn't generate it; the offering operates on energy the user already has. Sought state is what the energy is directed toward — the state the user is moving toward, conscious or not, articulable or not. Regulatory mechanism is the user's continuous tracking of whether the path is producing progress toward the sought state, with adjustments — continue, redirect, disengage — made on that basis.

The three are distinct primitives. Energy without a sought state is undirected. A sought state without energy is inert. Both without a regulatory mechanism would produce no adjustment when the path stops working.

Intent produces the engagement

Intent is the mechanical force that moves users along their timeline — what makes them pursuers rather than passive entities. Without it, there's no energy behind progression through any process the offering is part of. This is the foundational mechanism-claim about intent: it produces the interaction in the first place. Tap into it and the offering has access to the force already moving the user. Miss it and the offering has to manufacture momentum that was never going to come from the offering itself.

Sought state and resulting state

Sought state is user-side: a property of the user's directedness, present in the user prior to and independent of any particular transition the offering provides. Resulting state is transition-side: the terminus of a particular named transition, defined by completion criteria.

The two coincide when a transition is named such that its resulting state matches the user's sought state. That coincidence is the work intent-aligned design does — not a fact the framework can presuppose. A user has sought states whether or not any transition is well-aligned to them; a transition has a resulting state whether or not that resulting state serves any user. Treating the two as identical collapses the distinction the framework relies on for evaluating whether a transition serves the user at all.

Intent is fuzzy but theorizable

Intent is multi-layered: short-term and long-term pursuits operate simultaneously, intent shifts over the user lifetime, sometimes involves other people, sometimes changes within a single session. The framework doesn't claim intents are atomic units; it claims intent as a category is foundational to the framework and can be theorized rigorously enough to act on.

At runtime, intent detection reads ambient signals — search terms, referral context, campaign attribution, intent-revealing questions placed at the transition's entry — to form best-guess intent classifications. The detection is approximate inference, not deterministic classification, and the approximation is a feature rather than a limitation. Trajectory-reliability at the routing-relevant altitude is what the apparatus needs; ontological intent resolution is not.

Example

A freelancer's client demands an invoice tonight. The freelancer searches "send invoice fast" and lands on an invoicing tool's signup page. The components break down cleanly:

• Energy: deadline-driven attention, willingness to spend ten minutes setting something up.
• Sought state: an invoice sitting in the client's inbox.
• Regulatory mechanism: if the tool hasn't produced an invoice within a few minutes, the freelancer abandons it and reverts to a Word template or a PayPal request.

Another user signs up the same day to evaluate tools for replacing their accounting setup next quarter. Different sought state (a vetted recommendation), different energy profile (exploratory, not deadline-driven), different regulatory mechanism (willing to spend an hour comparing). Same offering, same signup cohort, two different intents — and the activation experience that serves the freelancer well won't serve the evaluator.

See also

• Energy
• Regulatory mechanism
• Multi-intent model
• Margin on efficacy
• Starting state
• Transition
• Path
• Approximate intent