Structure Intent

Alpha 0.1 · Published 7 November 2026 · Evolving

The structured intent workforce framework

AI changes how work is divided between people and machines. As implementation gets cheaper, human judgment becomes the scarce resource, and structured intent becomes the artifact around which that judgment is organized. This framework describes the work, who is accountable for it, and the competencies it relies on.

It is an alpha: a catalog of the work and its accountabilities, not a prediction of a future profession’s organization chart. The field is moving fast, and real projects will shape what comes next. Its approach follows the NICE Workforce Framework for Cybersecurity, which describes work and capabilities rather than job titles. It is not affiliated with NIST.

The human discipline determines intent

Deciding what a system should mean and do: what it is for, which trade-offs are acceptable, and what must never happen. It needs judgment, systems thinking, critical thinking, domain knowledge, authority and accountability.

The technical discipline structures intent

Expressing what was decided so machines can build from it and check against it. It needs modeling, formalization, evidence, verification and bindings.

A specification methodology covers only the second. Structured intent organizes both: human judgment and machine execution around a shared, durable expression of purpose.

The central value of an intent architect isn’t writing more specifications. It’s being able to understand complex situations, reason rigorously about them, and express what should be true in a form that people and machines can act on.

Principles

  1. Roles are not jobs. One person may hold several roles in a small organization; a large one may have several people per role.
  2. Authority is never delegated to an agent. AI agents can help perform much of the work. They cannot silently acquire the authority to decide what the organization values.
  3. Departments are represented. Structuring intent inside an enterprise is a team effort. Each role with a stake speaks for its department’s interests.
  4. Every role keeps learning. Learning is not a specialty. It is part of every role, all the time.

Competencies

Systems thinking and critical thinking are the foundation. They make every other competency useful.

1

Systems thinking

Seeing the whole system: its parts, relationships, dependencies, boundaries, feedback loops, stakeholders and unintended consequences.

2

Critical thinking

Testing what is believed or proposed: examining assumptions, evaluating evidence, spotting contradictions, separating fact from inference, and judging whether a conclusion is warranted.

Systems thinking discovers what the system actually is and what it must accomplish. Critical thinking tests whether that understanding is sound. Together they turn understanding into structured intent that can be trusted.

3

Research and rapid learning

Getting up to speed on an unfamiliar domain, system or market in days, using AI as much to learn as to build. Every role needs it, all the time.

4

Domain expertise

Knowing the practices, constraints, exceptions and consequences of an area of the business, and why its rules exist.

5

Abstraction and modeling

Finding the right level to describe a system, and the boundaries, layers and lines of sight between them.

6

Specification and formalization

Stating intent so neither a person nor a machine can misread it: clauses, examples, priorities and prohibitions.

7

Verification and evidence evaluation

Knowing what proven, tested, checked at runtime and assumed each establish, and when the evidence is enough for the risk.

8

Governance, judgment and accountability

Ruling on what is a promise, an accident or a bug, representing a department’s interests fairly, and owning the decision.

9

Technical translation and implementation

Carrying intent into technologies through bindings, using the strongest mechanism each one offers.

Work roles

Three kinds of contribution: the intent owner decides, domain experts inform, and intent practitioners do the structuring work. These need not be separate people, but each responsibility must be explicit. Each role lists its purpose, typical tasks, the competencies it relies on most, and what AI agents contribute.

Decide

The authority and accountability to determine what the system is meant to achieve.

Intent owner

Holds the business authority to rule on the intent, and is accountable for its validity.

  • Rule on whether behaviors are promises, accidents or bugs
  • Approve boundaries: what software decides and what people decide
  • Settle conflicting priorities between departments
  • Approve changes to the intent

Relies most on Systems thinking, Critical thinking, Governance, judgment and accountability

Agents contribute Summarize options, consequences and evidence for each ruling.

Inform

The knowledge that determines what the intent should say, from every area with a stake.

Domain expert

Brings knowledge the code doesn’t contain, and represents a department’s interests.

  • Explain why rules exist and who relies on them
  • Identify obligations the code never stated
  • Speak for their department when priorities conflict
  • Review clauses that touch their area

Relies most on Systems thinking, Critical thinking, Domain expertise

Agents contribute Draft questions, gather precedents and surface where the code contradicts practice.

Practice

Intent practitioners: they discover, structure, reconcile, verify, govern and bind what was decided into an actionable intent.

Intent architect

Designs the structured intent and brings the right people to each decision.

  • Draw boundaries and lines of sight between layers
  • Structure goals, guarantees, prohibitions, priorities and examples
  • Facilitate rulings and record who decided
  • Keep the intent independent of any implementation

Relies most on Systems thinking, Critical thinking, Abstraction and modeling, Specification and formalization

Agents contribute Draft clauses, propose boundaries, flag ambiguity and contradictions.

Intent discovery specialist

Lifts intent from existing systems, so actual behavior can be compared with intended behavior.

  • Extract candidate clauses from code, tests, logs and documents
  • Find where sources disagree
  • Separate likely promises from likely accidents
  • Prepare rulings for the intent owner

Relies most on Systems thinking, Critical thinking, Research and rapid learning, Technical translation and implementation

Agents contribute Read code at scale, extract candidate clauses and propose examples.

Intent integration architect

Merges intent from several systems into one.

  • Align overlapping capabilities across systems
  • Turn conflicts into explicit decisions
  • Define shared contracts between systems
  • Record which system each clause came from

Relies most on Systems thinking, Critical thinking, Abstraction and modeling, Governance, judgment and accountability

Agents contribute Lift each system in parallel and map overlaps and conflicts.

Intent verification engineer

Establishes how strongly each promise is kept, and measures the quality of the intent itself.

  • Set the evidence each clause needs for its risk
  • Design tests, proofs and runtime checks
  • Measure completeness, for example with deliberately broken versions
  • Report evidence honestly, including what is only assumed

Relies most on Critical thinking, Verification and evidence evaluation, Technical translation and implementation

Agents contribute Generate tests and proof attempts, run mutation testing, report gaps.

Intent governance lead

Keeps authority, accountability and change control working across many intents.

  • Maintain who rules on what, across systems
  • Track assumptions and open questions to closure
  • Run impact analysis when goals change
  • Set the maturity level each system should reach

Relies most on Systems thinking, Critical thinking, Governance, judgment and accountability

Agents contribute Follow lines of sight, produce impact reports and flag stale evidence.

Intent binding engineer

Lowers intent into implementations in specific technologies, and produces the evidence.

  • Map each clause to the strongest mechanism a technology offers
  • Decide what can be enforced statically and what needs runtime checks
  • Preserve intent across technology migrations
  • Direct AI coding agents against the intent

Relies most on Critical thinking, Specification and formalization, Technical translation and implementation

Agents contribute Generate implementations and tests from the intent.

Combining roles

Small organization

One intent architect also does discovery and verification. A senior leader is the intent owner. Domain experts are consulted per system.

Mid-size organization

Intent architects work across a portfolio, with dedicated verification and binding engineers. Owners and domain experts sit in each business area.

Large enterprise

Every role is staffed, with an intent governance lead coordinating authority, impact analysis and maturity targets across systems.

What comes next

Future versions will add task, knowledge and skill statements and proficiency levels, drawn from the work of early adopters, and will turn emerging responsibilities into formal roles only when practice shows they are needed. If you are structuring intent in your organization, your experience can shape the next version: open an issue on GitHub.

See also hiring intent architects and structured intent for leaders.

You may share and adapt this framework with credit: Structured intent workforce framework, alpha 0.1, jaynryan, structureintent.com (CC BY 4.0).