Skip to content
Intent DrivenProduct

The Manifesto

Making intent explicit.

An Agile & Scrum‑aligned manifesto for Intent/Spec‑Driven, AI‑assisted delivery.

Inspired by Agile, informed by Scrum, enabled by Spec‑Driven Development.

Preamble

What we have learned

  1. 1

    We are uncovering better ways of delivering valuable products by combining Agile principles, Scrum discipline, and AI‑assisted execution.

  2. 2

    Through this work, we have learned that when AI participates in delivery, ambiguity compounds faster than velocity.

  3. 3

    Speed is no longer constrained primarily by coding capacity. It is constrained by clarity, decisions, and intent upstream.

  4. 4

    Therefore, we commit to making intent explicit, decisions deliberate, and Scrum practices purposeful rather than ceremonial.

Core values

What we value more

While there is value in the items on the right, we value the items on the left more.

  • Intent over interpretation

    Clear product intent, constraints, and trade‑offs over inferred requirements.

  • Explicit decisions over emergent understanding

    Problem framing, evidence, and hypotheses over “we’ll figure it out later.”

  • Executable specifications over implicit alignment

    Living, testable specs over shared assumptions and tribal knowledge.

  • Outcomes over output

    Measurable customer and business impact over story throughput.

  • Empiricism with guardrails over speed without control

    Learning fast within constraints over unbounded experimentation.

  • Scrum discipline over Scrum theatre

    Purposeful use of Scrum events over ritual compliance.

Principles

Principles of Intent‑Driven, Spec‑Driven Agility

  1. Principle 01

    Intent is the primary product artifact

    Intent is not a conversation; it is an explicit and shared understanding of:

    • The user
    • The problem
    • The business impact
    • The constraints
    • The risks
    • The tradeoffs

    If intent is unclear, work should not start.

  2. Principle 02

    Specs are written after decisions, not to discover them

    Specifications are not the starting point.

    Before a spec exists, teams must have:

    • Hypothesis
    • Clear problem framing
    • Evidence and insight
    • One or more testable hypotheses

    Specs exist to lock decisions in, not to surface them during development.

  3. Principle 03

    Specifications are constraints/guardrails, not documentation

    Specs are not descriptions of work after the fact.

    They are the single source of truth that constrains execution for humans and AI.

    A good spec makes it easier to say no than yes.

  4. Principle 04

    Hypotheses are the backbone of scope

    Every meaningful product initiative is anchored in hypotheses that:

    • Are grounded in evidence
    • Define measurable outcomes
    • Enable tradeoffs
    • Filter scope

    If a feature does not support the hypothesis or mitigate its risk, it does not belong in the spec.

  5. Principle 05

    Code is an expression of intent, not the source of truth

    When intent or decisions change, specs change first.

    Implementation follows.

    Code without clear intent creates costly rework and hidden risk.

  6. Principle 06

    Change is welcome when intent remains explicit

    Adaptation is expected.

    Ambiguity is not.

    Changes are handled by:

    • Correctly classifying the change
    • Updating intent and specs
    • Reapproving when necessary
  7. Principle 07

    Product Owners are owners of intent, not ticket queues

    Product Owners are accountable for:

    • Hypothesis
    • Problem framing
    • Scope and tradeoffs
    • Spec quality

    Their leverage comes from clarity, not coordination.

  8. Principle 08

    AI accelerates clarity — or hides its absence

    AI is a force multiplier.

    Used well, it:

    • Accelerates synthesis
    • Improves consistency
    • Surfaces gaps

    Used poorly, it:

    • Fabricates confidence
    • Masks missing decisions
    • Scales ambiguity faster than humans ever could

    Humans remain accountable for decisions. AI must never stand in for unresolved thinking.

  9. Principle 09

    Quality is designed before it is tested

    Quality emerges from:

    • Clear constraints
    • Explicit acceptance criteria
    • Known risks and assumptions

    Testing verifies understanding — it does not replace it.

  10. Principle 10

    Continuous improvement applies to thinking, not just doing

    We continuously improve:

    • How we frame problems
    • How we express intent
    • How we use AI
    • How decisions flow through the system

Next

Hybrid Scrum, intentionally applied

Closing

This manifesto does not replace Agile or Scrum. It completes them for an AI‑assisted world.

When execution accelerates, clarity must arrive earlier.

We commit to paying the cost of thinking sooner because it is cheaper, faster, and safer there.