Skip to content

Overview: Why Strategy Before Artifacts Matters

The Central Claim

A product strategy should be complete enough that you could rebuild the entire product from it using different technology or a different team, and users who knew the old product should experience it as the same product.

This means:

  • Users need no training or adjustment period; familiar workflows remain intact
  • Input produces the same outcomes as before
  • Core navigation, business logic, and UX patterns are preserved
  • The product feels continuous, even if design details, wording, or underlying code differs

When this is true, your strategy has become your source of truth. Everything else – code, epics, stories, docs – derives from it.

Why It's Called MOMENTUM

The name emerges from two core insights:

  1. Strategy is the spine. At any given moment, you have a single source of truth: your product strategy. Everything else – epics, stories, designs, code – hangs from that spine. Knowing your strategy at any point in time means you have a complete understanding of intent, not scattered across meetings, docs, and code.

  2. Reliable strategy creates momentum. When you always know your strategy, you gain speed and confidence. You can make decisions quickly because constraints are explicit. You can onboard new team members instantly by pointing them to strategy. You can delegate confidently to AI or junior PMs because they have clear boundaries to work within. Unreliable, opaque strategy creates drag, confusion, and rework. Reliable strategy creates momentum.

Why This Matters Now

The Traditional Problem

Most teams have strategy buried in:

  • Slide decks that become outdated (and no one ever updates them after a management meeting where new insights were given)
  • Jira tickets, Wikis, Miro boards, photos of flipcharts from the last strategy offsite
  • Meetings, Slack conversations, and tribal knowledge
  • Implementation choices in the codebase (if you want to know what the product does, read the code)

As a result:

  • Strategy decays. Decisions that made sense 12 months ago are forgotten; teams make contradictory local choices.
  • New team members have no reference. Onboarding a new PM means either re-learning everything through conversation or tring to make sense of all the sources above.
  • Features drift from intent. "This seemed reasonable at the time" becomes the only rationale for complex features.

Why AI Amplifies This Problem

Did you ever ask AI for an implementation plan and got a random 6-month project plan with invented dates?

Large language models are extremely good at creating output fast. You can easily spin up working prototypes of new concepts within half an afternoon. But without a clear context and specifications, this quickly becomes AI slob which isn't really actionable.

The more un-supervised content is created, the harder it becomes for both, humans and LLMs to know what's important and actually part of the strategy.

Instead of making the strategy clearer, AI dillutes it - unless there are the right safeguards in place.

How MOMENTUM Works

MOMENTUM treats product strategy as a Single Source of Truth (SSOT) that is explicitly owned by the Product Manager and kept in a repository.

Strategy is written in Markdown, versioned with Git, and reviewed deliberately. It is not inferred from code, meetings, or tools. Everything else derives from it.

AI and agents are used to reduce manual effort – not to replace judgment. They help keep strategy, evidence, and downstream artifacts in sync so the PM can focus on intent and decisions rather than busy work.

The Three Layers

MOMENTUM structures product work into three layers, with human ownership at the top.

Layer 1 – Human Input

Humans own intent.

This includes:

  • Strategic decisions
  • Constraints and priorities
  • Outcomes from meetings and discovery
  • Corrections based on real-world feedback

LLM instructions and prompts exist only to structure, preserve, and propagate this input. They do not define product direction.

MOMENTUM comes with multiple prompts to help you as PM to iterate on your strategy. It helps to improve the strategy but it's always the PM who has the final word!

Layer 2 – Single Source of Truth

The SSOT captures what humans decided and why.

It consists of:

  • Strategy – the authoritative description of the product’s purpose, scope, constraints, and success criteria
  • Evidence – research, production learnings, and documented trade-offs that justify the strategy

This layer is durable, versioned, and optionally reviewed through pull requests. It changes intentionally.

Layer 3 – Output Artifacts

Artifacts are downstream representations of strategy:

  • Epics and stories
  • Documentation and release notes
  • Designs and prototypes - or even production code

They are replaceable and regenerable. Their role is to express strategy in executable forms, not to redefine it.

Direction of Flow

  • Human input feeds strategy.
  • Strategy cascades into artifacts and code.
  • AI helps keeping everything in sync.

This keeps strategy stable, intentional, and authoritative while still grounded in real-world outcomes.