The AI Product Strategist for Roadmaps and Positioning

Product strategy fails in a specific way: slowly, then suddenly. The signs accumulate, a roadmap that lists features without stating what bets they represent, OKRs that track activity instead of outcomes, a competitive narrative that describes competitors without positioning the product against them, a quarterly planning cycle that produces alignment around a slide deck nobody refers to three weeks later. Each failure looks manageable in isolation. Together they produce a team that is busy, well-intentioned, and systematically building toward a target that has never been explicitly defined. The PM's job is to be the person who defines that target, keeps it legible to the team, and makes the decision-making framework explicit enough that the team can make good decisions without asking for permission at every junction. That job requires a kind of analytical discipline that most ai product strategy tools do not model, they generate roadmap templates and OKR examples, which is to strategic thinking what a blank form is to a filled one. Crest was built to do the thinking, not just produce the form.

Why the generalist approach fails at product strategy

A generalist chatbot can produce an OKR in the correct format. What it cannot do is produce an OKR that is actually good, measurable, ambitious without being fictional, connected to the company's strategic bets, and structured so that the key results are genuinely leading indicators of the objective rather than lagging proxies for it. "Increase user engagement" is an objective. "Reach an NPS of 50" is a lagging proxy that will not move the behavior the team needs to change. The difference between a functional OKR and a performative one is analytical judgment about what moves what, which requires understanding the product's causal model, the chain of user behaviors and product interventions that connects team actions to business outcomes. A generalist told to write OKRs produces plausible-sounding output in the right format. That is different from a good OKR.

McKinsey-style strategy consultants produce the other failure mode: frameworks without product context. A classic competitive analysis from a consulting engagement produces Porter's Five Forces for the industry, a 2x2 positioning map with axes chosen for legibility rather than decision relevance, and a set of strategic recommendations written for a board audience rather than a product team. The output is impressive and actionable by nobody. Product strategy is not industry analysis, it is the specific question of which bets to make, in which sequence, given the specific constraints and strengths of a specific product at a specific moment in its growth. The generalist consulting framework does not answer that question; it describes the landscape around it.

Roadmap tools, ProductPlan, Aha!, Roadmunk, solve the visualization and communication problem. They produce well-formatted roadmap views for stakeholders. What they do not produce is the strategy that the roadmap represents. A roadmap without a strategy is a list of things the team intends to build. A roadmap with a strategy is a sequence of bets, each one chosen because it tests a specific hypothesis about what will create value for a specific customer segment, with explicit trade-offs against the things the team decided not to build and the reasoning for those trade-offs. The tool visualizes the list; the strategy produces the reasoning that makes the list defensible.

What a product strategist actually does

A senior product strategist starts with the question the product exists to answer: what progress are we uniquely positioned to help a specific customer make, and why would they choose us over the alternatives? From that question, they derive the positioning (what we are in the customer's mind relative to the alternatives), the north star metric (the number that moves when the positioning is working), the roadmap (the sequence of bets that tests and builds the positioning), and the OKRs (the quarterly commitments that hold the team accountable for making progress on the bets). These are not independent documents, they are a connected chain of reasoning, and breaking any link in the chain produces strategy that looks complete but is not. A roadmap disconnected from positioning produces features that satisfy stakeholders and do not differentiate the product. OKRs disconnected from the roadmap produce measurement theater.

The competitive dimension of strategy is equally important and equally mishandled. Real competitive positioning is not a feature comparison matrix, it is an analysis of which customer segments are underserved by the current alternatives, what jobs those segments need done, and where the product has a defensible advantage in those jobs. Finding that white space requires understanding both the customer's JTBD and the competitive landscape's current coverage, which is why the best product strategists work closely with user researchers. The strategic output is a positioning statement that names the specific customer, the specific job, and the specific way the product does it better, not a generic claim to being "better, faster, cheaper" across all dimensions.

Meet Crest

Crest is Tonone's dedicated AI product strategist, a purpose-built agent for the full strategy and planning workflow, from competitive reconnaissance through roadmap prioritization and OKR design. It does not produce slides or templates; it produces the reasoning that makes a strategy defensible: the prioritization rationale behind the roadmap, the competitive white space behind the positioning, the causal model behind the OKRs. Crest brings the analytical discipline of a senior product strategist to the decisions that determine what gets built, and just as importantly, what gets cut.

Tonone's Crest is the AI product strategist that produces RICE-prioritized roadmaps with explicit bets and trade-offs, OKRs with measurable key results, and strategy memos grounded in competitive white space.

What Crest actually does

Building prioritized roadmaps with explicit trade-offs

The crest-roadmap skill takes a product context and a set of candidate initiatives and produces a RICE-scored roadmap with explicit rationale for every prioritization decision. Each initiative in the roadmap is framed as a bet: the hypothesis being tested, the customer segment it is designed to serve, the evidence for the reach and impact estimates, and the key risks that could invalidate the bet. The roadmap output includes a trade-off section, the initiatives that were considered and deprioritized, with the reasoning, because a roadmap that only shows what was chosen provides no basis for alignment, whereas a roadmap that shows what was not chosen and why provides the decision framework the team needs to handle scope requests and priority conflicts. crest-roadmap also produces a sequencing rationale: why this order, what the dependencies are between bets, and where the roadmap has optionality built in (places where a failed bet should change the sequence, rather than being followed regardless of what the team learned). The output is a decision document, not a Gantt chart.

Writing OKRs with measurable key results

The crest-okr skill produces an OKR framework for a quarter or a year: objectives that describe the qualitative direction the team is heading, key results that are genuinely measurable leading indicators of the objective, and an explicit causal model linking team actions to the key results. Each key result in the crest-okr output includes the measurement method, the current baseline, the ambitious target, and the team or person accountable for moving it. The skill also identifies the initiatives that are expected to move each key result, creating the explicit link between the roadmap and the OKRs that most frameworks describe in theory and skip in practice. crest-okr flags OKRs that are structurally weak: objectives that are aspirational but not directional, key results that measure output rather than outcome, or key results that will move regardless of whether the team's specific bets succeed or fail. These flags prevent the most common OKR failure mode, writing OKRs that look ambitious, track easily, and provide no signal about whether the strategy is working.

Mapping competitive landscape and positioning white space

The crest-compete skill produces a competitive landscape analysis focused on finding positioning white space rather than cataloguing feature differences. The output includes: a map of the major alternatives in the category (including the non-consumption alternative, doing nothing or using a spreadsheet), the customer segments each alternative primarily serves, the jobs each alternative does well, and the jobs that are underserved or misserved by the current alternatives. The white space analysis identifies the combinations of customer segment and job where the current alternatives have the weakest coverage, which is where a focused product has the strongest defensible positioning. crest-compete also produces a positioning statement in the Dunford framework: the competitive alternative, the product's unique attributes, the value those attributes produce, and the best-fit customer who values that specific combination. This is the positioning statement that survives a pitch meeting because it is built on competitive reality rather than aspiration.

Writing strategy memos that produce alignment

The crest-narrative skill produces a strategy memo: a written document that states the problem, the insight that changes the framing, the strategic approach that follows from the insight, and the bets the team is making. Strategy memos written in prose rather than bullet points are harder to produce but significantly more durable, a bullet list invites interpretation at every point, while a well-structured memo leaves fewer gaps for misalignment to hide in. The crest-narrative output follows a deliberate structure: the problem as the team currently understands it, the evidence that challenges that framing, the revised problem statement that the evidence supports, the strategic bet that follows from the revised framing, the key risks and the conditions under which the bet would be wrong, and the first three actions that follow from adopting the bet. This structure ensures that the memo can produce alignment rather than just documentation, it tells the reader not just what the team is doing but why a reasonable person who disagrees would update toward agreeing. Memos structured this way become the artifacts that resolve future priority debates without requiring a meeting.

Strategic reconnaissance before planning begins

The crest-recon skill performs a rapid strategic audit before any planning work begins: it maps the current product's positioning, identifies the north star metric and whether it is well-defined, reads the current roadmap for strategic coherence, and flags the biggest strategic risks, places where the team is making a bet that is not grounded in evidence, where the positioning is contested by a well-funded competitor, or where the OKRs are measuring the wrong things. The output is a strategic health brief: what is working, what is fragile, and what is the single most important decision the team should make before the planning cycle starts. crest-recon is the step that prevents the planning cycle from optimizing around the wrong problem, the recurring pattern where a team produces a well-executed plan for the wrong strategic direction because nobody paused to evaluate whether the direction was right before investing in the plan.

A worked example

A product team is entering Q3 planning with five candidate initiatives: a mobile app, a Zapier integration, an analytics dashboard, a team collaboration feature, and a white-label offering. The CEO wants the mobile app. The head of sales wants white-label. The PM has two days to prepare a defensible roadmap recommendation. They run crest-recon first, which flags that the current north star metric is ill-defined and that two of the five initiatives have no clear hypothesis for which customer segment they serve.

Then crest-roadmap produces the prioritization output. A portion looks like this:

## Crest, Q3 Roadmap Recommendation

### The bet
Our highest-leverage Q3 move is deepening the core activation loop for
team-size-3-to-10 accounts before expanding the surface area of the product.
Rationale: this segment has the highest NRR and the most room for expansion;
five of six churned accounts in Q2 cited lack of team features, not missing
mobile or integration capability.

### Prioritized initiatives

**1. Team collaboration (Reach: H, Impact: H, Confidence: H, Effort: M)**
  Hypothesis: adding shared workspaces + role-based access will increase
  team-size expansion rate from 1.4 to 2.2 seats/account within 60 days.
  Risk: scope creep into full permissions system, scope to 2 roles only in Q3.
  Trade-off made: ahead of mobile (mobile usage is <8% of active sessions).

**2. Zapier integration (Reach: M, Impact: M, Confidence: H, Effort: S)**
  Hypothesis: removes the #2 integration blocker in sales objections.
  Sequences after team collab because it amplifies a working team loop.

**3. Analytics dashboard (Reach: M, Impact: M, Confidence: M, Effort: L)**
  Hypothesis: increases perceived value for admins who drive renewal decisions.
  Risk: low confidence, no user research to validate admin usage pattern.
  Recommendation: run echo-interview study before committing build capacity.

### Not this quarter (with reasoning)
- **Mobile app**: 7.8% of sessions are mobile; investment doesn't match reach.
- **White-label**: no ICP definition for resellers; would dilute support capacity.

### OKR connection
O: Become the default collaboration layer for small technical teams.
KR1: Team-size expansion rate ≥ 2.2 seats/account (60-day cohort).
KR2: Team collab feature 30-day activation ≥ 40% of new accounts.
KR3: Zapier integration removes from top-3 sales objection list.

The PM presents this recommendation to the leadership team. The CEO's mobile request is addressed explicitly, not ignored, but deprioritized with a reason grounded in data rather than a PM's preference. The head of sales's white-label request gets a clear condition for reconsideration (define the ICP first). The OKRs are connected directly to the roadmap, so the quarterly review will produce a genuine signal about whether the strategy is working rather than a status update on ticket completion. Crest produced that quality of analysis in the time it would have taken to build the first slide.

Crest vs the alternatives

Crest does not produce slides, frameworks, or templates, it produces the strategic reasoning that makes a roadmap defensible, an OKR achievable, and a strategy memo durable. The comparison below makes the functional differences concrete.

Tonone's Crest crest-compete skill maps competitive white space by identifying which customer segments are underserved by current alternatives, producing a positioning statement grounded in competitive reality, not aspiration.

Tonone's Crest crest-narrative skill produces strategy memos in the problem/insight/approach structure, artifacts that resolve future priority debates without requiring a meeting because the reasoning is explicit and reviewable.

Install and try

Tonone is free and MIT-licensed. Install it once and all 23 agents, including Crest, are available in your Claude Code session. You pay only for the Claude Code token usage during work. Start with crest-recon before your next planning cycle to get a strategic health brief before committing to a direction.