ai-product-positioning
Skill1 starsChanged 22 days agoMITView on GitHub
Produces positioning for an Artificial Intelligence (AI) product that survives the three hardest buyer objections - 'why not just use a general-purpose assistant', 'it is only a wrapper', and 'the model provider will ship this next' - with a scored differentiation ladder, proof-point inventory, and messaging hierarchy. Use when positioning a new AI product or feature; rewriting messaging that sounds like every other AI startup; preparing for investor or buyer questions about defensibility; or responding to a foundation-model provider launching an overlapping capability.
In this repository41 other files it gives its agents.
- ai-activation-optimizerSkill
- ai-content-quality-gateSkill
- ai-feature-experiment-designerSkill
- ai-launch-playbookSkill
- ai-pricing-margin-modelerSkill
- ai-search-visibility-optimizerSkill
- agent-autonomy-level-mapperSkill
- ai-failure-state-designerSkill
- ai-usability-test-plannerSkill
- ai-ux-auditSkill
- human-in-the-loop-designerSkill
- prompt-surface-designerSkill
- uncertainty-and-explainability-patternsSkill
- ai-agent-financial-analystSkill
- burn-rate-runway-plannerSkill
- capital-allocation-frameworkSkill
- dcf-valuation-builderSkill
- financial-statement-analyzerSkill
- saas-cohort-analyzerSkill
- competitive-response-modelerSkill
- mechanism-design-plannerSkill
- nash-equilibrium-solverSkill
- negotiation-strategistSkill
- repeated-game-strategistSkill
- ai-decision-engineSkill
- ai-product-strategySkill
- ai-stakeholder-translatorSkill
- customer-interview-synthesizerSkill
- jobs-to-be-done-extractorSkill
- opportunity-solution-treeSkill
- product-discovery-coachSkill
- product-roadmap-prioritizerSkill
- blue-ocean-strategy-canvasSkill
- go-to-market-strategySkill
- gtm-engineeringSkill
- market-sizingSkill
- okr-cascade-plannerSkill
- porters-five-forces-analyzerSkill
- pricing-strategy-analyzerSkill
- strategic-landscape-mapperSkill
- swot-tows-analyzerSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 17
No credible answer to "I can do this with a chat assistant and a good prompt"
- Distinctive: line 26
After a foundation-model provider or platform announces an overlapping capability
- Distinctive: line 64
Write a one-paragraph, evidence-backed answer to each.
- Distinctive: line 99
Fast-follow scenario: if a model provider ships the core capability free in 6 months, which pillar survives?
- Distinctive: line 120
Use "AI-powered", "revolutionary", or model names as the headline value proposition
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AI Product Positioning
If your pitch still works after the buyer opens a general-purpose assistant in another tab, you have positioning.
What this skill is
A positioning workflow built for products whose core capability is powered by models anyone can rent. It starts from the buyer's real alternatives (including a general-purpose assistant and doing nothing), scores where durable advantage actually comes from across five layers, and turns only the defensible layers into a message hierarchy with proof. It deliberately refuses "powered by AI" as a value proposition.
What it solves
- Messaging that leads with the model or "AI-powered" instead of the job and the outcome
- No credible answer to "I can do this with a chat assistant and a good prompt"
- Defensibility claims that collapse when a model provider ships the same feature
- Proof points that are demos rather than measured results
- Positioning that is identical to five funded competitors using the same models
When to invoke
- Pre-launch positioning for an AI-native product or a new AI feature in an existing product
- Rewriting a homepage or pitch deck that reads as generic AI marketing
- After a foundation-model provider or platform announces an overlapping capability
- Preparing sales for "why not just build it ourselves with an Application Programming Interface (API)"
- Investor or board questions about moat and fast-follow risk
Phase 1: Map the real alternatives
List what the buyer would actually do instead, and why they would pick it.
| Alternative | Example | Why a buyer picks it | Where it breaks for this job |
| General-purpose assistant | A chat assistant plus manual prompting | Already paid for, flexible | |
| Do it yourself | Internal build on a model API | Control, perceived low cost | |
| Incumbent adds AI | Existing system of record ships a feature | Already integrated, procurement done | |
| AI-native competitors | Similar startups | Similar pitch | |
| Status quo | People and spreadsheets | Zero change cost |
The "where it breaks" column is the raw material for positioning. If it is empty for the general-purpose assistant, stop and fix the product before the messaging.
Phase 2: Score the differentiation ladder
Rate each layer 0-3 on current strength and on durability against fast-follow (how long until a model provider or incumbent matches it).
| Layer | What counts | Strength (0-3) | Durability (0-3) | Evidence |
| Model | Fine-tuned or specialized models, measurable quality gap on the target task | |||
| Data | Proprietary or compounding data, feedback loops, domain corpora | |||
| Workflow | Embedded in the system where work happens, multi-step automation, integrations, approvals | |||
| Distribution | Existing customer base, channel partners, marketplace placement, community | |||
| Trust | Accuracy guarantees, security certifications, audit trails, human review, liability terms |
Layer score = Strength x Durability (max 9)Lead with the highest-scoring layer. Model-layer advantages usually score low on durability; workflow, data, and trust usually score highest. A layer with Strength 3 and Durability 0 is a launch talking point, not positioning.
Phase 3: Neutralize the three objections
Write a one-paragraph, evidence-backed answer to each. Test each answer on 3-5 target buyers.
| Objection | Weak answer (avoid) | Strong answer pattern |
| "Why not a general-purpose assistant?" | "Ours is smarter" | Name the steps, context, and integrations the assistant lacks; quantify time or error difference on a real task |
| "It is just a wrapper" | "We have proprietary AI" | Show the system around the model: data pipeline, evaluations, workflow, guardrails, outcomes |
| "The model provider will ship this" | "They will not" | Explain why your layer (workflow, data, trust, distribution) is outside the provider's incentive or reach; show you benefit when models improve |
Phase 4: Build the proof-point inventory
| Proof point | Type | Metric | Source | Verifiable by buyer? | Shelf life |
| Outcome (time saved, errors reduced, revenue) | |||||
| Evaluation result on a named task set | |||||
| Customer reference or case study | |||||
| Security or compliance attestation |
Rank proof by: outcome metric beats evaluation score beats testimonial beats demo. Flag any proof older than 6 months - model progress makes quality claims stale fast.
Phase 5: Messaging hierarchy
| Level | Content | Constraint |
| Positioning statement | For <segment> who <job>, <product> is the <category frame> that <outcome>, unlike <primary alternative>, because <top ladder layer> | Internal; no adjectives without proof |
| Headline | Outcome for the job, not the technology | Under 10 words; no "AI-powered" |
| Three pillars | One per top-scoring ladder layer | Each pillar maps to 2+ proof points |
| Proof row | Metrics, logos, attestations | Dated and sourced |
| Objection handlers | From Phase 3 | Used in sales and Frequently Asked Questions (FAQ) |
Choose the category frame deliberately: an existing category ("better <tool>"), a sub-category ("<tool> for <segment>"), or a new category. New categories need budget and patience; most AI products should anchor to an existing budget line.
Phase 6: Pressure-test and fast-follow plan
- Swap test: replace your brand with a competitor's in the headline - if it still reads true, rewrite
- Assistant test: have a skeptic attempt the core job with a general-purpose assistant for 30 minutes; document where it fails
- Fast-follow scenario: if a model provider ships the core capability free in 6 months, which pillar survives? Make that pillar more prominent now
Output
- Alternatives map including the general-purpose assistant and do-it-yourself options
- Scored differentiation ladder with evidence and durability ratings
- Written answers to the three core objections
- Proof-point inventory ranked by strength and freshness
- Messaging hierarchy: positioning statement, headline, three pillars, proof row, objection handlers
- Fast-follow scenario note with the surviving pillar identified
Operating rules
Always
- Position against the general-purpose assistant explicitly, even if no buyer names it
- Lead with the job and the measurable outcome, then the system that delivers it
- Score durability separately from current strength
- Tie every pillar to dated, verifiable proof
- Re-run the ladder whenever a major model or platform release lands
Never
- Use "AI-powered", "revolutionary", or model names as the headline value proposition
- Claim a model-quality moat without a repeatable evaluation behind it
- Dismiss the fast-follow risk instead of planning for it
- Promise accuracy levels the product does not reliably achieve in production
- Copy the category language of better-funded competitors and expect to win on features
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

