Build an orthogonal context brief for AI-powered product analysis

AI Gives Generic Answers Because You Give Generic Prompts

The most common complaint about AI tools is that they produce generic, surface-level output. The response reads like it could apply to any company, any product, any situation. But the problem is rarely the AI. The problem is the context you gave it.

The Problem

When you ask an AI assistant to "write a product strategy," it has to guess everything: your market, your customers, your constraints, your competitive position, your team's capabilities, your company's values. With no context, the AI falls back on the most probable response given its training data. That response is, by definition, generic. It is the average of everything it has seen.

Most product managers provide just enough context for the AI to start but not enough for it to be specific. They describe the problem but not the market. They name the product but not the user. They state the goal but not the constraints. The result is output that sounds plausible but is not useful.

A 2023 Nielsen Norman Group study found that AI-generated content with rich contextual prompts scored 73% higher on relevance assessments compared to the same queries without context. The quality ceiling of AI output is set by the quality floor of your input.

How This Prompt Works

The Orthogonal Context Brief is a structured pre-prompt that assembles the non-obvious context an AI needs to produce specific, useful output. "Orthogonal" because it focuses on the context dimensions that are perpendicular to the main question, the background information the AI cannot infer from the request itself.

The brief covers six dimensions:

• Domain context: Industry, market segment, regulatory environment
• User context: Who they are, what they have tried, why existing solutions fail them
• Business context: Revenue model, growth stage, strategic priorities
• Technical context: Platform constraints, integration requirements, technical debt
• Team context: Team size, skills, decision-making style
• Temporal context: Deadlines, market timing, competitive urgency

By front-loading this context, every subsequent prompt in a session benefits from specificity. The AI stops guessing and starts reasoning within your actual constraints.

According to a 2023 McKinsey report on generative AI in the enterprise, teams that use structured context frameworks report 2.6 times higher satisfaction with AI-generated outputs and reduce the edit-to-publish cycle by 45%.

When to Use It

• At the start of any AI work session to establish the operating context
• When switching domains to prevent the AI from carrying assumptions from a previous topic
• Before generating strategic documents where generic output is actively harmful
• When onboarding a new team member to an AI workflow, as a template for how to provide context

Common Pitfalls

• Over-contextualizing. There is a point of diminishing returns. A 3,000-word context brief for a simple email is wasteful. Match context depth to task complexity.
• Providing context once and assuming persistence. AI sessions have memory limits. For long conversations, periodically re-anchor the key context points.
• Confusing context with instructions. Context tells the AI about your situation. Instructions tell it what to do. Both are necessary; neither is sufficient alone.
• Neglecting temporal context. The same strategy question has different answers in January versus October, pre-funding versus post-funding, before a competitor's launch versus after. Time matters.

Sources

• Nielsen Norman Group. (2023). AI Prompt Quality and Output Relevance. https://www.nngroup.com/articles/ai-prompting/
• McKinsey & Company. (2023). The State of Generative AI in the Enterprise. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai/
• Mollick, E. (2023). *Co-Intelligence: Living and Working with AI*. Portfolio. https://www.oneusefulthing.org/