Context is Everything in AI: Ask the Right Questions

By Stephen Fahey

Context is Everything in AI: Ask the Right Questions

EXECUTIVE SUMMARY

The quality of outputs you receive from AI systems depends fundamentally on the quality of your inputs. As organizations rush to implement AI tools, many miss this critical truth: context framing—achieved through thoughtful questioning—determines success. This article presents a three-part framework for executives and knowledge workers to maximize AI value by approaching these systems as "cognitive partners" rather than magic solution providers. By establishing proper context, positioning AI as a specialized teacher, and applying cognitive load management techniques, organizations can extract significantly more value from their AI investments while maintaining human judgment at the center of decision-making.

The Context Challenge in Enterprise AI

When Mark Henderson, CIO of a mid-sized financial services firm, first introduced generative AI tools to his organization, the results were disappointing. Despite investing in premium enterprise accounts, his team received outputs that were generic, occasionally inaccurate, and rarely insightful. "We had access to the same technology as our competitors," Henderson recalls, "but we weren't seeing the transformative results we'd been promised."

The problem wasn't the AI—it was how they were framing their questions.

As AI becomes ubiquitous across industries, a clear pattern has emerged: organizations that methodically build contextual frameworks for AI interaction dramatically outperform those that approach these systems with vague requests or unrealistic expectations. Research from MIT's Sloan School of Management suggests that companies with structured AI interaction protocols report 37% higher satisfaction with AI outputs and 28% greater implementation success.

The solution begins with understanding a fundamental truth: context equals framework, and framework equals principles.

A Three-Part Framework for Contextual AI Interaction

1. Question Formulation: The Foundation of Context

"To get the right principles, you've got to ask the right questions," explains Stephen Fahey, founder of Anthrotech AI. "It doesn't mean it's rocket science, but if you look at the bigger picture, context equals a framework. That equals principles."

Effective AI interaction begins with thoughtful question formulation—a skill many executives haven't developed because traditional computer systems didn't require it. Unlike database queries or search engines, generative AI systems benefit from carefully constructed contextual information:

Poor Context: "Give me marketing strategies."

Strong Context: "I need differentiation strategies for a premium SaaS product ($80/month) targeting mid-sized financial services firms competing against lower-cost alternatives. Our primary value proposition is enhanced regulatory compliance and audit preparation."

The difference between these approaches is stark. The first will generate generic marketing advice; the second creates a framework for precisely relevant outputs.

Organizations seeing the greatest AI impact typically implement:

• Context templates for common business functions
• Question formulation training for teams working with AI
• Progressive refinement processes where initial outputs inform more specific follow-up queries

2. Positioning AI as a Specialized Teacher

Once you've established initial context, conceptualize the AI system not as a magical answer generator but as a specialized teacher with whom you're having a collaborative dialogue.

"Imagine the AI model is now a teacher," suggests Fahey. "You can put that in: you are a teacher, you're an AI expert, or you can just ask it the questions."

This mental model shift has proven transformational for organizations struggling with AI implementation. By viewing the interaction as a teaching relationship:

• Teams ask more profound follow-up questions
• They challenge outputs appropriately rather than accepting them uncritically
• They build upon initial responses rather than treating them as endpoints
• They maintain agency in the knowledge-building process

Ravi Mehta, former CPO at Tinder and current AI strategy consultant, recommends explicitly defining the AI's role at the beginning of important interactions: "Starting with 'I want you to act as an expert in [relevant domain]' establishes expectations and often improves the quality of the exchange."

3. Cognitive Load Management

Perhaps counterintuitively in our digital age, one of the most effective techniques for maximizing AI value involves stepping away from the screen.

"Ideally, you still want to make notes on pen and paper," Fahey advises. "Because I think what a lot of people are doing in tier-one systems of just talking to AI is... we need to focus more and our brains work better when we lower the cognitive load."

This insight aligns with cognitive science research showing that handwritten note-taking improves conceptual understanding and retention compared to digital documentation. For complex AI collaborations, implementing a structured workflow pays dividends:

1. Identify knowledge boundaries (what you know, what you need to know)
2. Build initial contextual frameworks based on existing knowledge
3. Engage AI with targeted questions to fill gaps and expand understanding
4. Document key insights manually to facilitate deeper processing
5. Iterate through this cycle until you've developed a framework that's "good enough for that situation"

Implementing Contextual AI Practices

Organizations that excel at AI implementation typically establish clear protocols for different interaction types. Consider developing guidelines for:

Strategic Questions: Require comprehensive context setting, multiple perspectives, and explicit assumptions.

Operational Questions: Create templates with standard contextual elements for routine operations.

Creative Explorations: Establish frameworks that encourage divergent thinking while maintaining relevance.

Capital One's AI Center of Excellence provides a compelling case study. The financial services giant implemented a "context canvas" approach for teams engaging with AI systems. This standardized framework prompts users to articulate:

• Business objective
• Existing knowledge and assumptions
• Constraints and requirements
• Key stakeholders and their needs
• Success metrics

Teams complete this canvas before significant AI interactions, resulting in what Chief Data Officer Linda Apsley describes as "dramatically more relevant and implementable outputs."

The Future of Contextual AI

As AI capabilities advance, the importance of contextual framing will only increase. Models will become more powerful at addressing precisely formulated challenges but will still struggle with vaguely defined problems.

"Step one will always be identify your knowledge, build your framework to give you context," Fahey emphasizes. "How do you build your framework? Well, go back and ask AI questions, make notes, and then keep going on that cycle until you simplify something that's good enough for that situation."

Organizations that master this contextual approach will find themselves with a sustainable competitive advantage not easily replicated by competitors who view AI merely as a push-button solution.

Questions for Leaders

As you implement AI tools throughout your organization, consider:

1. Do we have standardized approaches for constructing context in different business scenarios?
2. Have we trained our teams in effective question formulation for AI interactions?
3. Are we capturing and refining successful contextual frameworks for reuse?
4. How are we integrating traditional critical thinking with AI collaboration?

The answers to these questions will likely determine whether your AI investments deliver transformative value or merely incremental improvements.