Explore

Using AI Right Now: Executive Summary

The Big Picture

The AI landscape has shifted from "best model" to "best overall system." Three platforms now dominate for serious professional use, each requiring a $20/month investment for full capabilities.

Why It Matters

Competitive advantage: Early adopters gain significant efficiency improvements

Cost-effectiveness: $20/month investment can deliver substantial productivity gains

Feature maturity: AI systems now offer professional-grade capabilities including research, document analysis, and code execution

The Bottom Line

Choose one of three systems: Claude (best privacy), ChatGPT (most features), or Gemini (Google integration). Ignore free versions—they're demos, not professional tools.

Key Decisions

Platform Selection

Claude: Never trains on your data, excellent for sensitive work

ChatGPT: Full feature set, can disable training, most versatile

Gemini: Strong Google Workspace integration, weaker privacy controls

Model Tiers That Matter

Fast models: Quick questions, brainstorming (default setting)

Powerful models: Complex analysis, high-stakes work (manually select these)

Ultra-powerful: 20+ minute processing for hardest problems

Game-Changing Features

Deep Research

Produces comprehensive reports with actual citations

Significantly more accurate than basic AI responses

Impresses information professionals (lawyers, consultants, researchers)

Use cases: Competitive analysis, regulatory research, market studies

Document Analysis

Upload multiple file types simultaneously

Cross-document analysis and synthesis

Real-time collaboration on complex materials

Voice + Visual Mode

Point camera at documents, screens, or physical objects

Real-time analysis while mobile

Underutilized: Most people use like Siri instead of leveraging visual capabilities

Implementation Strategy

Immediate Actions (Week 1)

Subscribe to chosen platform ($20/month)

Configure privacy settings appropriately

Test with real work: Switch to powerful model, provide full context, engage interactively

Power User Techniques

Context loading: Upload relevant documents and background materials

Specific deliverables: Request exact outputs (reports, analyses, summaries)

Interactive refinement: Push back, ask follow-ups, explore alternatives

Branching: Edit prompts to explore different approaches

Critical Guardrails

What's Improved

Hallucination rates significantly decreased with newer models

Prompting complexity no longer critical—conversational approach works

Citation accuracy much better with research modes activated

Persistent Risks

Overconfident errors: Still gives wrong answers with high confidence

Capability confusion: May claim to do things it cannot actually do

Context limitations: Only knows what you tell it in current conversation

ROI Indicators

Research time: 70-80% reduction for comprehensive analysis

First draft quality: Professional-grade starting points for documents

Ideation volume: Generate 50+ options instead of 10

Verification accuracy: Better fact-checking when web search enabled

What's Next

The 80/20 rule applies: Most users treat AI like Google search (quick questions, no context, default settings). Power users provide context, use advanced models, and engage interactively—delivering dramatically better results.

Success metric: If you're not regularly surprised by AI capabilities, you're likely underutilizing the technology.

⁠

Source: "Using AI Right Now: A Quick Guide" by Ethan Mollick, June 2025

Want to print your doc?
This is not the way.

Try clicking the ··· in the right corner or using a keyboard shortcut (

CtrlP

) instead.