← all ideas

AI Better as a Consultant

The idea

The argument is that when you ask an AI model to act as a consultant or provide advice, the quality of the response often falls short of real-world expertise. However, if you simply ask the AI what real people are doing to solve specific problems, it can provide more accurate and practical solutions. The reasoning behind this is that the AI does not generate new advice but instead accesses patterns from its training data. This approach bypasses the limitations of the AI's advisory role and provides more actionable insights. By framing your queries in a way that asks for real-world examples, you can tap into the vast and diverse experiences encoded in the AI's training data, leading to more effective problem-solving strategies.

Why it works

• The argument is that when you ask the AI to act as a consultant, it tends to revert to a more generic, advisory tone, which can be less helpful.
• On the other hand, requesting the AI to identify how real businesses are solving the problem directly accesses the training data and provides concrete, pattern-based solutions.
• This approach bypasses the AI's limitations in acting as a true expert and instead taps into its ability to analyze and describe existing practices, leading to more practical and effective advice.

The playbook

1. Ask for real-world solutions: When you encounter a problem, frame your request by asking the AI to identify real-world solutions that businesses have implemented to address similar issues. For example, if you need advice on revenue management, ask, "What are actual revenue management companies doing to solve this problem?"

1. Compare with expert advice: After receiving the AI's response, compare it with advice from a general expert. This will help you understand the difference between theoretical advice and practical solutions. Ask the AI to act as an expert and provide advice, then ask it to identify real-world examples.

1. Test consistency: Regularly test the consistency of the advice by asking the AI to provide both expert advice and real-world examples. Observe which type of response is more actionable and relevant to your specific situation.

1. Identify patterns in training data: Recognize that the AI's ability to provide real-world examples is due to its access to training data. When you ask the AI to describe what businesses are actually doing, it bypasses the "helpful assistant" frame and directly accesses the training data, providing more accurate and relevant information.

1. Refine your prompts: As you continue to use this strategy, refine your prompts to get the best results. For instance, if you find that the AI provides more detailed and actionable advice when you specify the industry or context, adjust your prompts accordingly. This will help you get the most out of the AI and make your problem-solving process more effective.

Where people get it wrong

Where people get it wrong is in expecting AI to act as a personal advisor or coach. When you ask the AI to act as a consultant, it may provide generic advice based on its training data, which can be helpful but not always tailored to your specific situation. Instead, try asking the AI to identify real-world solutions and strategies that other businesses have used to address similar issues. This approach often yields more practical and actionable insights.

Another common mistake is assuming that AI will always provide the best solution. While the AI can offer valuable perspectives, it's crucial to critically evaluate its suggestions and cross-reference them with your own knowledge and experience. Relying solely on AI can lead to overlooking important details or missing the nuances of your specific case.

Lastly, some might underestimate the importance of context and domain-specific knowledge. The AI, while powerful, is still limited by its training data. Asking it to provide solutions based on real-world examples from specific industries can help bridge this gap. Always ensure that the AI's responses are relevant to your industry and context to avoid generic or irrelevant advice.

Do this next

• When facing a business problem, try asking the AI to describe real-world solutions instead of seeking advice. For instance, if you're struggling with customer retention, ask the AI, "What are the top strategies revenue management companies use to improve customer retention?"
• For a more targeted response, frame your question in a way that directs the AI to focus on real-world examples. For example, instead of asking, "What should I do to increase sales?" try asking, "How are companies in the tech industry managing their sales processes effectively?"
• Use the AI to gather data on how other businesses have tackled similar issues. You can ask, "Can you provide a list of successful marketing campaigns from the past year that increased brand awareness?" This can give you practical insights and ideas.
• Test the AI's capabilities by asking it to compare different approaches. For example, if you're unsure about a new marketing strategy, you can ask, "How do the marketing strategies of companies in the e-commerce sector compare to those in the travel industry?" This can help you make more informed decisions.