How to get the best answers in Robin AI Reports

Getting the best outputs in Robin AI starts with asking the right questions—but it doesn’t end there. Answer Types and Answer Previews play a crucial role in ensuring you get the insights you need in the right format, every time.

In this tutorial, we’ll go through how to fine-tune your report answers, test them in real time, and improve the overall accuracy and consistency of your reports.

‍

1. Why Refining Prompts Matters

When you refine your prompts in Robin AI Reports, you:

• eliminate ambiguity,
• ensure Reports deliver structured, reliable insights, and
• minimise rework and avoid wasted report runs.

‍

2. Understanding Answer Types

Robin AI Reports offer a range of Answer Types, allowing you to control the format of each response.

‍

How to Choose the Right Answer Type:

• Use structured formats (Yes/No, Number, Select, Multi-Select) when you want standardised, easily comparable answers.
• Use Text (Summary) or Text (Word or Phrase) when the answer cannot be represented in a more structured format, or where nuance or additional context is helpful.

‍

3. Using Answer Previews to Refine in Real-Time

Answer Previews let you test how your questions will perform before running the full report. This helps you:

• quickly identify whether the phrasing of your question results in the expected answer,
• adjust your Answer Type if you need more structure or detail, and
• check answers in real-time.

How to Use Answer Previews:
After you create a report and add your relevant documents

1. In the Template Builder, add your topic and question.
2. Select the Answer Type.
3. Use Answer Preview to immediately see Robin’s response to that question on a single contract.
4. Refine your question or change the answer type based on the output.
   • If the answer seems vague, consider moving to a structured answer type.
   • If you need more context, switch to Text (Summary).

4. Practical Example of Refining an Answer

Scenario: You need to find out whether a contract includes a liability cap, and if so, its amount.

Step 1: Add two separate questions.

1. Does the contract include a liability cap? (Answer Type: Yes/No)
2. What is the amount of the liability cap? (Answer Type: Number - Currency)

Step 2: Use Answer Preview to test both.

• If the answer to the first question is "No", the second should logically return "N/A."
• If both return valid responses, your structure is correct.

‍

5. Best Practices for Refining Prompts

✔ Be precise in phrasing your question.

✔ Use structured Answer Types for consistent data extraction and analysis, to ease any downstream decision-making.

✔ Use Summary answers where more context or explanation adds value.

✔ Test with Answer Previews to save time and improve accuracy before running a full report.

‍

Conclusion

Refining answers in Robin AI Reports ensures that your outputs are accurate, clear, and actionable. By choosing the right Answer Types and leveraging Answer Previews, you can confidently deliver reports that drive informed decisions.

For more tips, visit Robin AI University.

‍