This article summarises best practices for setting up the Custom Topics in Bliro's AI Note templates. You can add headers, questions, or even full "ChatGPT-like" prompts to your custom topics. To learn how to set up AI Note Templates in general read this article.
Bliro harnesses the latest AI technology to generate notes for your meetings. Every AI, including Bliro, has its boundaries, and the input quality significantly impacts the output quality. These suggestions are designed to help you develop tailored topics that deliver impactful insights.
❓ What are Bliro custom topics?
Bliro's custom topics turn general AI notes into fully custom meeting documentation that matches your exact needs. They are especially powerful for structured conversations like recruiting interviews or sales demo calls and for extracting specific information from conversations, such as the decisions that were made in the meeting or potential blockers that came up.
For optimal results, we suggest choosing between custom topics, a chronological summary, or AI topics for your templates. Chronological summaries and AI topics are particularly effective for capturing the essence of unstructured conversations. However, combining custom topics with chronological summaries or AI topics in the same template may reduce the detail and quality of the extracted information for each category.
🗓️ Focus on relevant details
The core of your custom topics should be the meeting transcripts. When adding new topics, make sure they're directly related to the kind of discussions you generally have to ensure the information gathered is both precise and valuable.
✍🏻 Be as specific as possible
Include details in your descriptions and add context to get more relevant outputs. Make sure to tell the AI if you are interested in information that concerns specific meeting participants (Pain Points -> Customer Pain Points, Wishes -> Wishes of the candidate concerning their new job).
Less effective ❌ | Better ✅ |
Problems | Customer Pain Points |
Start-Up Information | The Pain the Start-Up solves for their customers |
Design-related feedback | Design-related positive feedback, negative feedback, and ideas |
🧮 Avoid topics related to numerical analysis
Bliro's AI is not specifically built for number crunching. The clarity of numerical data in the transcripts plays a crucial role in the accuracy of such tasks. Instead of asking for calculations, aim for text-based insights where Bliro excels.
Examples to avoid:
Calculating the speaking time percentage of each participant
Comparing Q2 to Q3 signup numbers
Examples to use:
Inquiring about a potential customer's budget, if it was mentioned in the meeting
Asking for the Q3 user churn rate, if it was mentioned in the meeting
🙅🏻 Some topics can induce bias in the summaries. Beware!
We advise you to be mindful of ethical issues when extracting information with custom topics. Since AI learns from extensive datasets that might contain inherent biases, it's crucial to avoid questions that could perpetuate stereotypes or discrimination. Focusing your queries on "who and/or what" of the actual meeting discussions can help reduce bias.
Bad description ❌ | OK description ✅ |
Who should run this project? | Who did the team decide should run this project? |
Why should we hire this candidate? | Does this candidate meet the qualifications we listed for this role? |