Asking the right questions is an essential part of using AI effectively. The questions you ask determine the outputs you receive — and if your question isn’t structured correctly, you won’t receive valuable responses.
In this blog, we’ll cover:
- Understanding AI capabilities
- How to craft effective questions
- Common mistakes when asking AI questions
- Challenges and limitations of AI
- Improve project management with Jira’s AI capabilities
- How to ask AI questions: Frequently Asked Questions
Understanding AI capabilities
When starting with AI, you need to understand that different AI tools have varying capabilities, and getting the right results for your needs will depend on the software you use.
- Rule-based AI tools use a set of predetermined rules and facts to deliver pre-defined outputs. These tools are simple and effective but are the most limited in their functions.
- Machine learning AI tools like Atlassian Intelligence have access to a set of data without predetermined rules and facts, which lets them evolve with new information. Machine learning tools are more effective for larger projects and businesses because the more data they process, the better they perform.
- Neural networks aim to teach AI how to process data like the human brain does. Unlike other types of artificial intelligence, neural networks’ outputs aren’t limited by the data they have access to, so they can “predict” answers and solutions.
Atlassian AI capabilities — and the capabilities of other AI tools — keep improving. However, there are limitations regarding what you can ask or tell AI software to do, so learning to ask effective prompts will ensure you’re making the most of those limitations.
Certain questions are best suited for different AI tools, so it’s crucial to understand the type of AI tool you’re using. Here’s a quick breakdown:
- Yes/no questions: Most AI tools do a great job answering simple yes or no questions. For example, ask, “Is it going to rain in San Francisco today?”
- Multiple-choice questions: If you provide multiple potential answers to a single question, AI can choose the ‘correct’ one. For instance, you could ask, “Which type of wood is the most affordable: oak, pine, or maple?”
- Open-ended questions: Some AI tools are better at answering open-ended questions than others. For example, you could ask, “What can I do to improve my product/service?”
You can identify the tool best suited for your needs by consulting documentation or guides, which typically provide insights into the tool’s strengths and limitations. This ensures you leverage the tool appropriately and get the most accurate and relevant responses for the type of question you’re asking.
How to craft effective questions
Before using artificial intelligence with Atlassian, you must understand how to craft questions that deliver the best results. Here are some basic guidelines on how to ask AI questions that get the results you’re looking for.
Be as specific as possible
Specificity is one of the single most important qualities of an AI question. Remember, AI can’t read your mind or pick up on context clues, so if your prompts are vague, your results will be too.
Let’s say you want to use AI to learn more about your target audience. Avoid broad prompts such as “Tell me about my target audience.” AI can’t guess WHAT you want to know about your target audience. Instead, be specific about the information you’re looking to receive. You could ask a simple question like this:
“How much do my customers spend on average?”
If you’re looking for more detailed information about customers, you could try something like this:
“How much did customers in North America spend on average last year?”
You could also modify your question to determine the average customer lifetime value or see which products or features perform the best in a particular region.
Be concise, avoid slang
While being specific can help you get better results when you ask AI questions, concise language is just as important. Avoid using phrases or slang that AI might have difficulty understanding.
Here’s a question you might ask a friend or colleague if you wanted a list of restaurants in your area:
“Where can I find delicious food that’s not too bad for me?”
Phrases like “delicious” and “bad for me” are jargon (and open to personal interpretation) used in everyday conversation. AI doesn’t know what you find delicious, or what “not too bad” means in this context. Instead, try asking AI a more straightforward question:
“Can you recommend restaurants that serve healthy food in my area?”
The exact language you use matters whether you’re using AI to drive content creation or asking a simple question.
Add context and relevant information
If you want responses to your question to be even more specific and useful, you must include context or relevant information in your query.
Let’s say you’re asking AI to generate a list of PR trends. You could try a simple query like this:
“Can you tell me about PR trends?”
If you ask a question like that, you will get a mixture of trends from different sources — some of which may be outdated or only apply to certain industries. If you’re looking for more specific PR trends, you could try this:
“Can you give me a list of 2024 PR trends and predictions in the software development industry?”
You don’t want to include too much information and make your question confusing, but there should be guardrails around it to help narrow the results.
Structure questions for optimal results
The structure of your questions is just as important as the actual information they contain. Even if you include all the right details, AI will have difficulty interpreting your question if the wording is confusing.
Let’s say you’re asking a simple question about the weather this week. Your query might look like this:
“Will it be nice enough outside tomorrow that I can golf?”
The structure of this question is clunky and odd, which makes it challenging for AI to parse. Try something like this instead:
“What is the weather forecast for San Diego today?”
If you’re not getting the results you want, try rephrasing your question and seeing if that helps.
Keep experimenting with inputs
There’s a learning curve whether you’re using AI for project collaboration or content creation. Don’t be afraid to experiment with different inputs to determine what works for you.
So remember, if you’re having trouble asking AI questions, try the following:
- Experiment with phrasing
- Add more context
- Avoid jargon
You can learn more about using specific AI tools like Atlassian Intelligence by looking at documentation and tutorials.
Common mistakes when asking AI questions
If you’re just starting with AI, making mistakes is common. Let’s take a look at some of the most frequent mistakes people make when they ask AI questions:
- Ambiguity: You need to be clear about what you’re asking and how you’re asking it to get the best results. Avoid jargon or words with two meanings that AI can take out of context.
- Misunderstanding AI capabilities: AI can only do so much, so you can’t expect great results if you’re asking questions AI isn’t capable of answering.
- Over-generalization: AI may not offer a relevant answer if your question is too broad. Make sure you’re including context to get the results you’re looking for.
These mistakes become less common as you experiment and learn.
Challenges and limitations of AI
While AI tools can help you streamline simple tasks and improve productivity and collaboration, AI isn’t perfect.
Bias is a big issue with AI because these tools are trained using data from humans — and humans are imperfect. Since that means some human bias is inherent in AI, you need to know how to identify biased outputs.
For instance, AI-generated hiring recommendations might favor candidates of certain backgrounds if the training data contained biased historical hiring patterns. It’s essential to identify and address these biases in AI’s results.
As a consumer, you must understand that AI works better for specific use cases. Manage your expectations and work within AI’s limitations for the best results. For more on how companies like Atlassian handle AI and user data, you can explore their stance on AI and data protection.
Improve project management with Jira’s AI capabilities
One use case in which artificial intelligence can play a crucial role is project management.
Atlassian’s AI capabilities in Jira and Confluence help bring together cross-functional teams to tackle any project. Atlassian Intelligence integrates seamlessly with both to deliver relevant results for your team and current project — and you can even use AI to brainstorm, create, transform, and organize content.
Try Jira today to see how AI can help you improve productivity and collaboration.
How to ask AI questions: Frequently Asked Questions
Should you use natural language when asking AI questions?
Using natural language can help you get better results when asking AI questions. However, it’s also important to note that AI parses language differently, so avoid jargon and overly technical terms.
Are there any ethical considerations to keep in mind when asking AI questions?
When using AI, it’s essential to understand that these tools can provide biased results. Since many AI platforms use slightly outdated data or directly leverage information humans have created, the answers you receive may not be current or may contain human error. You should also consider the impact of AI on society — including how AI can contribute to a better society when used responsibly.
Why is it important to ask AI questions strategically?
You should ask AI questions strategically to ensure you’re getting valuable outputs. If you’re using AI for research, the phrasing of your question and the context you provide will determine the information you receive. When using AI to solve a problem or make a decision using data, strategically asking questions can help ensure positive results.