🤓Best practices in working with amberAI

With these best practices, you can make your work even easier and more efficient and get the most out of amberAI.

1. Using the right chat

Choose the right amberAI chat for your needs.  

• The amberAI General Chat is flexible and can deal with more complex questions. Chat histories are saved.

• The amberAI Document Chat answers specific questions about an open document.  

• The amberAI Bookmarks Chat is perfectly suited as a chatbot for you and your colleagues to pick up and reflect the knowledge within a group with collected files.

• amberAI Answer helps you to get to your search result faster for both simple and complex search queries if the entire database is to be searched.

2. Precise wording

Make sure that your enquiry is formulated clearly and precisely. So that amberAI understands what is meant and can provide you with the most relevant information accordingly. Avoid colloquial language and short forms. Here are a few examples:

3. Include relevant information

Include any relevant details that may be important to answering the question. This can include specific terms, dates, names, contextual information, etc. 

This allows the AI to understand what information it should limit itself to for answer generation.

For example:

❌ "Give me some sample customer enquiries."

• This is not a good enquiry as the question is too open-ended. The AI would give a much too long answer because it doesn't know exactly what specific information you need.

✅"Can you provide me with 5 sample customer enquiries from Zendesk?"

• This request is better because it helps amberAI provide accurate results tailored to your request.

4. Use of search operators

Use search operators to refine the results and make them more specific. This can be particularly useful if you are looking for very specific documents or information. 

For example: 

❌ "Projects January 2023" 

• This query is not precise enough, so amberSearch will show you everything possible from the month of January. In this case, you will have to rummage through the search results for a long time until you get to your desired file. 

✅"Projects January 2024 type:pptx -Project XY" 

• By adding the file type operator and the minus operator to the search query in this example, only PowerPoint presentations will be listed in the search results and ‘Project XY’ will be excluded. This means that the search results can be narrowed down very well and amberSearch will only show you the other files that are relevant to you. 

You can find more information about the operators here.
 

5. Using filters

Just like the operators, the search filters also help you to narrow down your search and get to your desired search result more quickly.  
Once you have entered a search query, you can choose from the following filters directly below your search line:

1. File types
2. Time
3. Language
4. Data sources

You do not need to change anything in your search query, but simply select the desired filters. For example, under file types, you select that only Word files are displayed and under time you limit the results to the last three months.

 You can find more information about the filters here.

6. Ask follow-up questions

Use the opportunity to ask follow-up questions to obtain further details or clarify ambiguities. Either use the follow-up questions suggested by amberAI or formulate your own.

7. Have complex texts summarised or reworded by amberAI

If you have to deal with a complex text and possibly also in another language, amberAI can help you to reformulate or translate this text into simpler words.

Simply write the following request to the amberAI General Chat:
"Summarise the following text in simple words in a maximum of 10 sentences: -inserted text-"

8. No analyses, but precise results

amberSearch is NOT a content analysis tool, it is a language model. What do we mean by this? amberSearch is not suitable for analysing content, e.g. in Excel tables, or for identifying a person or object for you via image search.

amberSearch is not an analysis tool, but a language model.

It can provide you with excellent results when searching for metadata or keywords. This means that amberAI can reproduce all the information available to it.

Let's illustrate this with an example:

❌"Tell me the number of projects we carried out in 2023."

• With this query, amberAI will give you an unreliable answer because amberAI itself does not draw any logical conclusions. To actually get a number, this would have to be explicitly noted somewhere in the available documents.

✅"Where can I find an overview of all projects from 2023?"

• This request is better, because this way you get a proper overview of amberAI.

9. Give feedback

Feel free to give us feedback at any time so that we can continuously improve the quality of amberAI for you. 

Here you can see how you can contact us.