Lumu’s AI Filter enables you to query incidents using natural language, similar to everyday conversation. With a single prompt, you can apply multiple filters simultaneously, allowing you to refine results and quickly find the information you need.

How to Access Lumu AI Filter

You can find Lumu’s AI Filter in the Incidents view. Scroll down to find the usual search bar and beside it, you will find the AI Filter button.

By clicking on it, you will turn on AI Filter to query for incidents using natural language.

Now, let’s take a look at how to use AI Filters effectively to get the best results.

How to Use Lumu AI Filter Effectively

Lumu’s AI Filter can deliver focused results, but it is continuously learning and evolving based on user input. As a result, there are some limitations to the types of queries they can process. This section outlines the query types that yield the most accurate and effective results.

Bear in mind that Lumu AI Filter uses all of the filters available in the traditional Lumu Incidents search bar to yield results such as threat types, number of endpoints, contacts, etc. This means that your queries will be interpreted using these filters.

Here are some recommended queries that you can use to test AI Filter and to build your own:

You can query for specific incident types in a specified time period. Here are some examples of this query used practically:

C2C incidents this month

Phishing incidents detected last week

Playback incidents found last month

You can also look for specific dates

Incidents created on October 15th

Malware incidents generated on January 2nd

You can query for incident types that match certain conditions such as a specific number of contacts, number of endpoints, etc. Here are some examples of this query used practically:

Filter malware incidents with at least 3 contacts

Network scan incidents with more than 20 contacts

Get C&C incidents with more than 5 endpoints

This can be used to create more general queries such as:

Top incidents by Endpoints reached

Top incidents by contacts/requests

Filter incidents created using Playback

You can also use this structure to ask questions

Which incidents haven't been automatically responded to?

You can use these queries as building blocks to make more complex and specific requests such as:

As you can see, this request stacks some of the conditions and parameters used above to create a more particular search which will yield very specific results.

These are the basics on how to use Lumu AI Filter. You can try building your own queries using these examples as a basis. Now, let’s take a look at the results.

Lumu AI Filter Results

After entering your query the corresponding filters will be applied, which you can validate by looking at the filters below the search bar(1).

If the results match your query, we encourage you to rate the filter’s accuracy positively(2). Likewise, if it misinterpreted your request, please rate it negatively so we can improve it based on your feedback.

If the filter was applied correctly and it yielded results, you should see that reflected below:

If the query fails to yield any results, you will see the following message:

Please verify whether the query was interpreted correctly by the AI Filter, and if it wasn’t, please send your feedback using the provided option so we can keep refining this feature and improve our results.