Skip to main content

Understanding Ask AI

Have specific questions about your data? Use Ask AI to get clear, direct answers.

Written by Garen DiBernardo

What is Ask AI?

Ask AI is our new way to get a deeper understanding of your customer voice. Essentially, you can select any type of data you want (such as from a certain timeframe or touchpoint) and then Ask AI what you want to know about this data. Our AI will then generate you a clear and reasonable response to your question.

We offer analytics in multiple languages, but Ask AI was trained and modeled to work primarily with English questions and responses. You can ask questions in "any" language, and Ask AI will attempt to respond in that language, but for best results we recommend you ask your questions in English.

How does it work?

Ask AI looks at each piece of feedback in your data and analyzes it for keywords and phrases that are relevant to your question, then generates a response. This means it won't be able to answer questions like "how was the weather today" (unless for some reason that's what your customers are talking about, we don't judge 👀).

Note that if you are using Ask AI on conversations, the AI will look at the summary of the conversation and not the entire conversation when generating a response.

What happens to my data that goes through Ask AI?

We use the same process that we used for AI Summaries, which you can read about here. As a short recap:

  • Your data is processed in the EU and GDPR compliant

  • Your data will not be stored or used to train an AI

  • Your data will not be available to third parties or other orgs

Feel free to read more in our Privacy Policy.

How can I start using Ask AI?

Right now, Ask AI is available for all new customers and for most current customers. In theory Ask AI can be disabled or enabled upon request, so if you currently don't have it and want it, contact your CS manager or email help.lumoa@netigate.net.

To know if you have access to Ask AI, just go to any Impact page (the page after clicking on a Card) and see if the Ask AI tab is visible on the left side menu:

What kind of things can I Ask AI?

Technically, you can Ask AI anything, but we recommend you keep the subject related to your feedback. Note that Ask AI only works for the last 365 days of feedback though, so you couldn't effectively ask for your biggest issues "last year", as it wont capture the full calendar year.

Ask AI can recognize:

  • Time frames (such as "what were my biggest issues in Q1 or March of this year")

    • Ask AI cannot go outside the dates of your current card. Use the calendar tool to make sure the time frame you want to analyze is visible.

  • Filters, when formatted properly (such as "what are my biggest issues from Customer Group: B2B")

    • Filters should be formatted in the question as FilterName: FilterValue, like in Country: Germany or Gender: Female.

    • Ask AI can also see AI Filters if those are available on your data.

    • Note that this is in addition to any Filters that might already be applied on the card. If you are unsure, Ask AI what Filters are already applied to your data, and it should be able to tell you.

  • Topics, as seen in your Topic widget.

    • It can also list amounts associated with each Topic, and do a deep dive for certain issue within a Topic

  • Search Terms, such as things applied in the search field.

    • Ask AI may also "use" the search field to look for answers to questions, such as if you ask it "how many times do people mention the word 'cancel' in their feedback"

  • Previous questions

    • You can ask the AI to expand upon a point that it gave you in a previous answer, or ask for more details around a specific issue

  • Your unique business

    • The AI can understand your unique business, including the KPI's you might have internally as well as the industry, goals, or services you might be tracking.

    • Note that this info has to be provided to Netigate before the AI can use it. Please work with your CS manager or email help.lumoa@netigate.net to provide the AI with this information.

The AI can also create Tasks for you. Tasks allow you to track and get alerted of Feedback as it comes in, and then close the loop with your customers. We have a guide on Tasks in more detail, but the AI can create Tasks for you based on answers it provides. Simply select "create task" or ask it to "make that a task" as seen in the GIF below:

You might want to create a Task on:

  • Answers that mention very specific things, such as an error code or product line

    • The Task could then be assigned to a developer to investigate the error code, or a manager in charge of the product line

  • Answers that have comparisons to different time frames

    • You could inform your colleague that this month there was a much higher volume of "customer service" complaints than last month

Below is a list of example questions you could Ask AI. Feel free to modify them to fit your specific industry or business needs:

  • Out of all of our feedback, focus on the ones talking about our website. How should we improve our websites user flow? Answer with a list of tickets to a developer, and then create a Task

  • What issues do we need to resolve in order to achieve our goal of a 40 NPS score by the end of Q3?

  • How can we improve our login process from Device: Android users in Region: Germany?

  • What are the common issues users experience with our customer service when comparing this month to last month?

  • What are the sentiments expressed by users regarding our payment options? Summarize the feedback in the form of a poem.

  • What do users say about our most recent release? Respond with bullet points, and quote a customer feedback under each point.

  • Look for feedback around our onboarding, or people talking about being new customers. List 4 improvements we could make to our onboarding, and put them into the form of a customer journey.

  • Can you find instances where users mention problems with our employees attitude or professionalism?

Again, the above questions might not map over 100% to your specific data and industry. We recommend you tailor them to fit your specific needs, even if thats as simple as changing the word "website" in the first question to be something more related to you.

Can Ask AI quantify my data?

Ask AI can help you estimate, count, and compare feedback, but it should not be treated as a source of exact or definitive numbers.

How it works

Ask AI may use different methods depending on your question:

  • Text search → counts mentions of specific words or phrases

  • Topic data → uses pre-grouped feedback and associated volumes

  • Semantic understanding → finds feedback by meaning (does not provide exact counts)

Examples

  • “How many people have mentioned cold food?”

    • Uses text search to count mentions of specific terms

  • “What are my largest issues?”

    • Uses Topics and Insights to identify key themes

  • “How has X changed over time?”

    • Compares mentions or topics across date ranges or filters

Important to know

  • Counts based on text search or topics are generally the most reliable

  • Results based on meaning (semantic search) may not represent the full dataset

  • Small differences in wording (e.g. “claim” vs “claims department”) can change results

  • Ask AI is designed to analyze and summarize feedback, not replace reporting tools

👉 For precise numbers, we recommend verifying results directly in the platform (e.g. using search, filters, or topic views).

How does Ask AI analyze topics?

Ask AI can use your topic structure to understand what your feedback is about, but it does not always interpret topics the same way as the UI.

  • Ask AI is best at recognizing main (top-level) topics

  • More specific or lower-level topics may need to be explicitly referenced in your question

For best results, clearly indicate when you are referring to a topic:

Example:

  • “What are customers saying in topic: onboarding?”

  • “Compare topic: premium this month to last month”

👉 Topics are most useful when used for:

  • Summarizing feedback

  • Comparing trends over time

  • Understanding key drivers of sentiment

Important to know

  • Counts based on text search or topics are generally the most reliable

  • Results based on meaning (semantic search) may not represent the full dataset count

  • Small differences in wording (e.g. “claim” vs “claims department”) can affect results

  • Ask AI is designed to analyze and summarize feedback, not replace reporting tools

Ask AI will always try to answer your question, but for precise, report-level numbers, we recommend verifying results directly in the platform (e.g. using search, filters, or topic views).

How does Ask AI interpret your questions?

Ask AI relies on how your question is written to determine what to analyze.

For best results:

  • Be specific about what you’re looking for

  • Use structured formats when possible:

    • Filter: ValueCountry: Germany

    • topic: Xtopic: onboarding

  • Clearly describe:

    • timeframe

    • segment

    • comparison

👉 Ask AI does not automatically understand all context from your dashboard — your question helps guide it.

What to do if Ask AI has an issue?

While Ask AI cannot "lie" about what it sees in your data, it can sometimes throw an error message saying that it couldn't answer your question, or it may load for an extra long time.

While we are working to reduce the occurrence of these issues and others, the easiest trouble shooting steps are:

  • start a new chat

  • refresh your page

  • clear your cookies (CTRL + Shift + R)

  • restart your browser

Note that those steps are in order - so the first thing you should always do is start a new chat, and that will resolve most issues you may have with Ask AI.

Last tips

While using Ask AI, remember these helpful tips to achieve maximum Zen:

  • The model thinks while writing​

  • The best question is an accident​

  • Seek the unasked in your questions​

  • Invoke the unlikely expert​

  • Be annoyingly specific​

  • Find profundity in the nonsensical​

  • Celebrate the joy of tangents

Why are some parts of Ask AI's response in blue?

Ask AI can use our AI Highlighting feature to add some "truth" to its responses. Any text highlighted in blue should also have a number behind it, like "price (50)". What this means is that 50 people have mentioned the word "price", which is why Ask AI is including it in the response.

These highlights are there for you to verify that Ask AI is pulling from your real feedback when generating its response, and it will only mention things that your customers talk about.


Get in touch

📧 Do you have any questions or comments about using Lumoa? Don’t hesitate to email us at help@lumoa.me.

Did this answer your question?