Excellent! this is awesome!
Just need to work on Text IQ now please :-)

Ditto on the TextiQ.  We need better ability to see what verbatim is added or removed when you make a change to an existing topic.  This has been requested by many people and is mentioned on many evolve tickets.

Would be keen to know what is happening behind the scenes on that.

Thanks

Rod Pestell

Great addition!!

Hello,

I noticed here and elsewhere on the Community that there’s a lot of interest regarding TextIQ. In addition to ​@Rod_Pestell comments, ​@KiwiChaps & ​@c.rose are there specific questions or challenges you have regarding TextIQ?

As we develop educational content and gather input for our product team, it would be helpful if you could briefly share any challenges you're experiencing with TextIQ or areas where you see room for improvement.

While I can't guarantee implementation of specific ideas, your feedback helps inform our ongoing development efforts.

​@AlonsoC - Sure thing - I have some ideas!

I think a big pain point for us is that TextIQ auto recommended topics are incredibly broad, and therefore, not very useful. I would love to be able to lean on the tool more for generative recommendations gather than the manual process of creating the queries. Today, I’ve accepted that TextIQ is manual and only as good as we make it, which I feel like is a huge missed opportunity especially in the age of AI.

For example, we are a B2C ecommerce company and some of our recommended topics include:

• “Website”
• “Items”
• “�Our Company’s Name]”
• “Customer”

These recommendations are unhelpful at best and take up meaningful screen space at worst.

I also echo the feedback on Evolve that there needs to be a better way to exclude certain one word comments from TextIQ - for example, sometimes customers only say “no” in our comments and we have no good way to filter for that circumstance and exclude it from analysis besides manually bucketing it.

If your teams would like to loop me into any UX or product research on TextIQ, feel free to shoot me a message. 

Hi ​@AlonsoC,

TextiQ as it stands is ok for setting up brand new topics but when it comes to appending and enhancing, you can’t see what it is adding.  All you get is x responses was added or removed…. you are left thinking “what are they?”, “are they the right ones?”, “has my change in selection been too narrow or too wide?”.  Please make it so that we can review the latest additions or subtractions and you’ll make the UI much more usuable.

Re generative AI - I have experimented with AI engines and asked them to review a selection of comments and have asked it to give me the top 20 topics with % breakdown.  The AI engine gives me almost instantly an accurate summary of the top topics.  This helps me then to create the topics in Qualtrics TextiQ and then it’s down to me to work out the syntax for the topics.    The whitelist of topics as referred to by ​@c.rose are too generic for many industries and I might also add that when you use them the syntax is often very brief and limited meaning that it won’t capture half as many comments you think it might.

Hope this helps and word spreads on this to raise awareness further with a view to help improve it. 

Thanks

Rod Pestell

Thank you ​@c.rose & ​@Rod_Pestell, I appreciate your feedback and thoughtful comments! I’m still in the phase of gathering feedback and challenges and will get back to you if I need further info.

So happy to see some comments form other users and realising I am not alone! I hesitate to reply, I begin typing and I am concerned about omitting important points. But as top of mind see my feedback on TextIQ:
 

• Set up Text IQ.
  • I agree with ​@c.rose’s answers. The recommended topics are too broad and not particularly useful, so I know the setup process is manual, which requires several hours/days.
  • Sentiment Anarlyis (red and green) are really 50/50, can’t really trust, specially if the comments are long - it becomes really random results
  • And there is currently no option to make bulk changes in sentiment analysis - so need to edit comments one per one.
  • ​​​​​​Currently unable to filter comments when creating topics (for example, by detractors or over the last 30 days) would be useful, especially for editing sentiment analysis with a bulk option.
  • It is challenging to handle cases where a customer comment like “Competitor X is doing better on Z” results in Z being highlighted positively in the analysis, although it should be negative since it refers to a competitor’s strength.
  • Tagging comments that consist solely of a specific phrase—such as 'N/A' or 'Not sure'—is currently unsupported. Ideally, TextIQ would allow the use of regex commands to enable this functionality (same as ​@c.rose )
  • Improved UX during topic editing would be helpful; currently, the tool shows only the number of comments added or removed, rather than specifying which comments are affected. Seeing the actual comments would help users confirm their changes. (Same as ​@Rod_Pestell)
     
• More than the set up is also the visualisation for users in CX dashboard:
  • Table IQ does not support percentage calculations or dynamic time-based comparisons.
  • It would be beneficial if the response ticker could highlight text based on filter bubbles, similar to the view in Text IQ, enabling users to quickly see the relevant topic within a comment.
  • Filtering topics by sentiment (e.g., filtering claims labelled as negative) is not currently possible.
  • Clicking on a topic bubble navigates directly to its subtopics, which is great, but drilling down from a parent topic on a bar graph displays all subtopics rather than aggregating under the parent topic, which can be confusing.
    
    Integration of Qualtrics Assist within the CX dashboard is promising and may make TextIQ irrelevant? Insight Explorer feels somewhat isolated and lacks flexibility.