AI vs Community: How Are You Getting Qualtrics Help?

​@Chee Heng_SZ, ​@AdamK12, ​@vgayraud, ​@pamelalbeck. ​@rochak_khandelwal, ​@Sachin Nandikol, ​@ahmedA as experienced members who frequently engage on the community, I'd be curious to hear your thoughts on how AI is impacting your Qualtrics problem-solving workflow.

That’s a great question ​@AlonsoC , especially as LLMs continue to develop. To be honest in my experience it’s two fundamentally different use cases. Being language-based, the AI models that we use are really good at synthesizing information, categorizing responses or other database elements, distilling a bigger document into the main idea, or even generating text. But to answer a specific question about best practices or insights, I still go to people. I hope that’s helpful!

It’s still absolutely essential to keep the human in the loop in a way that adds value, and while LLMs can make it easier to do tedious tasks, or tasks that humans do inefficiently, better, there’s still an important role for human beings with actual experience.

This perspective is definitely helpful! Thanks for sharing!

Good question! Honestly, most of my AI use isn’t directly Qualtrics-related. I use it more for general work tasks like: drafting emails, writing meeting notes, creating mock-ups or diagrams, putting together training materials or quick cheat sheets, etc.

That said, I also lean on AI as a coding assistant for the solutions I develop in Qualtrics. For example: JavaScript tweaks for surveys, workflows custom code tasks, API calls structure and formatting.

It saves me a good amount of time when I’m documenting things like implementation plans or handover guides.

The support chatbot AI is more hit-or-miss. It rarely solves the problem entirely for me, but it can give a decent validation checklist, especially useful if one already has a good grasp of the product and just want to double-check they didn’t miss something obvious.

On the flip side, I’ve also seen AI waste time. Usually when people unfamiliar with Qualtrics take its output at face value. One example: a stakeholder used AI to "solve" a CX Dashboard request I had already told him wasn’t possible. What I got back was a two-page-long Frankenstein solution involving CX and EX Dashboards, Discover/Clarabridge, plain Excel formulas, and Stats iQ. None of which could actually work together in the way it claimed.

As for the Community, that’s still my go-to for a bunch of things:

• Weird, platform-specific bugs or behavior, especially when something just stops working mysteriously.
• New or recently released features. AI isn’t usually trained on the latest quirks or changes.
• Super niche or custom use cases. Community members have usually "been there, done that" and have insights you won’t get anywhere else.

It’s really about using the strengths of each tool: AI for speed and structure, Community for depth and experience.

It’s an important question ​@AlonsoC. I am an AI first guy and AI tools are my first go to place for any problems I face. Like ​@AdamK12 pointed out, they are good at information systhesis: so if the similar problem has been asked earlier or pertains to the documents / blogs that have already been put up online, then they are good. As a safe practice, I have lately started asking AI tools for the evidence behind piece of info they provide.
For issues that are highly contexual: community is still the best place. I also feel that community becomes all the way more important considering it is one of feeding grounds for AI tools.

I agree with ​@rochak_khandelwal here. There’s a balance to be navigated between harnessing AI to surface insight from problems that have been addressed before, and using the experience that exists in any community to solve new problems.

​@rochak_khandelwal , there’s a lot here that resonates with me. In my experience, I used to lean entirely on the Adobe Community when learning Premiere Pro to develop 2-Minute Tips but have increasingly leaned on AI due to the quick and detailed answers it surfaces. Yet for highly contextual problems the Community is still the best resource.

And your safe practice of asking AI for evidence behind their answers is something I will definitely be using moving forward!

Totally agree! In a past BOM topic it was great to read how members appreciate the unique perspectives and clever workarounds to product issues that simply don’t exist in official documentation. Many solutions offered in this Community come down to the creativity of our members.

I agree that each tool has its purpose and strength! I do think you need at least some basic knowledge of a platform (like Qualtrics) before leaning entirely on AI to solve issues. Once you have that working understanding, AI serves as a great assistant, as you mentioned. But there's still that "you don't know what you don't know" problem, and bringing your questions to the Community can help identify blind spots that AI just can't catch.

This is so true--LLMs are based on logic and patterns, but don’t have any experience based on the specific topic they’re being asked about, so kinds of questions like solving for a dashboard issue aren’t something that it handles well. But ask it to write a 1.5 page summary on the causes of World War One and it will do decently well, without ever mentioning an ostritch.

I second ​@vgayraud, this sounds a lot like my situation:

Honestly, most of my AI use isn’t directly Qualtrics-related. I use it more for general work tasks like: drafting emails, writing meeting notes, creating mock-ups or diagrams, putting together training materials or quick cheat sheets, etc.  

The support chatbot AI is more hit-or-miss. It rarely solves the problem entirely for me, but it can give a decent validation checklist, especially useful if one already has a good grasp of the product and just want to double-check they didn’t miss something obvious.

As for the Community, that’s still my go-to for a bunch of things:

• Weird, platform-specific bugs or behavior, especially when something just stops working mysteriously.
• New or recently released features. AI isn’t usually trained on the latest quirks or changes.
• Super niche or custom use cases. Community members have usually "been there, done that" and have insights you won’t get anywhere else.

It’s really about using the strengths of each tool: AI for speed and structure, Community for depth and experience.

I will Google my Qualtrics questions and glance through the AI Overview before drilling down on the actual search results.  I haven’t noticed any errors in the AI Overviews.  I use Google for quick reference - There was a sperate discussion on the Community boards of wanting a bookmark feature and it can be a little tedious to search the boards.

I do keep an eye on all of communication/trainings from Qualtrics and the Community for AI topics/trends.

I have been using AI for support since the day it was launched. Initially, I had no issues with the suggestions it provided. However, I have noticed some instances where the AI could improve. For example, sometimes it suggests enabling or disabling an option, but when I check, that option is not available in the current version of the platform, it was from an older version. There’s definitely room for improvement, but until then, I feel the community remains the safest place for us to rely on.

Hi Sachin,

Thanks for this info! When you say ‘AI’ what specific platform(s) are you referring to? ChatGPT, Qualtrics Support/Community AI Chatbot, etc...

Thanks for this feedback, Pamela! Completely agree that AI is great at quickly surfacing relevant info!

One time I was stuck on a Qualtrics automation and needed some custom JavaScript. I went to AI first because I wanted quick, step-by-step help and something I could test right away. Once I had it working, I still wanted to know if others had faced the same thing in real projects, so I checked the Community. People there shared real-world tips and “gotchas” that AI couldn’t cover from experience.

GPT & Qualtrics Support

Thanks for sharing ​@gPandey_715! The AI-first-then-community approach seems to be the common theme here.  

I would go to AI (ChatGPT) for coding related solution when the scenario below occurs:

1) I don’t recall seeing a similar community thread topic

2) I don’t recall the title of the community thread

3) I need something quick for a minor section of a survey project like hiding question, greying out text boxes, etc. 

If it is just common Qualtrics features or how to structure the presentation of the survey, the community and the support documentation are preferred because AI tools could give wrong suggestions or different suggestions despite having the same question. Some users in the community provide screenshots for illustrations and highlight the factors involved during a troubleshooting. This helps other users understand the Qualtrics features better.

Sometimes past coding threads solution doesn’t really work due to certain functions being deprecated or new functions/versions are available. Having the coding experts in the community helps a lot in understanding more about the newer solution while AI tool generally generates a sample tedious code that will require a number of testing and editing before it can be used within Qualtrics.

Additionally, although users can post questions for other users to reply/answer if there are no existing past threads available, it is uncertain if there will be a response/reply/answer and by which date/time. It might be better to continue to use AI to do trial and errors after posting the question instead of just waiting endlessly for replies.

Personally, I feel that Community responses (for coding) are more accurate/efficient if they are within a certain timeframe. Once it goes beyond a certain timeframe, perhaps AI may provide a slightly more accurate response despite potentially requiring some minor tweaks on the user end.

Thank you to everyone who shared your experiences, wins, and challenges using AI and our Community! Your thoughts revealed a clear pattern: AI is primarily used for quick answers and rapid problem-solving while our Community serves as your go-to source for in-depth expert knowledge and nuanced discussions.

We'll use these insights to continue optimizing our Support and Community AI Chatbot and work with Gainsight to find new ways to quickly surface the answers that help you get your work done more efficiently. We'd love your continued feedback as we make these and others improvements to our Community!

This has been an interesting chat indeed--it dovetails with a podcast episode on the “AI in Business” podcast that talked about how the prime use cases for monetizing and adding value with AI seem to be on making “5-minute tasks” easier, faster, less tedious, etc. -- and reserving higher-order projects and activities for humans.

I’ve been using the AI with support since it launched and have really enjoyed it. It is really helpful for quick questions I have that I am in early exploration with but am not quite ready to make a full post for or reach out to my account team over.

Thanks for this detailed feedback ​@Chee Heng_SZ! Your comment highlights several important factors: 1) The value of visuals (screenshots) and sharing what was already tried to come to a solution 2) Your second point seems to be similar to what ​@pamelalbeck brought up (and this thread) about the usefulness of a bookmark feature on the Community! 3) Updating past posts with up-to-date info 4) Regarding your final paragraph, can you elaborate on why “AI may provide a slightly more accurate response” after a certain timeframe? What timeframe have you found where community responses shift from being more accurate to potentially less accurate than AI?

Appreciate any insight!

I’ll check out the podcast!

Hi ​@c.rose, thanks for your comment! What makes a question feel 'ready' for a full community post versus keeping it at the AI level?

Hi ​@AlonsoC ,

Feature/function being mentioned within community in the past may work at that point in time.

However, a feature/function may become obsolete/deprecated as time passes.

For example, a solution thread from 5 years ago (which likely is from another forum-like platform before it was shifted over to this current community platform) may not work or encounter unexpected missing section or that users wouldn’t be able to locate the feature with the new layout.

Thus, when comparing this with AI tool (still referring to ChatGPT), the AI tool may suggest a more updated “solution”, even if the “solution” may not be correct and requires some amendments.