Chat for Blackbaud AI—Share Your Experiences!

GettyImages-2205342212.jpg

Over the next four weeks, we will be creating a dedicated space in the community to talk about Chat for Blackbaud AI: how it’s working for you, and any solutions and challenges you’ve encountered.

Whether you’re just getting started or already experimenting with prompts, your experiences—what helped, what didn’t, and what you’d try again—play a critical role in shaping what comes next.

You’ll notice featured discussions at the top of the Raiser’s Edge NXT discussion page over the next four weeks—and yes, there are badges involved.

The first discussion is now up! We want to know, “What task did Chat for Blackbaud AI help you with this week?”

If you participate in the featured discussions, you will earn this badge for your profile page:

badge.png

New Community Feature 

You also will notice that the discussions are tagged with “Blackbaud AI.” That’s now a tag you can follow! Just click on the tag cloud on the discussions page, click “BlackbaudAI" and then “Follow.” 

How to Share Feedback (and Why It Matters) 

When Chat for Blackbaud AI gives you a response, you’ll see 👍 and 👎 options directly in the chat experience. That’s the best place to let us know what worked or what missed the mark. 

Here’s why this matters: 

  • Thumbs‑up and thumbs‑down feedback goes directly to the teams improving the chat experience 
  • A short note about why something worked—or didn’t—helps us make improvements faster 
  • Community discussions are for sharing experiences, not troubleshooting prompt errors 

Think of it this way: 

Community = conversation and learning 
Thumbs 👍 / 👎 = product improvement 

Both matter—and they work best together.  

Learning From Super Users 

Throughout the month, you’ll also see posts from Blackbaud Community All‑Stars who are already piloting Chat for Blackbaud AI. You can respond to their discussions and earn the badge, too! 

If you’ve tried Chat for Blackbaud AI—even once—you have something valuable to share. We can’t wait to learn from you! 

And if you’re brand new, these resources will help you get started:  

You can also see what’s possible when AI meets real fundraising work in our upcoming webinar.  

➡️ Register now to save your spot!  

Comments

  • Karen Diener
    Karen Diener Community All-Star
    Tenth Anniversary Kudos 5 2026 Spring PUB: Raiser's Edge NXT Premium and AI First Reply

    I may follow up with some of the other discussions, but my early testing of AI chat was not positive.

    A new RE client was asking questions, so I was showing them how it works. We used prompts from the prompt library, and the AI responses were wrong. It would say something like "you have not received gifts from this donor in five years" as well as "the last gift you received from this donor was in 2025" in the same response. When we switched to a different constituent record, the AI chat did not update. It still kept referring to the previous constituent. Needless to say, they were not impressed.

    I expected that it would at least be trained so that prompts from the library would work.

  • I agree with @Karen Diener. Unfortunately on the couple of occasions I have activated AI chat to test it myself it did not answer anything correctly or gave contradictory answers, so I have turned it off and will not be sharing it with our team.

    Either way, I would much prefer to have improved functionality in the system (e.g., acknowledgement letters) than an AI chat bot, even if it could answer things correctly!

  • Thanks for this feedback! I've passed it along to my team and also want to check to see if you provided the thumbs down feedback to the chat along with an explanation on what was wrong. Thanks, Karen!

  • Karen Diener
    Karen Diener Community All-Star
    Tenth Anniversary Kudos 5 2026 Spring PUB: Raiser's Edge NXT Premium and AI First Reply

    Thanks Crystal. I do believe I gave it the thumbs down and explained why, but I was online with a new RENXT customer so may not have taken their time to explain the flaws. Again - I feel like we should not get completely incorrect answers when using a prompt from Blackbaud's library.

Categories