Starting September 1, 2026, Alchemer Pulse is upgrading the AI model that reads and tags your open-text feedback. This upgrade applies automatically to all Pulse customers at no additional cost, and it requires no action on your part.
Because the new model reads feedback more accurately, some of the metrics you track in Pulse, including net sentiment, negativity index, and top themes, may shift when the new model goes live. This is expected, and it's a good thing. Your customers' experience hasn't changed; Pulse's read of that experience just got sharper.
When the Upgrade Takes Effect
The new model goes live on September 1, 2026. From that day forward, all new feedback is tagged by the next-generation model automatically. Your historical data stays as-is unless you ask us to reprocess it, which you can do anytime after September 1.
What's Changing
- The AI model powering Pulse. We're upgrading to a next-generation model built specifically for this kind of open-text analysis and improved accuracy.
- How much of your feedback gets tagged. The new model captures more of your open-text responses, so fewer land untagged or in a catch-all "Other/Miscellaneous" theme.
- How closely tags match each comment. Themes and sentiment line up more precisely with what each customer actually wrote.
- Your metrics, as a result. Because net sentiment, negativity index, and similar metrics are calculated from the underlying tags, they may shift once the new model is live. Your top themes may also reorder as previously under-tagged themes become more visible.
How much any of this moves is specific to your account and your data, so we can't predict an exact size or direction ahead of time. What we can tell you is that any shift reflects better coverage and accuracy, not a change in your customers' actual behavior or experience.
What's Not Changing
- How you use Pulse. There is no change to the product functionality and no action required.
- Your themes and theme structure. Your existing themes stay exactly as they are. The new model tags feedback against them, just more accurately.
- Your historical data. Past feedback and the tags already applied to it stay as they were, unless you request reprocessing through your Customer Success Manager.
- Your customers' experience. Your customers' experience hasn't changed. Pulse's read of that experience just got sharper.
How the New Model Reads Feedback More Accurately
The next-generation model was built using state-of-the-art architecture specifically for the kind of analysis Pulse does: reading an open-text comment, identifying the different themes, and tagging the nuanced sentiment attached to each one.
The new model reads feedback more the way a trained human analyst would. In validation testing, it agreed with expert human tagging more often than the previous model, across both precision (how correct the tags are) and recall (how much feedback gets tagged at all).
Before and After: What Sharper Tagging Looks Like
The example below shows the same customer comment tagged by the previous model and by the next-generation model. This customer praised tech support and ease of use. The next-generation model tags it to Technical Support and Documentation and User Interface and Navigation, mapping each point to the specific area the customer named. The result is tagging that lands closer to the customer's actual words.
How to Compare the Change on Your Own Data
In Pulse Reports, you can compare two time periods side by side. Set one period to before September 1, 2026 and the other to after, and you'll be able to see how themes, sentiment, and scores differ between the old and new model.
How to Explain a Metric Shift to Your Stakeholders
If you report these metrics to your team or leadership, here's a short version you can pass along: "Our feedback analysis platform upgraded to a more accurate AI model on September 1. Some of our sentiment and theme metrics shifted as a result. This reflects a sharper read of the same customer feedback, not a change in how our customers feel or behave."
What Happens to Your Historical Data
The new model applies to new incoming feedback only. Your historical data and the tags already applied to it are not automatically changed, so your past numbers stay exactly as they were.
One side effect of that: if you look at a trend line that spans the cutover date, you may see a visible step where the new model takes over, since data before and after that point was tagged by two different model generations on a slightly different basis.
If you'd rather have your entire history on one consistent basis, you can request that we reprocess your historical data with the new model. Reach out to your Customer Success Manager to submit that request.
Frequently Asked Questions
Do I need to do anything?
No. The upgrade is automatic. There is nothing you need to set up or turn on.
Will this change how I use Pulse day to day?
No. Nothing about the Pulse interface, workflow, or how you review feedback is changing. Only the tagging model underneath it is being upgraded.
How much will my metrics move?
It varies by account, since every organization's feedback data is different, so we can't give a specific prediction. Any change you see reflects Pulse capturing your customer voice more completely and accurately, not a change in your customers' underlying sentiment.
Will my historical trends still line up with new data?
Your history isn't reprocessed automatically, so old numbers won't change. That does mean the point where the new model takes over may show as a step in an otherwise continuous trend line. If you want your full history on a single consistent basis, ask your Customer Success Manager about reprocessing your historical data with the new model.
Can I opt out of the upgrade or delay it?
The upgrade applies automatically to all Pulse accounts so that every customer benefits from more accurate tagging. If you have a specific reporting-period concern, talk to your Customer Success Manager about timing your historical reprocessing.
