17 November, 2025

Scorecards

You can now add weights to the criteria selected in each scorecard. Weights are simply used as multiplier when calculating the final pass rate of evaluations, i.e. 0-100%. This allows you to:

• Re-use the same criteria among multiple scorecards with different impact on scoring, without having the duplicate criteria and only adjust the scores.

• Use "Critical" criteria which can fail the entire evaluation and, at the same time, do not count toward the final score if passed succesfully. This allows you to implement a list of critical checks with no scoring impact.

Tickets

We have a new "Customer details" section which includes the basic customer details, customer notes (if enabled) and all the selected customer fields.

Training

Fixed repetitive AI customer response in case of multiple back to back messages sent by the learner.

You can now also simply send a message using command/ctrl+enter.

AutoQA

Improve AutoQA KB retrieval to allow fo faster response when 100's of articles are linked.

Iterated on AutoQA's comment output to make them less repetitive.

Workloads

When you update a workload, you will be notified about which changes are effective immediately and which ones will be effective from the next cycle.

We've also added a new option to randomly select, as reviewee, any agent who participated in the ticket. This allows you to evaluate all your agents, even if their participation in the tickets was minimal.

Data ingestion

We now ingest customer, i.e. contact, notes from Intercom. You can include them in the UI and pass to the AutoQA by enabling the following option in Settings > Workflows.