Sola, our AI assistant for people and team insights, has a major round of upgrades. This release improves how Sola understands your company data (HRIS), custom competencies, personalizes its answers, and stays configurable from Settings.

Full Use of Your Company Data
Sola can now use your company data fields directly when answering multi-part questions. Alongside this, we built a dedicated flow so admins can add, edit, or AI-generate context for their own custom HRIS and competency columns, so Sola knows how each one should be used.

More Ways to Search and Filter
Sola's user search now supports combining multiple filter types in a single query — such as Core Drivers, Core Values, name, and role match — along with additional columns like Core Reasoning scores and HRIS data.

Terminology & Preferences
Admins can override how Sola refers to common concepts by defining up to 10 word pairs — for example, telling Sola to say "clients" instead of "customers," or "workspace" instead of "project." This lets Sola's language match the terms your organization actually uses, rather than defaulting to generic phrasing.

Sensitive Topics to Avoid
Admins can specify topics Sola should avoid entirely (e.g., politics, religion, or legal advice), with a character limit of up to 500.

Competencies Context
Just like Company Data fields, each competency can have AI-generated or custom context describing how Sola should interpret and talk about it.

Sola Now Speaks Directly to You
Sola previously answered in the third person even when you asked about your own results. Sola now recognizes who's logged in and responds directly — "Your logical reasoning score is..." instead of "Adam's score is...".

Continuing a Conversation After Stopping It
Previously, if you stopped a Sola response mid-generation and then tried to send a new message in the same thread, you could get stuck on "This conversation is still being processed." This has been resolved so you can pick the conversation back up right away.

"Network Error" Message Resolved
Sola's UI occasionally showed a "Network error" even when the response was still streaming and completed successfully behind the scenes. This false alarm has been fixed.