July 24, 2026

Clearer Ask Layr citations, smoother loading and actions, more reliable Slack sync, and a landing product demo video.

Product

Ask Layr answers now use numbered citation refs ([1], [2]) that match the supporting evidence list, instead of raw machine IDs.
The Ask supporting-evidence panel is denser for long citation lists: filter by source when needed, page through results, and jump from an answer marker to the matching evidence row.
High-traffic product surfaces use layout-preserving skeleton loading so waits are easier to tell from empty states.
Primary mutation actions show clearer pending, success, and error feedback while work is in flight.
Review inbox, opportunity comments, and integration source scopes use server pagination so large lists stay browsable.
Opportunity detail no longer crashes when evidence dates serialize across the page boundary.
Alert action buttons stay aligned under the alert content instead of overlapping the border.

Slack sync is more resilient under rate limits: feedback flushes per channel, partial history can complete instead of stalling empty, and bot-authored text messages are ingested.
Slack Ask answers use the same numbered citation refs as in-app Ask.
Jira OAuth webhook deliveries authenticate more securely so connector events stay trustworthy.
Image evidence ingestion retries transient vision network failures instead of dropping the fragment.
Text evidence redacts common PII patterns (email, phone, payment card) before storage and embedding.