Skip to content
Operon AI
Changelog

Everything we shipped

Published every other week, including the unglamorous work. If something we changed affects your configuration, your deployment lead tells you before it lands.

July 22, 2026v4.9Feature

Multi-location capacity balancing

  • Scheduling Agent can now shift a booking to a sister location when the requested one is full, with configurable distance and revenue thresholds.
  • New per-location override editor in the Brain for hours, pricing, and escalation policy.
  • Capacity heatmap added to the Analytics workspace.
July 8, 2026v4.8Feature

Spanish, Portuguese, and Vietnamese voice paths

  • Three new production voice languages with vertical-tuned vocabularies.
  • Automatic language detection within the first 1.2 seconds of a call.
  • Mid-call language switching without dropping context.
June 24, 2026v4.7Security

Zero-retention model tier

  • Enterprise plans can now route inference through a zero-retention path where no prompt or completion is persisted by any model provider.
  • Per-field redaction policy for PHI and PCI data in transcripts.
  • Audit log export to Snowflake and BigQuery.
June 10, 2026v4.6Feature

Revenue Agent general availability

  • Outbound reactivation campaigns against dormant customer lists.
  • Automatic quote follow-up on a configurable cadence with a hard stop rule.
  • Full TCPA guardrails including quiet hours, consent checks, and per-state frequency caps.
May 27, 2026v4.5Performance

Latency: median first-token down to 380ms

  • Speculative synthesis now covers 94% of turns, up from 61%.
  • Regional inference in us-east, us-west, and eu-central.
  • Cold-start elimination on the retrieval path.
May 13, 2026v4.4Integration

Salesforce and HubSpot bidirectional sync

  • Two-way field mapping with conflict resolution rules.
  • Call recordings and structured summaries written to the activity timeline.
  • Custom object support.