A fitting close to the quarter: agents moving deeper into operations, and the model layer reshuffling again.
TELUS scales ElevenLabs voice agents into support
On June 22, TELUS Digital became a preferred implementation partner for ElevenAgents (ElevenLabs' voice-agent platform), deploying voice AI alongside frontline customer-care teams and integrating across Genesys, Twilio, Amazon Connect, Zendesk, and Salesforce. A cited proof point: a TELUS Communications pilot cut 30-day home-internet cancellations by more than half. Voice agents are now a serious part of the contact-center stack.
Betting on agents that learn from games
On June 25, General Intuition raised $320M at a $2.3B valuation to train agentic models on video-game action data — an unusual bet on where general-purpose agent capability might come from next.
Major model news: GPT-5.6 and Mythos 5's partial return
OpenAI previewed the GPT-5.6 family on June 26 — Sol (flagship reasoning/agentic), Terra (balanced, ~2× cheaper than GPT-5.5), and Luna (fast) — with a new "ultra" mode that uses subagents. And on June 27, the US Commerce Department authorized Anthropic to redeploy Mythos 5 to roughly 100 US organizations (federal agencies and critical-infrastructure firms), a partial off-ramp from the month's export-control saga, with Fable 5 still restricted for general use.
What it means if you're picking tools
Voice + support is one of the clearest agent ROI stories right now, and the model tiers underneath keep multiplying — match the tier to the job. See voice & audio and customer support.
Sources: TELUS × ElevenLabs (PR Newswire) · General Intuition $320M (TechCrunch) · GPT-5.6 preview (Axios) · Mythos 5 restoration (Euronews)