Launched this week, Verity Score records what six AI assistants say about a Shopify store and preserves each answer with its sources. The app drafts corrections from a merchant’s text, publishes approved fixes to Shopify, and tracks visits and orders attributed to AI.
V7 converts company files into a Context Graph that agents can query, using GPT-5.6 models and GPT-6 Astra for harder queries. V7 reports accuracy gains, lower costs, and faster, auditable workflows across finance, insurance and asset management.
Authors Arthur Zucker, Simon Brandeis, Luc Georges and Lysandre report that tokenizers v1 preserves v0.23 output while improving encode speed by 3–30× on an Apple M4 Max and scaling at 76% across eight workers. Changes include bitcannon splitting, a thread-local word cache and a no-alloc merge loop.
Hex says GPT‑6 Astra helps its data agents turn analysis into interactive, shareable visualizations by handling complex transformations and checking that results match business questions. Caitlin Colgrove, Hex co‑founder and CTO, attributes improved analytical judgment to the model.
Context.dev launched Answers by Context.dev, an API that combines web scraping, enrichment and LLM-style research into one response. The company says the service returns structured JSON with source URLs and offers SDKs for TypeScript, Python and Ruby.
Zella records screen or camera and performs on-device editing on Mac and iPhone. The free tier offers 1080p output with no watermark and no account; Pro is a one-time $89 payment that unlocks both devices. Auto-Polish and Make it Viral automate edits and pacing.
Profound said on September 15, 2026 that it raised $180 million in a Series D at a $1.8 billion valuation, under seven months after a $96 million Series C. The company, part of the GEO/AEO wave, says revenue tripled in six months and it serves over 1,000 enterprise customers.
OpenAI announced on September 16, 2026 new AI capabilities for ChatGPT Ads, including Sponsored Agents that let users start business-run conversations after clicking an ad. The update also adds AI-driven creative tools in Ads Manager and integrations with HubSpot and Shopify.
IBM researchers found that average accuracy can hide run-to-run variability: a ReAct agent using GPT-4.1 posted Mean@5 77.4% but Pass^5 only 53.0% — a 24.4-point gap. Their Consistency Analyzer and generated guidelines cut the gap to 12.0 points without lowering mean accuracy.
Fyxer combined OpenAI models with more than 500,000 hours of annotated executive assistant workflows and 30–50 specialized models to draft replies in each user’s voice. The system reports 53% of AI drafts are accepted and over 90% user retention at 90 days.