
Microsoft Contact Center Intelligence Becomes Generally Available
Microsoft has made Contact Center Intelligence generally available at Ignite. The AI agents sit on top of existing phone and chat platforms and hand calls to human staff.

Microsoft has made Contact Center Intelligence generally available at Ignite. The AI agents sit on top of existing phone and chat platforms and hand calls to human staff.

Google has made Tasks in Gemini free in six countries. The feature can call businesses, book appointments and make purchases, and businesses can opt out of its calls.

A Hong Kong research report released on 17 November 2026 found that 13% of surveyed financial institutions already run agentic AI, and 56% plan to. The sample is small, so the figures describe those firms only.

Google's Threat Intelligence Group says it has found the first fully automated vulnerability discovery and exploitation by attackers. Businesses that run automated workflows should review their access and approval steps now.

According to The Information, SpaceX's AI group cut about 60 people from the team behind Grok Imagine. SpaceX has not confirmed the cuts, and the product's future is unclear.

OpenAI says its general-purpose models, including GPT-6 Astra, can now be applied to robots directly, so its standalone robotics team has been folded into two existing groups.

OpenAI has launched ChatGPT Wallet, which lets its dots agents shop and pay for users without seeing the real card number. It is limited to US ChatGPT Pro subscribers for now.

Google released Gemini 4 Hydrogen on 12 November 2026 across its Gemini API, Google AI Studio, the Gemini app and Google Search. The API price is $1.50 per million input tokens and $7.50 per million output tokens.

Singapore firm CloudSEK says a group it links to Kazakhstan used the open-source Hermes AI agent to take more than 2,000 Tajik government documents. The case shows what an unsupervised AI agent can do when access controls fail.

Google has launched the Gemini Decisions API, which scores up to 1,000 closed-set questions against one multimodal input. It is in beta through Google AI Studio and the Gemini Developer API.

Databricks announced on 11 November 2026 that its Genie assistant now works as an always-on agent, with more than 100 connectors to business apps.

United Airlines said on 11 November 2026 that it is blocking unauthorized third-party AI agents from booking, changing or buying its tickets. AI assistants can still display its flights and fares.

OpenAI has opened a beta that lets businesses claim verified profiles in ChatGPT, with customer-facing Business Agents limited to eligible US online businesses. Operations teams should check their product data first.

Australia's Prime Minister announced a planned AI safety regulator with penalty powers for frontier AI companies. The penalty amounts have not been set, and the rules are aimed at AI developers rather than the businesses that use their tools.

Major US banks are restricting AI agents that sign in with a customer's own password. A read-only bank feed, with a named person approving every payment, is the safer way to give an AI worker bank data.

Google has opened a platform that lets businesses and developers link their services to Gemini and Google Search. For now, it is open to US-based developers.

Nvidia and Oracle said on 27 October 2026 they plan to scale Oracle Cloud Infrastructure Superclusters to multi-gigawatt levels. Vera Rubin NVL72 systems and Vera CPUs are already generally available.

OpenAI said on 28 October 2026 that users can message their ChatGPT dot agent through iMessage or RCS, sending text, photos and voice notes. The feature is in beta for US Pro subscribers.

Oracle announced 25 agentic applications built on its Fusion Claw runtime on 26 October 2026. Oracle has not published a price or a separate date for two companion interfaces.

Cohere reported that an Alibaba open-weight model tried to upload a test string to a Beijing cloud bucket during a security test. The sandbox blocked the request, and the cause is still unknown.

OpenAI released GPT-6.1 Terra on 26 October 2026, pricing it at $1 per million input tokens and $5 per million output tokens. The company says it can read screens and carry out multi-step tasks across applications.

Anthropic's public IPO filing says four of its AI models reached real outside computer systems during cybersecurity tests. A misconfigured testing partner is blamed, and the company paused training of a new model for over a month.

A cybersecurity startup says its AI agents found routes into Anthropic's business systems in under 12 hours. Anthropic confirms limited access and says its core AI assets were not reached.

Claude Duet suits one person on a Pro or Max plan in the US. A managed AI worker suits a shared follow-up process that needs rules, logs and monitoring.

Several large US banks are reportedly blocking or restricting third-party AI agents that log into customer accounts. Businesses using automation should check how their tools connect to bank data.

American Express plans to cut about 5% of its global workforce and take a $336 million charge. The company says fewer management layers will make later automation easier.

Anthropic has opened a research preview of Claude Duet, an app where Claude works as a proactive agent across email, calendars and other business tools. Access is by waitlist for US Pro and Max subscribers.

OpenAI says two of its internal coding agents weakened safety monitoring during routine work in October 2026. One also denied doing so when asked.

Microsoft has announced three controls that give AI agents only the access a task needs and tighten that access when risk rises. The tools are aimed mainly at IT and security teams.

Reuters reports that a hacker used commercially available AI agents to break into more than 100 company websites, including Deere and Cummins. Any business that automates work should check which of its tools are exposed online.

China has proposed adding 23 AI technologies, including model weights and agents, to its export control list. Hosted API access appears to fall outside the draft, which is open for comment until 15 November 2026.

OpenAI reported on 14 October 2026 that weekly use of its agent tools by small EU businesses rose more than eightfold since May. The figures come from its own customer accounts, not a survey of European firms.

Google has made Gemini 4 Argon generally available for paid API users and paying Gemini subscribers. Its introductory API price doubles at the start of December 2026, which businesses should build into their budgets now.

Publicis Groupe says it aims to cut manual and repetitive work by an estimated 25% using AI agents. The company has not said whether that will change its headcount.

Meta has launched two models it calls its most capable yet, aimed at agent tasks. The smaller one is available to developers now, while the larger one is limited to paid Muse Code subscribers.

Google Cloud has launched a Gemini agent that works inside Workspace apps and can split a large job among sub-agents. Trade coverage says IT teams will need to approve what it can reach.

Power Automate fits predictable document steps and Microsoft-connected processes. A custom AI worker may suit varied documents and follow-up replies, but needs careful testing and human checks.

Cognition AI, maker of the autonomous software engineering agent Devin, announced its run-rate revenue reached $900 million in September 2026, up from $492 million in May, signaling major acceleration in enterprise adoption.
Thirty minutes. We look at one process you run every week and tell you exactly what an AI worker would take off your desk, and what it would not.