AI Radar for Companies — Friday, October 30, 2026
noticias ia para empresas claude proteccion de datos

AI Radar for Companies — Friday, October 30, 2026

· CompaniesAutomation

Claude tells managers how teams use AI, a Madrid court clarifies an AI-linked layoff, Google earns less per search and Cohere raises up to C$7 billion.

AI isn't just doing work inside companies anymore. Now it's measuring it, too. Anthropic is giving executives reports that read a sample of each team's conversations with Claude, and in Spain, a ruling from Madrid's regional High Court on a collective layoff (an ERE) at a video game studio clarifies how far AI can go as grounds for dismissal. And money tells two stories: Google has more searches than ever but is struggling to monetize them, while Cohere is raising the largest private round in Canadian history, at a C$35.7 billion valuation, with the owner of Lidl as lead investor. Catch up on Thursday's radar (Oct. 29).

Claude Enterprise now tells managers how each team uses AI: tell your staff before you switch it on

Anthropic has opened "smart reports," in-app surveys and pooled group budgets to Claude Enterprise customers. To run a report, you pick a team, a date range and the products (chat, Claude Code or Cowork): Claude analyzes a sample of the team's conversations, groups them into workstreams, and shows how much each cost, what was delivered, and where people get stuck, such as a missing connector. According to Mashable, reports can cover up to the last 28 days, sessions can be sorted by cost, and the report flags as "inefficiencies" those that produced nothing useful or were personal or off-topic; there are 10 free reports per month. Anthropic warns in writing that they are not intended to evaluate individuals or make employment decisions and, according to the same outlet, the tool refuses questions about health, religion, or union activity. The takeaway: in Spain, the vendor's disclaimer isn't enough. Spain's data protection law (Article 87 of the LOPDGDD) requires employers to set written rules for the use of digital devices, with the participation of workers' representatives, and to inform staff. The EU AI Act also treats systems that monitor or evaluate worker performance as high-risk. Before the first report, add to your AI policy what gets analyzed, who sees it and what it won't be used for. Then use it for what it's good at: the three most repeated tasks are the first ones to turn into a standard procedure, and that's the step that takes you from ASSISTED to INTEGRATED on the ladder. The pooled budget, with alerts at 50, 75, 95, and 100%, is the per-team cap we called for in the Oct 27 flash. Source · Source

Madrid's High Court and the first layoff case involving AI: the economic grounds matter, not the tool

Several Spanish outlets ran headlines on Thursday saying the courts had backed an AI-driven layoff "for the first time," but the ruling is more nuanced than that. Ruling 864/2026 from the employment chamber of Madrid's High Court, dated October 8, decides the appeal of a writer from a video game studio dismissed in an ERE that affected 13 of its 83 workers. The company swung from a €646,297 profit in 2023 to a €3,546,027 loss in 2024, its revenue fell from €7.99 million to €4.60 million, and in October 2024 it notified a court that it was insolvent. AI shows up in the dismissal letter of another worker, a 3D modeler: the company explained that images generated with AI made a "new production process" necessary and that two senior artists were enough. The lower court ruled that dismissal fair, it was never appealed to the High Court, and the company itself framed AI as a consequence of its situation and what its clients requested, not as a direct cause. By contrast, a court in Bilbao ruled it was unfair dismissal to replace an employee with AI with no other justification. The takeaway: if automation is going to change jobs, what holds up in court is a documented change in how you produce and the numbers behind it, not "we bought an AI." And before getting there, measure what tasks the tool actually covers: it's the same data that tells you whether redeploying people beats letting them go. Source

Google is getting more searches than ever but earning less from them: ads are moving into its AI answers

Alphabet reported third-quarter results after Wednesday's close, and on Thursday its shares were down about 12%, on track for their worst post-earnings drop since the 2004 IPO. Search ad revenue grew 12% to $63.35 billion, below expectations for the second straight quarter, even as Google says queries are growing at the fastest rate in five years. The company mainly blamed currency swings, but Chief Business Officer Philipp Schindler acknowledged that lower monetization in AI Mode also played a part, and said Google is scaling ads in AI Overviews and AI Mode, including in international markets. Free cash flow was negative $9.07 billion on data center spending, and Google Cloud revenue more than doubled to nearly $31 billion. The takeaway: if your business runs on Google Ads or SEO, the Google click is changing shape, and your ads may start appearing inside AI-generated responses. This month, split the leads in your reports by source: ads, organic search and assistants like ChatGPT, and check that your website clearly states prices, hours, and what you sell, because that is what an AI reads when it answers for you. Source · Source

Cohere raises up to C$7 billion with Lidl's owner in the lead: the money is backing AI you control

Cohere, the Canadian AI company for regulated sectors, has closed the first part of its Series E, which targets up to C$7 billion at a C$35.7 billion valuation. It is the largest private round by a Canadian tech company. It is led by Schwarz Group, the owner of Lidl and Kaufland, which after Cohere's combination with Germany's Aleph Alpha wants to run its models on its own European cloud, STACKIT. New backers include the Government of Canada, Bell Ventures and Bullhound, while Nvidia, Inovia and Prosperity7 are returning. Cohere says it ended September with annual recurring revenue up 165% year over year, and it has grown its customer base 20x, to 300, including RBC, PwC, Fujitsu, and TD Bank. Its selling point is the ability to deploy AI in a private cloud, on your own servers, or without an internet connection. The takeaway: "sovereign AI" is moving from government talking point to product catalog, with a major European retailer behind it. If you work with health, financial, or customer data that you don't want to leave the EU, you can now ask in any proposal for the model to run on European servers under a European contract. Also ask what happens to your data and your fine-tuning if you switch vendors: that clause matters most as the market consolidates. Source

What to watch next

Amazon reports after the bell on Thursday: watch how fast AWS grows and how much it spends on data centers, because that drives the cloud prices you'll pay. Also watch whether Anthropic publishes full documentation for the reports and whether other vendors copy the feature.