The most important news pieces from the AI world are here for you, whether you’re an AI enthusiast or just wish to keep an eye on what’s going on.
Focus Topic: AI is moving into the tools and data you already use
For years, using AI at work meant copying numbers out of your spreadsheet, pasting them into a chat window and copying the answer back. This week, several launches pointed the other way: the AI comes to where your work already is.
The clearest example is the new Claude sidebar inside Google Docs, Sheets and Slides. Claude can edit the open file directly or ask for approval before each change. That second option matters more than it sounds. In a budget model or a reconciliation file, “the AI changed a cell” and “the AI proposed a change and I approved it” are two very different control situations. In the same week Anthropic also launched Claude Dashboards, which turns your data into interactive dashboards.
OpenAI is moving in the same direction from the other end. With intelligent UI, ChatGPT can answer with interactive charts, maps and tools instead of a wall of text. OpenAI is also testing a Meetings plugin for the Mac app that turns meetings into notes and next steps based on each user’s past work. The chat window is slowly becoming a workspace.
The most interesting part for me is close to home. In under three months, Pipedrive, LHV and Creditinfo have opened MCP servers that bring their data and services straight into customers’ AI assistants. MCP is the open standard for how an assistant like ChatGPT or Claude connects to company software. One well-built server works across many AI tools, the same way a website opens in any browser. The article compares it to companies building their first websites 25 years ago. Creditinfo Estonia’s head Oksana Tolmatshova put it well: the company may become less visible to clients, but far more valuable.
Think about what that means for a finance team. Your bank data, your CRM pipeline and a credit check on a new client can all be asked about in plain language from the same chat, without logging into three systems and exporting three files. The work moves from collecting data to checking it and deciding.
There is a catch, and it is the boring finance one: access and control. When the assistant can read your bank account and edit your forecast file, the questions become who connected what, which account it uses, and whether changes are approved or automatic. My simple advice: start with read-only access, turn on approval before every change, and write down which connections your team has switched on. Most of us already have rules for who can see payroll in Excel. The same rules now need to cover the AI.
LLMs & AI Models
- Anthropic launched Claude Haiku 5.5, a budget model at $0.10 per million input tokens that beats GPT-6 Luna on several benchmarks.
- Mistral Large 4 (“Le Chonk”) and Reflection’s Beam are two serious Western open models in one week, but China’s Kimi K3 still leads.
- ChatGPT can now answer with interactive charts, maps and tools instead of plain text, thanks to its new intelligent UI.
- Claude now works inside Google Docs, Sheets and Slides, and Claude Dashboards turns your data into interactive dashboards.
- OpenAI is testing a Meetings plugin for the ChatGPT Mac app that turns meetings into notes and next steps.
- Claude Code got mods, small add-ons such as a guard that pauses risky delete commands.
- Musk’s Grok Bot will route tasks to Claude and other rival models, effectively making xAI an Anthropic customer.
New Tools
- Hark Pro from Figure AI’s founder is a proactive assistant that researches, shops and books for you on its own cloud computer.
- Google rolled out SynthID Detector worldwide, so anyone can check images, video and audio for AI watermarks.
- Anthropic launched OSS Scanner, a free service that scans open-source projects for vulnerabilities and suggests fixes.
- Google Labs Playground lets U.S. adults describe a game in chat and play it minutes later.
- Meta released Muse Gadgets, open-source code for DIY devices that plug into its Muse agent.
👉 Explore these tools: Hark Pro | SynthID Detector | OSS Scanner | Google Playground | Muse Gadgets
Other Quick Picks
- OpenAI’s safety report lead quit, calling the culture “broken”, and 3 fired researchers hit back.
- OpenAI released 722 math papers from an internal model, and mathematicians are angry.
- Biohub grew to a $1.8B effort with the U.S., DeepMind and Meta to build AI that simulates human cells.
- a16z’s Top 100 AI apps: only 4.5% of Americans pay for a chatbot, but the top 1% spend $903 a month.
- SemiAnalysis finds Claude plans give about 5x more value per dollar than ChatGPT plans.
- OpenAI’s textGrain puts an invisible watermark on ChatGPT text in the EU to meet AI Act rules.
- Anthropic will ban “sustained and needless” abuse of Claude from 12 November.
- Sam Altman told Politico the world “should accept some bad things happening” for AI’s benefit.
- Norway plans to ban AI smart glasses from gyms, schools, playgrounds and beaches.
- Utah lets an AI app diagnose acne and prescribe creams on its own, a U.S. first.
- Nous Research raised $90M, and its open Hermes agent now handles ~2.5% of global tokens.
- Trump set up a “Super Intelligence Force”, and Musk may rename SpaceXAI to SpaceXSI.
🇪🇪 AI News from Estonia
- Pipedrive, LHV and Creditinfo brought their data into AI assistants via MCP servers, the “new website” for reaching clients.
- Jarmo Tuisk’s report sets out three AI paths for Estonia and says to judge success by impact, not people trained.
- The Data Protection Inspectorate warns schools not to feed pupils’ grades or health data into AI.
- EIS will also fund AI projects that first missed out on the €20,000 grant after 200+ companies applied.
Want to stay in the loop?
Straight to your inbox: practical AI updates, finance use cases, tools to try, and upcoming trainings.
I cut the noise and send only what’s actually useful.
🎙️ AIPowerment Podcast #93 is out: September’s key AI news, from GPT-6 Astra and the AI shopping-agent war to Estonia’s first AI-agent payment.
🎧 Listen to AIPowerment Podcast on Spotify, Apple Podcasts, and YouTube.
💼 Exploring how to make AI work in finance? Let’s talk practical use cases – connect with Gerlyn Tiigemäe for expert guidance.
📚 If you would like to participate in one of my trainings or listen to speaking engagements, here are the upcoming ones:
- I am planning an e-course to teach about AI in Finance, so definitely sign up for the Inside Finance Newsletter to know about the launch 😉
- Training “AI võimalused finantsvaldkonnas” @ Äripäeva Akadeemia, next trainings for beginners @ 11.11 – information; for intermediate level @ 26.10 – information
- Training “AI võimalused projektijuhtidele” @ Äripäeva Akadeemia, next trainings @ 14.12 – information
- I am doing a couple of trainings together with Eesti.ai algatus, which are all free, so check out their website to find suitable trainings here.
