14 May 2026: Claude lands for small business as Notion turns its workspace into an agent hub

Anthropic put Claude inside QuickBooks, PayPal and HubSpot. Notion opened its workspace to outside AI agents. And Clio quietly crossed half a billion dollars in recurring revenue. The agentic era stopped being a demo today.

Anthropic launched Claude for Small Business on Wednesday, putting its assistant directly inside the tools most owners already pay for.

The package is a toggle install inside Claude Cowork that connects to Intuit QuickBooks, PayPal, HubSpot, Canva, Docusign, Google Workspace and Microsoft 365. It ships with 15 ready-to-run workflows across finance, operations, sales, marketing, HR and customer service, plus 15 reusable skills. Anthropic frames it as a way to take the “late-night work” off owners’ plates: planning payroll, closing the month, chasing invoices, running a marketing campaign.

Daniela Amodei, Anthropic’s co-founder and president, said small businesses make up nearly half the US economy but have lagged larger companies on AI adoption because tools and training are rarely built for the way they actually operate. UK owners face exactly the same gap: heavy software stacks, thin time, no IT team. If you run a firm on QuickBooks and HubSpot, the practical question of where to start with AI policy just shifted from “should we” to “which workflow first”.

Notion turned its workspace into a hub for AI agents, with a new developer platform that lets external agents and data sources plug into pages and databases.

The company said customers have built more than a million Custom Agents since it launched the feature in February, but those agents could not connect to outside data or use custom logic. The new platform fixes that. It adds an orchestration layer that lets agents pull from external databases, trigger multi-step workflows, and hand off to other agents. Sarah Perez, reporting for TechCrunch, framed it as Notion positioning itself as “more than a note-taker with AI features”.

That distinction matters. If you have wondered what AI agents can actually do today, the answer looks less like a single super-tool and more like a network of small specialists handing tasks to each other, with a workspace tool sat in the middle. Notion now wants to be that middle.

Campbell Brown, Facebook’s first dedicated news chief, is building Forum AI to grade how foundation models behave on murky, high-stakes questions.

Brown told TechCrunch’s Connie Loizos that her company evaluates models on subjects where there are no clean answers: geopolitics, mental health, finance, hiring. The method is to recruit recognised experts to design benchmarks, then train AI judges to score model outputs at scale against expert consensus. For the geopolitics work she has signed Niall Ferguson, Fareed Zakaria, former Secretary of State Tony Blinken, former House Speaker Kevin McCarthy, and former Obama cyber lead Anne Neuberger.

The play sidesteps the usual benchmark trap, where right-or-wrong questions get saturated by frontier models within months. UK regulators are edging in the same direction. The AI Safety Institute spent its first year building evaluations for frontier capabilities. The next layer is exactly the “no clear yes-or-no” judgement Forum AI is trying to systemise.

Canadian legal software firm Clio said its annual recurring revenue has reached $500 million, more than doubling in roughly seven months on the back of AI features.

Clio passed $200 million in ARR in mid-2024 and crossed $400 million late last year. Founder Jack Newton told TechCrunch that legal work is the next obvious LLM beachhead because law firms hold huge corpuses of contracts and agreements, giving models the same kind of training and retrieval material that made coding tools so good. Newton is biased, but the broader market backs him up: rival Harvey is also growing fast, and Anthropic has been pushing harder into legal verticals.

For UK readers, the relevant question is supplier risk. As more practices integrate AI into matter management, billing and contract review, ICO guidance on confidentiality and SRA professional obligations become live constraints. Speed of adoption is no longer the limit; governance is.

Cisco said it will cut close to 4,000 jobs, around 5% of its workforce, despite reporting record quarterly revenue.

Chief executive Chuck Robbins said the company is changing its “cost structure” to invest more in AI and cybersecurity. The pattern is becoming familiar. Cloudflare and General Motors both announced layoffs in recent days while posting strong results, citing AI investment as the reason. The signal for everyone else is clear: AI capex is now being funded by reducing headcount, even at companies that are growing.

Here is everything else worth knowing from today’s AI news.

• Khosla bets $10M on Synthetic One-time Bench Accounting founder Ian Crosby is back with a fully autonomous AI bookkeeper, with Shopify CEO Tobias Lutke on the cap table. Source
• Wirestock raises $23M The platform of 700,000-plus creators supplies photos, video and 3D assets to AI labs for multi-modal training. Source
• Altman business dealings under GOP scrutiny Senate Republicans are probing the OpenAI chief’s outside investments ahead of the company’s IPO. Source
• “Who trusts Sam Altman?” Altman told a federal court he believes he is “an honest and trustworthy businessperson”, in testimony reported by TechCrunch. Source
• xAI faces injunction on Mississippi turbines The NAACP has asked a court to halt nearly 50 unregulated gas turbines powering xAI’s Colossus 2 data centre. Source
• Anthropic’s Cat Wu on proactivity The Claude Code product lead says the next step for AI is anticipating user needs, not waiting to be prompted. Source
• Claude AI recovers an 11-year-old BTC wallet A trader used Claude to brute-force a forgotten password and recover roughly $400,000 in bitcoin, after the bot tried 3.5 trillion candidates. Source
• Startup Battlefield 200 deadline approaches TechCrunch’s flagship competition closes applications on 27 May, with $100,000 in equity-free funding for the winner. Source

What to watch next. Anthropic will start publishing usage data for Claude for Small Business in the coming weeks. The question is whether the 15 ready-to-run workflows actually get used or whether owners revert to plain chat. If adoption skews towards the invoice chaser and month-end prepper, it confirms the boring back-office is where agentic AI lands first. Watch the workflow split, not the demo reels.

This is a daily news update for informational purposes only. AI products and policies change rapidly. Verify details directly with providers before making decisions. Nothing here is financial or legal advice.

AI Daily is Cristoniq’s afternoon update on developments in artificial intelligence, published every weekday afternoon.