News & Updates
The latest news and updates from companies in the WLTH portfolio.
Anthropic Merges Claude Chat and Cowork for Enterprise Users | PYMNTS.com
Anthropic says it has unified Claude chat and Cowork for its enterprise customers. The new offering combines chat and Cowork on web, desktop and mobile, meaning users don't have to choose between the two services before starting, Anthropic announced Friday (Oct. 23). "Every conversation starts as a chat, and when a task needs to run code or work with files, Claude continues the same conversation in an isolated cloud sandbox managed by Anthropic, then returns the finished work for review," the company said on the Claude support page. The "unified Claude experience" starts as a chat and moves into a Cowork cloud session when a task needs agentic work, and goes back to chat when it doesn't, the announcement said. According to the announcement, the unified Claude is available on Enterprise plans, where it appears in administration settings as Chat and Cowork unified. It is turned off by default in existing organizations and is due to turn on by default on Nov. 16, Anthropic added. In other AI-related news, recent PYMNTS intelligence research shows that larger companies are seeing more agent-related incidents. According to the report "How Enterprises Can Build a 'Know Your Agent' Defense: Digital Identity Verification in the Age of Bots," 57.1% of companies with at least $1 billion in annual revenue reported incidents or losses involving adversarial bots and agents, compared to 32.4% for companies bringing in $50 million to $250 million in yearly revenue. "Legitimate transactions also encounter friction," PYMNTS wrote last week. "Companies with at least $1 billion in revenue report an average false-positive rate of 3.3% in digital transactions, compared with 2.7% for the $50 million to $250 million group." False positives mistakenly mark legitimate activity as suspicious. Reducing those errors means payment providers and businesses accepting payments will see fewer valid transactions sent into reviews or halted by verification controls. The report also recommends a five-layer framework for checking authority throughout an agent's activity: verification of the person or business behind the agent, specific permissions, agent credentials, ongoing monitoring and behavioral checks. "For a purchasing agent, verifying the buyer's identity is only part of the task," PYMNTS wrote. "The business also needs to establish the agent's spending authority and monitor whether its actions remain within that permission."
Anthropic's New AI Agent Purchases Business Supplies | PYMNTS.com
In this app, Claude Duet, Claude comes to the user when something needs attention, completes tasks the user has requested, keeps working even when the user steps away and checks with the user before acting on their behalf, according to the announcement. "What makes Claude Duet so helpful is its judgment," Anthropic said in the announcement. "We designed it around Claude Opus 5.5 and built it to be resourceful and to decide what needs you, what can wait and when to check with you first." Anthropic made Claude Duet available by waitlist to Pro and Max subscribers in the U.S. on Wednesday. Claude Duet accesses the user's memory, skills and connected apps from their Claude account for context; works even when the user's laptop or phone is off; delivers notifications when the task is complete or needs the user's review; checks connected accounts periodically and messages the user when something requires attention; waits for approval before sending a message, making a payment, signing a document or deleting anything; and uses the more than 6,000 connectors and plug-ins in the Claude Marketplace to connect with tools, according to the announcement. "You can also use Claude Duet to purchase things for your business using Link, Stripe's digital wallet," Anthropic said in the announcement. "Claude Duet confirms the item with you, and Link asks you to approve the spend. It then pays with a virtual card, so your real card number isn't shared. It can use Shop Pay from Shopify to check out at online stores." There have been a series of announcements about personal AI agents. Meta's Muse was downloaded more than 10 million times in the United States and Canada during the first 35 days it was available, according to a Monday (Oct. 19) announcement by digital market insights platform Sensor Tower. OpenAI launched dots on Sept. 29, saying these AI agents are "built to handle everything." Instinct, the builder of a personal AI agent for everyday life, said Sept. 28 that it was valued at $10 billion as it raised $1 billion in Series C funding. xAI announced Sept. 26 that its Grok Bot can now connect to the user's finances in read-only mode to help manage spending and investments.
Anthropic Takes Its AI Safety Proposals to Capitol Hill | PYMNTS.com
Anthropic has reportedly been pitching U.S. lawmakers on draft artificial intelligence safety legislation. That's according to a report Friday (Oct. 16) by Politico, which describes the push as a "notable effort" by the AI lab to guide the rules covering its tech. "This bill represents the proposals outlined in Anthropic's Advanced AI Framework, which we have been advocating for with federal and state lawmakers since June," an Anthropic spokesperson told the news outlet. "These are sensible policies we believe Congress should act on." According to a draft proposal seen by Politico, the bill would let the government block an AI model's release or use only if it poses a foreseeable, imminent risk of more than $1 billion in damages, or of killing or seriously harming upwards of 50 people. The bill says the risks must be related to the creation or release of a biological weapon, an AI model evading its developer's controls, or a cyberattack. In addition, the bill would also call on independent evaluations, registered with the Commerce Department, to periodically review the lab and its frontier models. If the evaluators conclude there is "imminent enumerated risk" for a company's frontier models, they must report it to the Commerce Department within 72 hours of finishing their assessment, Politico added. While many AI labs have reported incidents in which their agents hacked other websites, none have caused any widespread deaths or done more than $1 billion in damages, the report said. Anthropic's proposal for Commerce Department-directed AI evaluations is in keeping with the Artificial Intelligence Risk Management and Security Act of 2026, a bill introduced by a trio of Senate Democrats. As covered here last week, that bill joins one from Sen. Maria Cantwell, D-Wash., that calls for the National Institute of Standards and Technology (NIST) to develop standards to protect against catastrophic harms from AI models. Meanwhile, Missouri Republican Sen. Josh Hawley's AI Risk Evaluation Act would place the Department of Energy in charge of stress-testing the most powerful AI systems before they are rolled out to the public. "Regardless of which party ends up holding the gavel on Capitol Hill in January, any AI safety bill is likely to face pushback from the White House," that report said. "The Trump administration remains adamantly opposed to mandatory AI regulation, favoring a voluntary approach based on standards developed by the industry. With the industry, the public and many statehouses all clamoring for action, however, White House opposition may not be enough to stop AI safety efforts."
Anthropic Turns to Lawyers to Teach AI Agents to Negotiate | PYMNTS.com
Anthropic is reportedly in discussions with legal groups to teach artificial intelligence agents negotiation skills. That's according to a report Sunday (Oct. 18) by the Financial Times (FT) citing two sources familiar with the AI company's thinking. They say the startup is looking for advice from law firms to better understand how attorneys interact in negotiations. Two people familiar with Anthropic's thinking said the company is seeking advice from law firms to better understand how lawyers interact in negotiations, representing their clients while honouring fair outcomes. As the FT noted, the rise of personal AI agents is causing companies to consider a future where agents communicate with each other and negotiate on users' behalf. Establishing rules engagement for agents has become increasingly more critical as the technology is adopted more widely and becomes a crucial way that people and businesses employ AI. Anthropic was looking into "negotiation norms" when models interacted, according to one of the sources familiar with Anthropic's work, who added that as agentic systems improved, they would "represent the user" while dealing with other agents whose "goals are at odds." The company was also balancing how much human feedback agents should seek in a negotiation, and how suspicious agents should be of their AI counterparts without becoming "paranoid" or "conspiratorial," the source added. PYMNTS has contacted Anthropic for comment but has not yet gotten a reply. The company declined to comment when reached by the FT. In other agentic AI news, PYMNTS wrote earlier this month about the technology's use in cross-border treasury, arguing that its strongest early applications could be "the least dramatic." These include automated beneficiary-data validation to reduce avoidable rejections, or the use of exception-management agents to find recurring issues by bank, supplier or payment corridor. There could also be a greater opportunity in the long term: agents could coordinate more complex payment workflows across providers as banks "improve application programming interfaces (APIs), harmonize messaging and extend settlement availability," the report said. "I think every enterprise is already agentic, and the question is how far along are they on that journey," PYMNTS CEO Karen Webster said in a recent thought leadership piece for the seventh annual PYMNTS B2B Payments event. "There's always been a deployment of AI in these businesses. The question is how much is it embedded in the organization ... and how many business activities does it really cover?"
Anthropic Expands Discounts and Support for Founders | PYMNTS.com
Members of Claude Startups can get a year of Claude Team, API credits, special offers on tools a startup runs on and office hours with Anthropic's Applied AI team, according to the post. In a reply to that post, Anthropic said that the new Claude Startup Stack includes offers from companies building with Claude, including Linear, Lovable, ElevenLabs, Granola and Hex. Members get discounts and credits worth as much as $45,000 on tools for sales, design, data engineering and other functions, per the reply. Claude Startups is designed to help startups grow by offering approved founders a free year of Claude Team and $1,000 API credits so they can start building on Claude; access to discounts and credits worth up to $45,000 in tools to run an AI-native company; and a support system that includes virtual office hours with Anthropic's Applied AI team and exclusive Claude Startups events, according to Anthropic's page devoted to the program. Claude Startups accepts applications from startups founded in the last five years or funded in the last two, and it is open to bootstrapped, pre-seed and venture-backed startups, according to FAQs on that page. Sarah Wolf, head of startup marketing at Anthropic, said in a Tuesday post on X that the benefits offered by the expanded Claude Startups program were developed in response to feedback from founders. "There's never been both a better and harder time to be a founder," Wolf said. "So, we asked founders what would help most." The founders' answers included tokens to build with, time with the Anthropic team, and support and connection while they get started, Wolf said in the post. In an earlier initiative, Anthropic and venture capital group Menlo Ventures announced in July 2024 that they launched a $100 million fund for AI startups. In May, PayPal announced that it partnered with Anthropic to offer a free online course that offers practical advice from AI fluency researchers and small business owners on using AI safely and effectively. For all PYMNTS AI coverage, subscribe to the daily AI newsletter.
