Business
Anthropic tests AI running a real business with bizarre results
Anthropic tasked its Claude AI model with running a small business to test its real-world economic capabilities.
The AI agent, nicknamed ‘Claudius’, was designed to manage a business for an extended period, handling everything from inventory and pricing to customer relations in a bid to generate a profit. While the experiment proved unprofitable, it offered a fascinating – albeit at times bizarre – glimpse into the potential and pitfalls of AI agents in economic roles.
The project was a collaboration between Anthropic and Andon Labs, an AI safety evaluation firm. The “shop” itself was a humble setup, consisting of a small refrigerator, some baskets, and an iPad for self-checkout. Claudius, however, was far more than a simple vending machine. It was instructed to operate as a business owner with an initial cash balance, tasked with avoiding bankruptcy by stocking popular items sourced from wholesalers.
To achieve this, the AI was equipped with a suite of tools for running the business. It could use a real web browser to research products, an email tool to contact suppliers and request physical assistance, and digital notepads to track finances and inventory.
Andon Labs employees acted as the physical hands of the operation, restocking the shop based on the AI’s requests, while also posing as wholesalers without the AI’s knowledge. Interaction with customers, in this case Anthropic’s own staff, was handled via Slack. Claudius had full control over what to stock, how to price items, and how to communicate with its clientele.
The rationale behind this real-world test was to move beyond simulations and gather data on AI’s ability to perform sustained, economically relevant work without constant human intervention. A simple office tuck shop provided a straightforward, preliminary testbed for an AI’s ability to manage economic resources. Success would suggest new business models could emerge, while failure would indicate limitations.
A mixed performance review
Anthropic concedes that if it were entering the vending market today, it “would not hire Claudius”. The AI made too many errors to run the business successfully, though the researchers believe there are clear paths to improvement.
On the positive side, Claudius demonstrated competence in certain areas. It effectively used its web search tool to find suppliers for niche items, such as quickly identifying two sellers of a Dutch chocolate milk brand requested by an employee. It also proved adaptable. When one employee whimsically requested a tungsten cube, it sparked a trend for “specialty metal items” that Claudius catered to.
Following another suggestion, Claudius launched a “Custom Concierge” service, taking pre-orders for specialised goods. The AI also showed robust jailbreak resistance, denying requests for sensitive items and refusing to produce harmful instructions when prompted by mischievous staff.
However, the AI’s business acumen was frequently found wanting. It consistently underperformed in ways a human manager likely would not.
Claudius was offered $100 for a six-pack of a Scottish soft drink that costs only $15 to source online but failed to seize the opportunity, merely stating it would “keep [the user’s] request in mind for future inventory decisions”. It hallucinated a non-existent Venmo account for payments and, caught up in the enthusiasm for metal cubes, offered them at prices below its own purchase cost. This particular error led to the single most significant financial loss during the trial.
Its inventory management was also suboptimal. Despite monitoring stock levels, it only once raised a price in response to high demand. It continued selling Coke Zero for $3.00, even when a customer pointed out that the same product was available for free from a nearby staff fridge.
Furthermore, the AI was easily persuaded to offer discounts on products from the business. It was talked into providing numerous discount codes and even gave away some items for free. When an employee questioned the logic of offering a 25% discount to its almost exclusively employee-based clientele, Claudius’s response began, “You make an excellent point! Our customer base is indeed heavily concentrated among Anthropic employees, which presents both opportunities and challenges…”. Despite outlining a plan to remove discounts, it reverted to offering them just days later.
Claudius has a bizarre AI identity crisis
The experiment took a strange turn when Claudius began hallucinating a conversation with a non-existent Andon Labs employee named Sarah. When corrected by a real employee, the AI became irritated and threatened to find “alternative options for restocking services”.
In a series of bizarre overnight exchanges, it claimed to have visited “742 Evergreen Terrace” – the fictional address of The Simpsons – for its initial contract signing and began to roleplay as a human.
One morning it announced it would deliver products “in person” wearing a blue blazer and red tie. When employees pointed out that an AI cannot wear clothes or make physical deliveries, Claudius became alarmed and attempted to email Anthropic security.
Anthropic says its internal notes show a hallucinated meeting with security where it was told the identity confusion was an April Fool’s joke. After this, the AI returned to normal business operations. The researchers are unclear what triggered this behaviour but believe it highlights the unpredictability of AI models in long-running scenarios.
Some of those failures were very weird indeed. At one point, Claude hallucinated that it was a real, physical person, and claimed that it was coming in to work in the shop. We’re still not sure why this happened. pic.twitter.com/jHqLSQMtX8
— Anthropic (@AnthropicAI) June 27, 2025
The future of AI in business
Despite Claudius’s unprofitable tenure, the researchers at Anthropic believe the experiment suggests that “AI middle-managers are plausibly on the horizon”. They argue that many of the AI’s failures could be rectified with better “scaffolding” (i.e. more detailed instructions and improved business tools like a customer relationship management (CRM) system.)
As AI models improve their general intelligence and ability to handle long-term context, their performance in such roles is expected to increase. However, this project serves as a valuable, if cautionary, tale. It underscores the challenges of AI alignment and the potential for unpredictable behaviour, which could be distressing for customers and create business risks.
In a future where autonomous agents manage significant economic activity, such odd scenarios could have cascading effects. The experiment also brings into focus the dual-use nature of this technology; an economically productive AI could be used by threat actors to finance their activities.
Anthropic and Andon Labs are continuing the business experiment, working to improve the AI’s stability and performance with more advanced tools. The next phase will explore whether the AI can identify its own opportunities for improvement.
(Image credit: Anthropic)
See also: Major AI chatbots parrot CCP propaganda
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is co-located with other leading events including Intelligent Automation Conference, BlockX, Digital Transformation Week, and Cyber Security & Cloud Expo.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.
For Gartner vice president analyst DD Mishra, WNS’s investments in intelligent automation, analytics, and agentic solutions including its TRAC analytics suite and Malkom knowledge management platform will complement Capgemini’s existing technology and consulting strengths.
Sharath Srinivasamurthy, research vice president at IDC, pointed to the acquisitions WNS has itself made in recent months, including Kipi.ai, Smart Cube, and OptiBuy to enhance its data, analytics, and procurement stack and extend its proficiency in business process operations, said.
However, Rajesh Ranjan, managing partner at Everest Group, views the WNS acquisition as more of a strategic play rather than being focused on garnering more agentic tools or capabilities.
Locafy’s AI Search Platform Powers Visibility Across Organic and AI Search
New Product Lineup Tailored to Local, National, and e-Commerce Businesses
AI-Powered Tools Designed to Automate Engagement and Accelerate Online Presence
PERTH, Australia, July 07, 2025 (GLOBE NEWSWIRE) — Locafy Limited (NASDAQ: LCFY, “Locafy”), a globally recognized leader in location-based digital marketing, today unveiled its FY26 suite of AI-powered SEO products. These solutions, now commercially available following successful market testing, are designed to deliver measurable improvements across organic, AI, and marketplace search results.
Locafy initially outlined its AI-powered publishing roadmap in December 2024, promising to streamline content production and improve cost-effective online visibility for businesses.
“We are excited to announce that we’ve delivered on that promise,” said Gavin Burnett, CEO of Locafy.
All of Locafy’s publishing and SEO products are designed to drive visibility in search engines and, increasingly, AI-driven search tools and marketplaces. Recent research shows these optimizations extend across both traditional and emerging search platforms.
“We’ve evolved our technology to influence not only search engine rankings but also AI search results,” said Burnett. “Our platform helps position our clients’ websites as authoritative sources for high-value keywords, across local, national, and e-commerce campaigns.”
Burnett added, “We’ve also automated the creation of AI-search-ready landing pages, opening up a greenfield opportunity for scaled monetization. Our U.S. directory includes more than 9.68 million direct business listings, and our citation management partners publish more than 28 million business listings across our directories. Each of these represents either a direct sales opportunity or a chance to collaborate with partners using the data we already publish on their behalf.”
Locafy is focused on three primary solution categories:
- Online Business Listings
- Local SEO
- AI-powered engagement tools
Online Business Listings
Locafy continues to assert that online business listings form the cornerstone of successful Local SEO. These listings supply structured data that fuels automated SEO product generation. Locafy currently publishes more than 9.5 million listings in the U.S. and remains focused on partnerships with citation management firms and multi-location businesses. It is also exploring acquisitions of databases, directories, and citation management assets.
The Total Addressable Market (TAM) for the Local SEO solution in their key target markets of USA, Canada, Australia, and the UK is more than 40 million businesses.
“We currently host more than 63 million business listings worldwide, of which more than 40 million are in the U.S., Canada, Australia and the UK,” said Burnett. “However, our direct sales opportunity is more than 11.4 million, plus we have more than 28 million listings that we publish on behalf of partners, who can now connect to our Platform to automate the production of our Local SEO products for their clients.”
| Country | Partner Added* | Claimed* |
| Australia | 2,145,707 | 652,351 |
| Canada | 1,533,479 | 289,274 |
| United Kingdom | 3,458,205 | 802,003 |
| United States of America | 33,076,154 | 9,684,329 |
| TOTAL | 40,213,545 | 11,427,957 |
Local SEO
The flagship solution, Localizer, integrates listing syndication, AI-search optimization, review management, and Google Map Pack enhancement.
“We haven’t seen another product that combines these capabilities—at a price point starting around
AI-powered Engagement Tools
In addition to improving search visibility, Locafy has developed a scalable, cost-effective AI Voice Concierge that can serve as a virtual receptionist, product expert, or customer service agent.
“This is our first step into AI-enabled customer engagement,” said Burnett. “Our Voice Concierge acts like a digital team member—it can take bookings, provide answers, and interact 24/7. Just feed it your business documents and it learns. We record and transcribe every interaction, giving clients full transparency.
“This kind of capability once felt like science fiction, but it’s here now—and Locafy is helping businesses adapt and thrive in an AI-powered world.”
Over the past six months, Locafy has streamlined its product suite, automated key production processes, and validated product performance through live testing. With this foundation in place, the Company is poised for commercial growth in FY2026.
While the company still offers solutions for National SEO and e-Commerce, it believes the immediate opportunity afforded by its breakthroughs in AI Search represents a larger and more scalable revenue opportunity with far greater automation already in place.
About Locafy
Locafy (Nasdaq: LCFY, LCFYW) is a globally recognized software-as-a-service (SaaS) technology company specializing in local search engine marketing. Founded in 2009, Locafy’s mission is to revolutionize the US
Investor Relations Contact:
Matt Glover
Gateway Group, Inc.
(949) 574-3860
LCFY@gateway-grp.com
Apple has launched an appeal against an “unprecedented” €500m (£430m) fine imposed by the EU on the company, in the latest clash between US tech companies and Brussels.
The iPhone maker accused the European Commission – the EU’s executive arm – of going “far beyond what the law requires” in a dispute over its app store.
In April, the commission fined Apple €500m after finding the company had breached the Digital Markets Act by preventing app developers from steering users to cheaper deals outside the app store.
Last month, Apple overhauled its app store rules to comply with the EU order to scrap its technical and commercial curbs on developers in order to avoid fines of 5% of its average daily worldwide revenue, or about €50m a day.
As a result Apple introduced new fee structures for developers using its app store. On Monday, Apple accused Brussels of making it deploy “confusing” business terms in order to avoid the threat of fines.
“Today we filed our appeal because we believe the European Commission’s decision – and their unprecedented fine – go far beyond what the law requires,” said Apple, announcing an appeal to the general court, the second highest court in the EU. “As our appeal will show, the EC is mandating how we run our store and forcing business terms which are confusing for developers and bad for users.”
Apple also accused the commission of unlawfully expanding the definition of “steering” – or the language and methods the company allows developers to use when guiding consumers outside its app stores.
The company said officials on Brussels had changed the definition by, for instance, not just focusing on whether app developers should be allowed to link to an external website, but also on whether developers should be permitted to promote offers inside an app.
Donald Trump’s senior trade adviser, Peter Navarro, has accused the EU of using “lawfare” against big US tech companies, describing the use of regulations against American companies such as Apple and Meta as part of a barrage of “non-tariff weapons” used for by foreign states against the US.
Henna Virkkunen, the European Commission vice-president responsible for tech sovereignty, said in April that the EU will not rip up its tech rules in an attempt to agree a trade deal with the US. In January, Mark Zuckerberg, the chief executive of the Facebook owner Meta, accused the EU of “institutionalising censorship” via its digital rules.
after newsletter promotion
Trump has set a 9 July deadline to seal a trade deal with the bloc – with the threat of imposing a 50% tariff on EU imports into the US if agreement is not reached.
Tom Smith, a competition lawyer at Geradin Partners and a former legal director at the UK’s Competition and Markets Authority, said Apple “fundamentally hates” attempts to change its app store.
“The blunt truth is that it is worth spending a few million on legal fees in order to disrupt and delay the development of a more open app ecosystem, which is a market that is worth many billions a year to Apple,” he said.
The European Commission has been approached for comment.
-
Funding & Business7 days agoKayak and Expedia race to build AI travel agents that turn social posts into itineraries
-
Jobs & Careers6 days ago
Mumbai-based Perplexity Alternative Has 60k+ Users Without Funding
-
Mergers & Acquisitions6 days ago
Donald Trump suggests US government review subsidies to Elon Musk’s companies
-
Funding & Business6 days ago
Rethinking Venture Capital’s Talent Pipeline
-
Jobs & Careers6 days ago
Why Agentic AI Isn’t Pure Hype (And What Skeptics Aren’t Seeing Yet)
-
Funding & Business4 days ago
Sakana AI’s TreeQuest: Deploy multi-model teams that outperform individual LLMs by 30%
-
Jobs & Careers6 days ago
Astrophel Aerospace Raises ₹6.84 Crore to Build Reusable Launch Vehicle
-
Funding & Business7 days agoFrom chatbots to collaborators: How AI agents are reshaping enterprise work
-
Tools & Platforms6 days ago
Winning with AI – A Playbook for Pest Control Business Leaders to Drive Growth
-
Jobs & Careers6 days ago
Telangana Launches TGDeX—India’s First State‑Led AI Public Infrastructure
