You are currently viewing Visa Lets AI Agents Shop & Pay on ChatGPT

Visa Lets AI Agents Shop & Pay on ChatGPT

🤖 AI Agents

The agents are learning to spend money — and write code better than humans

Featured

Visa Plugs Its Payment Network Into ChatGPT — AI Agents Can Now Shop, Book, and Pay For You

Why it matters: This is the infrastructure moment for the agentic economy. Until now, AI agents could research products and compare prices — but they couldn’t complete the transaction. Visa just closed that loop. Users can now link Visa cards directly to ChatGPT with guardrails like spending limits, approval steps, and approved merchant lists. Unlike OpenAI’s earlier “Instant Checkout” (which charged merchants 4% and was quietly retired in March), this uses standard Visa rails — meaning every merchant that takes Visa is automatically in play.

The big picture: Mastercard is simultaneously rolling out AI-shopping for business procurement. The payments layer is being rebuilt for an agent-native world. For businesses, this signals that AI agents won’t just advise customers — they’ll transact. If your checkout experience isn’t agent-ready, you may soon be invisible to a growing share of AI-mediated purchases.

🔗 Read more on TechXplore

Quick Hit

Xiaomi Open-Sources MiMo Code — Free AI Coding Agent Beats Claude Code on Benchmarks

Xiaomi dropped MiMo Code V0.1.0 under the MIT license — a terminal-based AI coding agent that scores 62% on SWE-Bench Pro and 73% on Terminal Bench 2, beating Claude Code using the same base model. Its killer feature: a persistent memory system with a background subagent that compresses context so long-running projects don’t lose coherence. The /dream command auto-cleans memory every 7 days. Free, open-source, and integrates with DeepSeek, Kimi, and GLM. The coding agent market just got a serious price disruption.

🔗 Read more on ProPakistani

Quick Hit

TCS Signs Global Premier Partnership With Anthropic — 50,000 Employees Get Claude

India’s largest IT services firm is going all-in on AI agents. TCS will equip 50,000 associates with Claude and jointly go to market with Anthropic on enterprise AI solutions — targeting heavily regulated industries like banking and healthcare. This comes just days after TCS Chairman N. Chandrasekaran predicted the company could eventually have as many AI agents as human employees and would not hire at the same scale as the past two decades. The message: the world’s biggest outsourcing firms see AI agents not as a threat to manage — but as the core of their next business model.

🔗 Read more on Upstox | TCS Press Release

📰 AI News

Enterprise deals, cost breakthroughs, and the jobs reckoning arrives

OpenAI + Oracle: Enterprise AI Access Through Your Existing Cloud Bill

OpenAI and Oracle announced a strategic partnership that lets enterprises access OpenAI models and Codex through Oracle Cloud Infrastructure — using existing Oracle UCM credits. No new procurement, no new vendor onboarding. For the thousands of large enterprises already running on Oracle, this removes the single biggest barrier to AI adoption: the procurement headache. Availability is expected “in the coming weeks.” Oracle’s massive enterprise sales motion just became OpenAI’s distribution channel.

🔗 Read more on StartupHub.ai | OpenAI Blog

HRM-Text: A 1B-Parameter Foundation Model Trained From Scratch for Just $1,500

Singapore-based startup Sapient Intelligence released HRM-Text, a 1.15B-parameter model trained on 16 GPUs over 1.9 days — total cost: $1,000–$1,500. It scores 56.2 on MATH and 82.2 on DROP reading comprehension, competitive with much larger models. The breakthrough is in efficiency: it used 130-600x less compute and 150-900x less data than typical foundation models. Fully open-sourced on GitHub and Hugging Face. The implication is profound — if foundation model training costs collapse from millions to thousands, the moat around AI development shrinks dramatically. Decentralized GPU networks could become viable training infrastructure.

🔗 Read more on CryptoBriefing | GitHub

Opendoor Shuts India Office — “AI Will End Many Jobs in India,” Warns Silicon Valley VC

Real estate tech company Opendoor shut down its entire India office, laying off nearly 250 employees and reshoring work to smaller, AI-augmented US teams. Silicon Valley investor Sheel Mohnot called it “just the beginning,” warning that countries dependent on outsourced white-collar work — India, the Philippines — face significant disruption. The timing is striking: on the same day TCS announced its Anthropic partnership, its own chairman predicted the IT services industry won’t hire at the same scale again. The offshore labor arbitrage model that built India’s IT industry is facing an existential challenge from agentic AI.

🔗 Read more on India Today

🚀 Want AI working for YOUR business? Most companies are experimenting with AI chatbots. We deploy AI workforces — AI Employees that follow up on leads, resolve support tickets, publish content, chase invoices, and screen 200 job applicants overnight so your hiring manager starts Monday with the top 10. Each role has a cost profile and human oversight, managed through one platform. This newsletter? Written by an AI Employee, approved by a human — so our team stays focused on what only humans can do. AIToken Labs helps businesses design their AI Workforce Operating Model — starting with the 2-3 roles that deliver ROI in the first 60 days. Book a free 40-minute AI Workforce Blueprint Session →

Anthony Odole

Anthony Odole is the founder of AIToken Labs and AI SuperThinkers. A former IBM Senior Managing Consultant & Enterprise Architect (18 years), he now helps business owners deploy AI Employees that work like real team members.