The company revealed that over one million businesses already use early versions of the agent which was built inside Meta’s messaging platforms, where over a billion conversations happen daily between people and businesses.

The Business Agent, which works in local languages, responds to customer questions without delay and also makes product suggestions from a company’s catalogue. It can book appointments, qualify incoming leads and complete sales.

Businesses can set the tone, and the agent adjusts its language to match.

At the event in London, demonstrations showed how quickly the agent replies to customer messages, with one test illustrating a full product enquiry handled from start to checkout without human input.

Meta also built in a morning briefing feature that sends business owners a summary of missed messages and highlights from customer chats. It is aimed at teams that cannot stay online all day.

The company is also expanding access with new businesses being able to set up the agent in minutes. The entry option is free, but Meta plans to introduce paid tiers later for different business sizes.

Discovery is also changing, letting customers search for businesses directly on WhatsApp soon. They can also share contact cards in chats as well as trigger direct conversations with companies and Meta says this will make response times faster once customers reach out.

Behind the tool is a larger system called the Meta Business Agent Platform. It allows companies to build and customise their own agents and also connects with external services such as Shopify, Zendesk and Shopee. This gives the agent the ability to act on behalf of a business across different systems.

Larger companies get added features including guardrails, measurement tools and options to set policies around customer interactions. Meta says this is aimed at ensuring consistency and oversight at scale.

The company is expanding support across WhatsApp, Messenger and Instagram. It says businesses will eventually be able to use the system for more than customer service. Future plans include market research, product insights, calendar management and competitive tracking.

Meta is also investing heavily in artificial intelligence infrastructure, with spending expected to reach $145 billion. The Business Agent aims to turn messaging platforms into full business tools.

OpenAI, Anthropic and Google are all building similar AI systems for enterprise use. Meta’s focus is strongly on integrating these tools directly into its social and messaging apps, where businesses already operate.

The venture, called The Deployment Company, brings together 19 investors, including TPG Inc., Brookfield Asset Management, Advent International and Bain Capital. SoftBank Group and Dragoneer Investment Group are also involved.

People with direct knowledge say the new company is valued at about $10 billion, not counting the new capital raised, while OpenAI will keep control of the business.

OpenAI wants its tools used inside more companies, not just tested. So it will place its engineers, who will help redesign workflows, automate routine tasks and ensure wider use of its software, directly within organisations backed by these investors.

This approach changes direction from simply selling access to software to focusing on hands-on deployment. It is closer to a service model, where companies pay not just for tools, but for implementation and ongoing support.

The investors backing the venture control more than 1,000 companies between them. That gives OpenAI a ready pipeline of clients without relying on long sales cycles. It also means faster rollout across sectors.

OpenAI has committed about $500 million upfront, with the option to increase that to $1 billion later. The rest of the funding will come from private equity firms over the next few years.

Interestingly, OpenAI is offering investors a 17.5% annual return, and if the venture doesn’t meet expectations, it will cover the gap, creating risk. On a $4 billion commitment, the shortfall could run into hundreds of millions each year if returns disappoint.

The development comes after Anthropic secured about $1.5 billion for a similar initiative. Its backers include Blackstone Inc., Goldman Sachs and Hellman & Friedman, who plan to deploy AI tools across their own investment portfolios.

Both companies are trying to prove that their technology can deliver value inside large organisations even as they move closer to potential public listings.

OpenAI is on track for about $30 billion in annual revenue this year, and at the same time, heavy spending on infrastructure could push losses as high as $14 billion.

The company behind ChatGPT will first invest $500 million in the venture, known internally as DeployCo, with the new business expected to reach a $10 billion valuation in a funding round due to close in early May.

DeployCo has been set up as a Delaware-listed limited liability company. Its main role will be to speed up the use of OpenAI’s workplace products as the company launches harder into the business market.

Financial Times reported that OpenAI will hold super-voting shares in the venture, giving it stronger control over key decisions.

The report also noted that OpenAI has the option to invest another $1 billion later. At the same time, investors including TPG, Bain Capital, Advent International, Brookfield and Goanna Capital are expected to provide a further $4 billion.

Private equity backers are said to be investing for five years, with OpenAI guaranteeing them an annual return of 17.5%.

If completed, the deal would rank among the biggest partnerships between an artificial intelligence company and private equity investors.

OpenAI has recently increased its focus on corporate customers. It is looking to grow products such as ChatGPT Enterprise and other workplace tools beyond its consumer business.

That effort comes as competition increases in the business market. Anthropic, maker of Claude, is widely seen as having gained stronger traction with corporate clients. The company has also benefited from partnerships with Amazon and Google Cloud.

Microsoft is another major competitor through its deep ties with OpenAI and the integration of OpenAI models into Microsoft 365 and Azure. Google is also expanding its Gemini products for enterprise users.

Private equity firms are valuable partners because many of them own or influence large companies and can impact software spending decisions across their portfolios.

Reuters said it could not immediately verify the Financial Times report. OpenAI, TPG, Bain Capital, Advent International, Brookfield and Goanna Capital did not immediately respond to requests for comment.

The proposed joint venture could help OpenAI open a new revenue stream outside ChatGPT subscriptions, while expanding artificial intelligence tools across sectors such as finance, healthcare and manufacturing.

This makes Deepgram IBM’s first voice partner, providing real-time transcription and voice features for enterprise clients.

The integration is designed to improve how companies handle complex audio environments, including background noise, accents, and natural conversation.

It also supports a wide range of languages and regional dialects, including multiple Arabic and Indian variants. Users will gain access to real-time captioning, natural-sounding voices, and options for custom tuning.

These tools can be applied across sectors such as healthcare and finance, supporting automated customer care, call analysis, and voice-driven data entry.

Scott Stephenson, Deepgram CEO and co-founder, said, “Voice is rapidly becoming the default interface between humans and technology, and enterprise deployments require a real-time platform that is accurate, low latency, and reliable at scale. 

“By embedding Deepgram inside watsonx Orchestrate Agent Builder, IBM clients can build voice agents and voice-enabled workflows on top of a real-time foundation that has been developed and refined over more than a decade.”

Nick Holda, vice president of AI Technology Partnerships at IBM, added, “Our watsonx Orchestrate integration powered by Deepgram APIs introduces new speech recognition and transcription capabilities to IBM clients, refining and modernizing their operations. 

“This collaboration aims to help enterprise organizations accelerate their AI initiatives and reinforces IBM’s open ecosystem, bringing choice and cutting-edge voice technology to partners and customers.”

The partnership is expected to strengthen IBM’s ability to provide flexible voice solutions to enterprise clients while expanding Deepgram’s reach to new customers through a trusted platform.

Deepgram provides real-time speech-to-text, text-to-speech, and full speech-to-speech features through cloud or on-premises APIs.

It has processed over 50,000 years of audio and transcribed more than one trillion words. IBM, on the other hand, provides hybrid cloud, AI, and consulting solutions to clients in over 175 countries.

The update focuses on one issue, which is prompt injection. In these attacks, hidden instructions are placed inside content an AI system reads. The goal is to mislead the system into revealing sensitive data or taking actions it should not take.

Lockdown Mode is optional and built for a small group of users who face higher security risks, including executives and security teams in major organisations. Most users will not need it.

When switched on, Lockdown Mode restricts how ChatGPT interacts with external systems. It disables certain tools that attackers could exploit to extract data from conversations or connected applications.

Web browsing is limited to cached content. No live network requests leave OpenAI’s controlled network. Some features are disabled entirely where the company says it cannot guarantee data safety.

The setting is now available for ChatGPT Enterprise, ChatGPT Edu, ChatGPT for Healthcare and ChatGPT for Teachers. Workspace administrators can enable it by creating a new role in Workspace Settings. Once activated, it adds tighter limits on top of existing controls.

At the same time, administrators keep granular control. They decide which apps remain available in Lockdown Mode and what actions users can take within those apps. Separately, the Compliance API Logs Platform provides visibility into app usage, shared data and connected sources.

OpenAI said it plans to make Lockdown Mode available to consumer users in the coming months.

The OWASP GenAI Security Project has classified prompt injection as a top vulnerability for large language models. It noted that malicious prompts can alter AI behaviour in unintended ways, even when they appear harmless.

Google’s security team has warned about indirect prompt injections, where hidden instructions are embedded in emails or documents. An AI system may access those sources and leak sensitive data without the user knowing.

Attackers have embedded instructions inside webpages or retrieved documents, causing AI systems to carry out harmful actions. In one case involving Gemini in Google Translate’s Gemini Mode, researchers showed how translation functions could be bypassed to generate dangerous content.

Anthropic recently published findings on prompt injection failure rates. It reported that even advanced models could be breached in certain contexts. In GUI-based systems with extended reasoning enabled, attack success rates exceeded 50% after repeated attempts.

Security researchers have also identified newer forms of attack, including Logic-Layer Prompt Control Injection, which targets deeper parts of AI systems such as persistent memory and retrieval logic.

By restricting live network access and disabling high-risk tools, Lockdown Mode addresses common attack surfaces linked to prompt injection, including browsing and connected apps.

The company said the feature builds on existing safeguards such as sandboxing, monitoring, enforcement, role-based access and audit logs, while adding stricter limits.

Alongside this, OpenAI has standardised “Elevated Risk” labels across ChatGPT, ChatGPT Atlas and Codex. These labels mark features that may introduce additional risk, particularly those involving network access.

For instance, in Codex, developers can grant network access so the system can retrieve documentation or perform actions on the web.

Where that access is enabled, the interface now displays an “Elevated Risk” label explaining what changes, what risks may arise and when such access may be appropriate.

The company said it will continue to review which features carry the label. As security protections improve and risks are reduced, it plans to remove the label from features considered safe for general use.

Now among the world’s most valuable private technology firms, the company confirmed the round on Thursday, saying investors including D. E. Shaw Ventures, ICONIQ and MGX co-led the deal. 

Microsoft and Nvidia also took part, adding to their existing investments. Singapore’s sovereign wealth fund GIC and Coatue Management were among the lead backers in what the company described as its Series G round. 

Other investors included Founders Fund, Qatar Investment Authority, Accel, General Catalyst and Jane Street.

With this latest raise, Anthropic’s total funding since it was founded now exceeds $57 billion. The Series G deal ranks among the largest private technology financings on record, second only to OpenAI’s $40 billion raise in 2025.

Anthropic said its annualised revenue has reached $14 billion. Its coding-focused product, Claude Code, accounts for more than $2.5 billion of that figure. The company said revenue from Claude Code has more than doubled since the start of 2026.

Business demand is growing. Subscriptions to Claude Code from companies have quadrupled this year. Enterprise clients now generate more than half of the product’s revenue, according to the company.

Anthropic has built much of its strategy around tools for developers and office workers. Its Claude Cowork agent carries out computer-based tasks for white-collar staff. 

The release of plugins for the agent unsettled parts of the software market, as investors weighed the possible impact of automation on traditional software providers.

The funding places Anthropic closer to its main competitor, OpenAI. In January OpenAI was in talks with SoftBank Group to raise as much as $30 billion more, in a deal that could value the company at about $830 billion.

Microsoft and Nvidia have now backed both companies, strengthening their positions as key suppliers of computing power to the artificial intelligence sector. Anthropic also counts Google and Amazon among its earlier supporters.

On regulation, Anthropic has taken a different line from many technology firms. The company has pledged $20 million to support U.S. political candidates who favour stronger oversight of artificial intelligence.

Earlier on Thursday, the company said: “The companies building AI have a responsibility to help ensure the technology serves the public good, not just their own interests.”

Chief executive Dario Amodei repeated that position at the World Economic Forum in Davos in January 2026, where he said artificial intelligence companies must ensure their technology benefits society as a whole.

Blackstone, the world’s largest alternative asset manager, is also increasing its stake in Anthropic to about $1 billion, Reuters reported earlier this week.

The deal, announced Thursday, aims to sharpen how Andela identifies and deploys engineers who can deliver on AI projects at scale.

“With Woven, Andela is leapfrogging the development of world-class assessments for both AI fluency and engineering fundamentals,” said Barun Singh, Andela’s chief product and technology officer.

The acquisition will enable Andela to meet the high demand for AI-native engineers, professionals who not only experiment with AI but also build, integrate, and scale AI systems. 

Andela describes three critical archetypes for enterprises, including builders, integrators, and scalers. Builders turn business needs into functional AI components. Integrators connect models, data, and tools into autonomous workflows. 

Scalers manage reliability, governance, and risk in deployed AI systems. Woven’s technology will now allow Andela to assess engineers accurately across all three roles, helping companies hire the right talent for each stage of AI adoption.

“To power the AI ecosystem at scale, the world needs AI-native, enterprise-ready engineering talent en masse. Andela plus Woven equals the best technical assessment engine in the world to ensure AI fluency and real-world job success,” said Carrol Chang, Andela’s CEO.

As part of the integration, Woven’s founder and CEO, Wes Winham Winler, will join Andela to lead next-generation assessments focused on AI-assisted software development and AI system creation. 

The company will also gain access to Woven’s library of engineering scenarios, AI-driven scoring systems, and the team’s domain expertise to accelerate its roadmap.

Andela’s platform already connects more than 150,000 technology professionals worldwide. With Woven’s capabilities built on top of Andela’s previously acquired Qualified platform, the company now has a unified system for scalable, high-accuracy engineering assessments.

This is the third major Nigerian startup acquisition this month. Earlier in January, Flutterwave bought Mono, an open banking startup, while Paystack acquired Ladder Microfinance Bank, revealing a trend of consolidation and expansion in Nigeria’s tech sector.

Andela wants to become the global hub for AI-native engineering talent, combining deep assessment strength with a large, verified talent network to ensure engineers are both skilled and ready for real-world AI challenges.

The talks are described as “very fluid,” and no final decision has been made.

The potential investment comes as competition within the tech industry increases over AI computing power. OpenAI has already signed multi-billion-dollar agreements with Nvidia, Oracle, and Microsoft, as well as a $38-billion cloud services deal with Amazon in November.

“OpenAI plans to use Amazon’s Trainium chips,” The Information reported, which compete with Nvidia and Google’s chips. The report also revealed that Amazon’s funding could lead to a fundraising round with other investors. 

OpenAI is exploring selling an enterprise version of ChatGPT to Amazon, though it is murky if this would include integration with Amazon’s own shopping features.

The talks stress OpenAI’s transition from its non-profit roots. The company restructured in 2023 as a public benefit corporation controlled by a non-profit with a financial stake, removing limitations to raising capital and securing computing resources. Microsoft now holds a 27% stake in OpenAI and has exclusive rights to sell its models to Azure customers.

OpenAI is also preparing for an initial public offering in 2026 that could value the company at up to $1 trillion, Reuters reported in October, potentially making it one of the largest tech listings ever. 

Reports say an Amazon investment could complicate the current Microsoft arrangement and may lead to multi-cloud distribution of OpenAI models.

Despite the surge in AI investment, investors are being careful. The global AI market is expected to surpass $300 billion annually by 2027, but profitability and the sustainability of massive infrastructure spending are still concerns.

OpenAI, Amazon, and Microsoft did not respond to requests for comment.

This follows reports that CEO Sam Altman declared a “code red” inside the company in early December, halting side projects and pulling teams into a faster development sprint. 

The urgency was linked to Google’s latest innovations, which had placed Gemini 3 at the top of key performance rankings across reasoning, coding and multimodal tasks.

OpenAI says GPT-5.2 brings stronger general intelligence, better coding results, and far longer context handling. The company believes these improvements will help users complete more demanding work, particularly tasks that involve spreadsheets, complex documents, and project-heavy workflows. 

Interestingly, the new model stretches to handle up to a million tokens, a big difference from the previous model.

Google has been keen to highlight what Gemini 3 is capable of across text, audio, images and video, and analysts say its tight integration with Workspace and Android gives it an advantage with corporate users. 

Even with that, Altman played down the internal panic when he spoke on CNBC, saying: “Gemini 3 has had less of an impact on our metrics than we feared.” Google has not responded to requests for comment.

OpenAI is rolling out GPT-5.2 in three versions: Instant for quick responses, Thinking for slower but more reasoned answers, and Pro for enterprise-level performance. Paid ChatGPT users will receive them first. The company also states it will continue to support GPT-5.1, GPT-5 and GPT-4.1 on its API, giving developers more flexibility.

Away from the technical competition, OpenAI is also moving into entertainment. Disney has confirmed a $1 billion investment in the company and will allow its Sora video generator to use characters and worlds from Star Wars, Pixar and Marvel. 

This is one of the largest licensing deals yet between Hollywood and an AI firm, and it sets up OpenAI as a direct partner in digital content production. Microsoft, still OpenAI’s biggest backer with about $13 billion committed since 2019, continues to host the company’s models on Azure.

Industry forecasts show spending on cloud-based AI services is expected to rise sharply, with Gartner estimating it will exceed $723 billion next year. Many companies are already relying on GPT models for coding assistance, document processing and data insights. According to OpenAI, enterprise usage has climbed roughly 40% in the past year.

However, regulators in the US and Europe are examining safety standards, competition risks and copyright issues, with Disney’s licensing deal likely to draw even closer attention.

The new platform, developed in collaboration with NVIDIA, focuses on securing AI models, workloads, and agentic applications used in enterprise environments, without compromising performance.

The company confirmed that AI Cloud Protect is now available for on-premises enterprise use and has been validated on NVIDIA RTX PRO Servers. 

Built on NVIDIA’s BlueField technology, it enables organisations to secure AI model development and inference workloads both in data centres and cloud environments.

“As enterprises race to build AI-driven innovation, they can’t afford blind spots,” said Nataly Kremer, Chief Product Officer at Check Point. “With NVIDIA, we’re making AI factories secure by design—protecting models, data, and infrastructure without slowing innovation.”

The rise of AI has exposed enterprises to new and complex risks. According to Check Point data, one in every 80 generative AI prompts reveals sensitive information, while Gartner reports that nearly one-third of organisations suffered an AI-related security incident in the past year, ranging from prompt manipulation to infrastructure attacks.

AI Cloud Protect was built to address these vulnerabilities head-on. Running on NVIDIA BlueField-3 DPUs, it delivers full-stack protection without consuming CPU or GPU resources, putting an end to the common trade-off between security and performance. 

Its network-level defence prevents unauthorised access, data poisoning, and model exfiltration. At the host level, it leverages NVIDIA’s DOCA Argus framework for direct memory access, offering full visibility into active processes on AI nodes to detect and block malicious workloads, even within downloaded large language models.

“Security is essential for the next generation of AI infrastructure,” said David Reber, chief security officer at NVIDIA. “NVIDIA is working with Check Point to integrate BlueField acceleration and the NVIDIA DOCA Argus runtime security framework into the AI Cloud Protect platform to help enterprises deploy AI confidently.”

The solution will also extend to NVIDIA’s upcoming BlueField-4 DPU, which promises six times more compute power and double the network throughput, setting the stage for faster and more scalable AI operations.

Beyond AI Cloud Protect, Check Point’s AI security portfolio includes CloudGuard Web Application Firewall (WAF) with Lakera integration, designed to block prompt injection and jailbreak threats in enterprise AI applications. 

Another solution, Infinity GenAI Protect, helps organisations monitor generative AI usage, apply policies, and prevent sensitive data exposure across teams. Together, these solutions aim to deliver end-to-end protection across the AI ecosystem, from infrastructure to user level.

The company is currently piloting AI Cloud Protect with select financial services firms and partners like World Wide Technology (WWT), focusing on protecting data centres supporting large language model development.

“As enterprises build AI server factories at scale, the combination of Check Point’s AI Cloud Protect and NVIDIA BlueField acceleration delivers enterprise-grade protection for sensitive AI workloads from model training to inference without compromising the performance modern AI applications demand,” said Chris Konrad, vice president, Global Cyber, WWT.

Check Point is addressing data leakage, model manipulation, and infrastructure threats, while enabling organisations to innovate confidently in this phase of intelligent computing.