The company said on Monday that the update will let Copilot’s Researcher agent draw responses from both OpenAI’s GPT and Anthropic’s Claude models within a single workflow. Until now, it relied on one model for each task.

With the new “Critique” feature, one model handles the initial draft while another reviews and refines it before the final output is produced. 

Microsoft said this setup is designed to improve accuracy and reduce errors that can appear in AI-generated responses.

Microsoft is also adding a feature called “Model Council”, which lets users compare answers from different AI systems side by side. This gives a clearer view of where responses match or differ, and helps users decide which output to rely on.

The company says the changes are aimed at improving speed, quality and overall productivity for users working on complex tasks.

At the same time, Microsoft is expanding access to its Copilot Cowork tool, which is still in early release under its Frontier programme. The tool is built for longer, multi-step tasks. Users can describe what they want done, and the system creates a plan, works through connected steps and shows progress along the way.

Copilot Cowork can also carry out routine work such as scheduling, preparing briefings and handling repeat tasks like monthly reviews. Microsoft said early users are already applying it to planning, document creation and preparation for executive meetings.

The company has been focused on strengthening its Copilot platform as competition grows from other AI tools, including Google’s Gemini and Anthropic’s Claude-based systems.

Within that surge, the drive between Google and Microsoft has become one to pull focus on, not just for technology leadership, but for how artificial intelligence (AI) turns into profitable business, especially with committed spending.

Two Companies, Two Directions

Even before you look deeper, you’d notice both companies are building similar systems, but you’d see the difference in how those systems are used.

Google is pushing its models into products people already use every day, including Search, Android, and YouTube. Its Gemini platform has crossed 750 million monthly users, giving it reach that few competitors can match.

Microsoft is taking a different route which is more structured. Its Copilot tools are built into Word, Excel, Teams and other workplace software. The idea is to make businesses pay for productivity.

That difference is where we place our attention. Google has scale, while Microsoft has pricing.

The Competition is Infrastructure

It is easy to focus on apps and chat interfaces, but that is not where the case is being decided.

It is in infrastructure you’d find the competition; data centres, chips, and computing power.

Alphabet, Google’s parent company, plans to spend $175 billion to $185 billion in 2026 alone, largely on servers, networking and AI capacity.

Microsoft is also increasing spending, with its capital expenditure expected to move towards $100 billion or more, driven by demand for cloud and AI services.

This level of investment changes the nature of the industry. AI is not just software, it is capital-intensive, closer to energy or telecoms than traditional tech.

I would put it this way, whoever controls compute, controls the market.

Products: Gemini vs Copilot

The difference in strategy becomes better to grasp at the product level.

Google’s Gemini is built for wide use, sitting inside search results, mobile devices and developer tools. Updates have been frequent, with new versions released through 2025 and early 2026 to improve reasoning and performance.

Microsoft’s Copilot is more targeted, focusing on workplace tasks, writing documents, analysing spreadsheets, and summarising meetings.

But adoption?

Microsoft has around 15 million paid Copilot users, a small share of its Microsoft 365 base of hundreds of millions.

That gap stresses the fact that interest in AI tools is high. Paying for them is still limited.

Cloud: Where the Money Actually Comes From

The revenue engine is behind the scenes. Google Cloud has been expanding, with revenue growth close to 48% year-on-year, driven largely by demand for AI workloads.

Microsoft Azure is however a larger business, with strong growth tied directly to AI usage and enterprise demand.

This is where the competition becomes tougher because companies are not just using AI tools, they are renting computing power to run them.

Cloud turns AI into something billable.

Spending is Increasing Faster Than Returns

There is, nonetheless, an imbalance.

Microsoft is targeting $25 billion in AI-related revenue by 2026, supported by Copilot and Azure services.

Google is already seeing profits in advertising and cloud from its AI rollout.

But both are spending far ahead of what they are earning.

Even within Microsoft’s ecosystem, only a small percentage of users are paying for AI features, despite heavy investment and promotion.

So when does this start paying off?

It is Important to note that Investors are not ignoring the risk.

Google’s decision to increase spending has already triggered mixed reactions in the market, even as its core business stands strong.

Microsoft is facing a different issue, which is adoption. Copilot is growing, but not at a pace that fully justifies the scale of investment yet.

So the market is in a strange position, believing in the long-term potential, but watching the short-term numbers carefully.

Here the Bigger Question Comes

This has gone beyond a competition between two companies. Will the current level of investment produce the kind of productivity being promised?

The comparison with past technology cycles is unavoidable. Large amounts of capital are being deployed ahead of proven returns. That does not automatically mean a bubble, but it does introduce risk.

Right now, demand for computing power is strong, but what we don’t know is whether that demand will remain strong enough to justify the infrastructure being built.

Who is Ahead?

The answer depends on how you measure it.

Google is ahead when it comes to reach. Its products touch billions of users, and its AI systems are already embedded into everyday digital activity.

Microsoft comes top in structure. It has a clearer path to monetisation through enterprise software and cloud services.

Google and Microsoft are strong when it comes to AI, both are spending heavily, but neither has fully solved the same problem, which is turning scale into sustained profit.

So, let’s not look at who builds the better model between Google and Microsoft or who comes top in AI spending, but who can turn artificial intelligence into a reliable business before the cost of building it becomes harder to justify.

That is where this growth will be decided.

What happened instead is a lack of clarity in Microsoft’s rebranding. Over the past week, reports have spread across Reddit, Hacker News and X saying that Microsoft Office has been renamed “Microsoft 365 Copilot”. It sounds believable, but it is also wrong.

Gareth Oystryk, senior director of Marketing, Microsoft 365, clarified: “We have not made any recent naming changes to our Office apps. Word, Excel, and PowerPoint, the Office apps within the Microsoft 365 productivity suite, remain unchanged.

In November 2022, we renamed only the Office “hub” app for web and mobile to the Microsoft 365 app. In January 2025, we updated it to the Microsoft 365 Copilot app to reflect its role in bringing Copilot and Microsoft 365 productivity experiences together in one place.”

The source of the confusion was that Office.com redirected users towards what Microsoft calls the Microsoft 365 Copilot app. 

This is not a new office suite, it’s a hub. One place to open Word, Excel, PowerPoint and Outlook, alongside Copilot features, under a single banner.

Microsoft set this path years ago. In 2022, the company dropped the Office name at the brand level and replaced it with Microsoft 365. The familiar apps never disappeared. What changed was the wrapper around them. 

Then, as Copilot became central to Microsoft’s pitch, the old Office app was renamed again, this time as the Microsoft 365 Copilot app.

That is where the language tripped people up. On Microsoft’s own site, users are greeted with the line: “The Microsoft 365 Copilot app (formerly Office)…”. Read quickly, it looks final, almost like an end notice for Office itself. That single phrase was enough to birth the idea that Office had been fully replaced.

It has not. Word is still Word. Excel still opens spreadsheets. PowerPoint still builds slides. Businesses and individuals can even buy Office 2024 as a standalone package, without cloud tools and without Copilot. That option alone should settle the case.

The timing made things worse. Microsoft has recently pushed the Copilot name harder than ever, positioning it as the front door to productivity. Office.com now reiterates that message. The result is a blurred line between the apps people use every day and the hub Microsoft wants them to launch from.

What surprises me is not that users are confused, but that Microsoft seems surprised by the reaction. So far, the company has not publicly stepped in to clear the air. Silence, in this case, has allowed assumptions to do the talking.

There are risks here. Some users now believe Office no longer exists. Others assume Copilot is mandatory or that familiar tools are being phased out. For businesses, especially cautious ones, that kind of uncertainty slows decisions and feeds distrust.

This episode is about unclear messaging layered on top of years of background rebrands. Microsoft may see Microsoft 365 Copilot as a neat, unified story. Many users see something else entirely, a once-simple product family wrapped in names that no longer explain themselves.

Office, whatever Microsoft chooses to call it at the top level, is still very much alive. The problem is that Microsoft’s own words have made that harder to see.

Unlike earlier Copilot functions that mainly assisted with edits or summaries, Agent Mode takes on multi-step processes, which include generating, testing, and refining outputs until they reach professional quality. 

In Excel, this means the system no longer just suggests formulas, but runs complete analyses, builds visualisations, and even fixes errors without manual intervention. Microsoft claims this is comparable to handing over work to a trained analyst while you remain in control.

Sumit Chauhan, corporate vice president of Microsoft’s Office Product Group, said: “Today we’re bringing vibe working to Microsoft 365 Copilot with Agent Mode in Office apps and Office Agent in Copilot chat. In the same way, vibe coding has transformed software development, the latest reasoning models in Copilot unlock agentic productivity for Office artifacts.”

The new experience extends beyond Excel. In Word, Agent Mode is designed to turn writing into a conversation. A user can ask Copilot to draft a monthly report, highlight insights, or reformat documents according to corporate style guidelines. Instead of just delivering a block of text, the tool asks clarifying questions and proposes refinements, creating what Microsoft calls “vibe writing.”

Office Agent, meanwhile, takes the same approach to PowerPoint and Word but directly from Copilot chat. A simple instruction, such as preparing a deck on consumer trends, prompts the system to clarify the brief, gather external data, draft slides, and show previews before handing over the final presentation. 

Chauhan noted: “It’s not just simple assistive short answers, but board-ready presentations or documents. It’s work, quite frankly, that a first-year consultant would do, delivered in minutes.”

In internal testing using the SpreadsheetBench benchmark, Agent Mode in Excel scored an accuracy rate of 57.2%, higher than competitor tools like Shortcut.ai and Claude Files, though still below human-level accuracy of 71.3%. 

Microsoft stresses that every output is auditable and refreshable, an essential safeguard given Excel’s role in critical business operations.

While OpenAI’s models continue to power Agent Mode in Excel and Word, Anthropic’s models are now responsible for Office Agent in Copilot chat. This dual approach, Microsoft argues, allows it to combine strengths from different model families depending on the task. “We are committed to OpenAI, but we are starting to explore with the model family to understand the strength that different models bring,” Chauhan said.

The features are being released gradually. Agent Mode is now available in the web versions of Word and Excel for Microsoft 365 Copilot subscribers, with desktop support expected soon. Office Agent is live in the United States for personal and family subscribers through the Frontier programme, starting with English-language support.

Productivity tools are no longer limited to editing files, they’re becoming full partners in the work process. Chauhan highlighted, “Productivity is our DNA, we’re Office. While others will try to replicate us, there is no substitute for the real thing.”

Before now, businesses paid $30 per user monthly for Microsoft 365 Copilot, with an additional $20 for access to sales, service, or finance tools. The new arrangement drops the total cost to a flat $30 per user, removing what many saw as pricing limitations to adoption. The tools will be available through the Microsoft 365 Copilot Agent Store.

The change is aimed at simplifying subscriptions and enhancing adoption among enterprises. Microsoft says more than 100 million people already use Copilot, including over 70% of Fortune 500 firms. Analysts estimate Office Copilot alone is generating over $1 billion annually.

Alongside this change, Microsoft has confirmed it will use Anthropic’s Claude Sonnet 4 models within Microsoft 365 Copilot. Tests showed Claude outperformed OpenAI’s GPT-5 in Excel automation, large-scale text processing, and PowerPoint generation. Despite Microsoft’s deep ties to OpenAI, it will pay Anthropic via Amazon Web Services to secure access.

The company is also preparing to launch Agent 365 at its Ignite event. The suite will help businesses oversee AI agents while maintaining security, compliance, and governance standards.

For enterprise users, the integrations mean smoother workflows: sales teams can prepare for meetings with instant CRM insights, service teams can draft customer-ready responses in less time, and finance teams can reconcile data and produce reports with fewer manual steps.

Shaun Worsley, manager, Sales Process & Technology at Sandvik Coromant, said: “The role-based sales solution can transform the way our teams work. Now that it’s part of Microsoft 365 Copilot, we can scale AI across the business much faster to accelerate adoption.”

In bringing everything under one subscription and enlarging its model choices, Microsoft is strengthening its hold in enterprise AI. With Anthropic, OpenAI, Meta, Mistral, and others already integrated into Azure AI Foundry, the company is building a diverse and resilient ecosystem.

This year alone, Microsoft has sunk $64 billion into infrastructure, much of it driving the AI growth through data centres powering services like Copilot in Microsoft 365. 

While most tech firms are cautiously navigating unstable markets, Microsoft’s share price has surged over 30%, a sign investors are betting on the company’s aggressive AI focus.

Behind the scenes, Microsoft appears to be recalibrating its alliance with OpenAI. Though the two remain close, with Microsoft still a strategic backer, OpenAI has been given leeway to partner with others, including Oracle, on the massive Stargate data centre project in Texas.

What’s happening is Microsoft is gradually placing itself as a neutral technology provider, what some are calling an “arms dealer” in the AI wars, rather than locking into exclusive alliances. This neutrality allows it to offer AI tools across industries without being boxed in by one partner’s limitations or priorities.

Meanwhile, demand for AI services in Azure, Microsoft’s cloud platform, is climbing. CEO Satya Nadella has suggested that AI costs can be drastically reduced through algorithmic efficiency. “Once it settles on an algorithm and begins to optimise it, Microsoft can obtain 10 times better performance for the same computing costs,” he said. That’s the kind of return that could redefine tech margins.

The company is also being tactical about how it handles computational surges. Instead of building more expensive data centres, Microsoft is leaning on “neocloud” providers like CoreWeave. These firms specialise in delivering Nvidia-powered AI infrastructure on demand. It’s a leaner, faster, more flexible approach to scaling.

Cantor Fitzgerald analyst Thomas Blakey said: “If they have to flex up in some way, they’ve been consistently saying that they’re going to shift away from buying more data centres and dirt and cement and they’re going to leave that to the neoclouds.”

The Microsoft Build 2025 conference, running until May 22, is not just a developer gathering this year. It’s a moment of clarity about Microsoft’s vision to take over the AI stack, monetise it, and use every tool, from GitHub to Azure, to keep developers building inside its ecosystem.

These tools, which integrate deep reasoning, will roll out in April under the company’s newly launched Frontier program.

Researcher is built to simplify complex, multi-step research tasks by pulling from both internal company data and external sources. Microsoft has built it as a tool that can develop business strategies, conduct competitive market analysis, and generate detailed reports with outstanding depth.

Unlike traditional AI search tools, Researcher is designed to operate seamlessly with external data connectors, including Salesforce, ServiceNow, and Confluence. This means businesses can extract and analyse insights from multiple platforms without switching between applications. 

Microsoft claims Researcher offers a level of analysis that surpasses previous AI tools, making it a valuable asset for professionals handling large-scale research and strategy development.

For organisations dealing with extensive datasets, Analyst offers a solid solution. Built on OpenAI’s o3-mini reasoning model, it mimics the analytical approach of a data scientist, working iteratively to refine its outputs.

Microsoft highlights its ability to convert raw data into actionable insights, including revenue forecasts, customer behaviour trends, and demand predictions.

An interesting feature of Analyst is its ability to run Python scripts, allowing businesses to execute complex data queries while maintaining full visibility into the AI’s processes. Microsoft asserts that this transparency ensures accuracy and reliability, reducing concerns over AI-generated errors.

Alongside Researcher and Analyst, Microsoft has expanded its Copilot Studio platform to support the creation of autonomous AI agents. These agents can now independently perform tasks, initiate workflows, and handle business operations with minimal human intervention.

The new features in Copilot Studio allow IT teams to build and govern AI-driven workflows while maintaining control over data security and access permissions. Microsoft claims this system ensures enterprises can confidently integrate AI without compromising regulatory or security requirements.

Microsoft 365 Copilot customers will gain access to Researcher and Analyst through the Frontier program, an initiative that allows early adopters to test upcoming AI innovations before they reach general availability.

The company acknowledges that AI models, including those behind Researcher and Analyst, are not flawless. There are ongoing challenges in preventing AI from generating inaccurate information or misinterpreting data. However, Microsoft says its approach—grounding AI outputs in enterprise data and allowing user oversight—helps mitigate these risks.

With these developments, Microsoft is embedding AI deeper into enterprise workflows, aiming to bolster how businesses leverage artificial intelligence for research, analysis, and automation. 

After successful testing in select markets, these tools are now being rolled out to the majority of Microsoft’s 84 million consumer subscribers, aiming to boost productivity and creativity globally while simplifying daily tasks and improving design capabilities.

Subscribers can now rely on Copilot to assist with various activities such as organising weekly meal plans, summarising email threads and creating great presentations. 

Similarly, Microsoft Designer enables users to generate and edit visuals, enabling seamless image creation and editing directly within apps like Word and PowerPoint.

Copilot’s Features

The integration of Copilot definitely changes how users engage with Microsoft 365 apps:

• Meal Preparation: Generate recipes focused on dietary preferences in Word.
• Budget Management: Use Excel to analyse and optimise household budgets.
• Presentations: PowerPoint users can input details, and Copilot will draft slides and content.
• Email Summaries: Outlook provides condensed summaries of lengthy email threads.
• Task Organisation: OneNote helps organise freeform notes into structured task lists.

Microsoft Designer for Visuals

Designer, included in the subscription, enables users to:

• Edit images effortlessly, such as removing unwanted objects.
• Generate professional-grade visuals by describing a concept or choosing pre-set ideas.

Subscribers also receive a monthly AI credit allocation to utilise these features, with upgrades available for extensive usage.

Flexibility and Pricing

To support the expanded benefits, Microsoft is increasing subscription prices by $3 per month in the US, with changes effective at renewal for existing users. For those preferring simpler plans, Basic and Classic options remain available, maintaining core features without AI integrations.

Privacy

Microsoft has assured users that data such as prompts and documents used within Copilot will not be used to train AI models, reiterating transparency and privacy.

Access to New Features

These upgrades will be automatically available to existing subscribers upon updating to the latest version, ensuring a seamless transition to the enhanced suite of tools.

With Copilot and Designer now integral to Microsoft 365, users can look forward to a more intuitive and efficient experience across their favourite apps.

“The investment we are making with Microsoft takes our long-standing partnership to the next level, helping our clients accelerate and integrate AI at scale as a catalyst for reinvention,” said Julie Sweet, chair and CEO, Accenture. “The powerful combination of Microsoft Copilot, agentic and generative AI technologies, and Avanade’s and Accenture’s deep industry and functional expertise will help enable enterprises to reimagine their processes and operations, discover innovative ways of working, strengthen their security and open up new pathways to value.”

The Copilot business transformation practice will provide the technology and industry experience clients need to help accelerate and expand their use of Copilot and agents across the enterprise.

The companies will collaborate on AI and Copilot agent templates, extensions, plugins and connectors to help organizations leverage their data and gen AI to reduce costs, improve efficiencies and drive growth.

The companies will help clients leverage Microsoft’s gen AI-powered security capabilities to elevate the efficiency of their security operations and help analysts make better, faster decisions.

The practice consists of 5,000 professionals from Accenture and Avanade, supported by Microsoft product specialists who will work closely with the Accenture Center for Advanced AI.

The team will tap into more than 50,000 Microsoft Copilot-trained professionals across Accenture and Avanade to help deliver industry and functional transformation.

Clients will have access to specialized engineers and early access to the latest technologies, including the Accenture AI Refinery, to create custom models and solutions. Accenture will leverage learnings from its use and deployment of 100,000 Microsoft 365 Copilots and agents, with a commitment to expand to approximately 200,000 users.

“We are pleased to deepen our collaboration with Accenture to help our mutual customers develop AI-first business processes responsibly and securely, while helping them drive market differentiation,” said Judson Althoff, executive vice president and chief commercial officer at Microsoft. “By bringing together Copilots and human ambition, paired with the autonomous capabilities of an agent, we can accelerate AI transformation for organizations across industries and help them realize successful business outcomes through pragmatic innovation.”

Accenture and Avanade have already developed agents to streamline workflows, boost revenue and enhance customer experience.

For example, a supplier discovery and risk agent can deliver real-time market insights, agile supply chain responses and better vendor selection, which could result in up to 15% cost savings.

A procure-to-pay agent could improve efficiency by up to 40% and enhance vendor relations and satisfaction by addressing urgent payment requirements and avoiding disruptions of key services.

They have also invested in building functional and industry Copilot solutions that are delivering client value, including:

Microsoft 365 Copilot

The companies have developed Microsoft 365 Copilot solutions such as a readiness assessment and value accelerator, among others, to help clients effectively implement and successfully adopt and scale the technology for maximum value.

They are helping Repsol, a global multi-energy company, deploy Microsoft 365 Copilot and a four month study has shown impressive results. Employees are saving a gross average of 121 minutes per week, enhancing productivity and the quality of outputs by 16%.

Additionally, 61% favor the AI-enhanced workflows, noting improved confidence and motivation. Productivity gains were most significant in technical and professional roles, with engineering roles saving nearly 97 hours annually per employee.

Maintenance roles saw even greater benefits, saving about 225 minutes weekly, which significantly enhanced the precision and consistency required in their tasks. These results underscore the transformative potential of generative AI in optimizing business operations and boosting employee satisfaction across the board.

Industry and Functional Copilots

Together, the companies have developed Copilots for a variety of industries and functions including finance, manufacturing, supply chain, retail and consumer goods and healthcare.

For example, they developed a Manufacturing Copilot powered by Microsoft Manufacturing Data Solutions in Fabric, to transform the shop floor and the way factory workers operate.

Custom Copilots and Agents

The companies are using Microsoft Copilot Studio to extend foundational Copilots, like Microsoft 365 Copilot and industry and role-based Copilots, with enterprise knowledge to create autonomous agents that act on certain triggers or on the intent of the user to implement AI-powered, responsible and efficient workflows.

Avanade and Accenture helped Bricorama, a French home improvement retailer, create a custom home improvement Copilot called “pAInt.”

The pAInt app uses Azure OpenAI to guide customers through their projects.

Udacity, part of Accenture LearnVantage, and Microsoft have co-created a new Azure Generative AI Engineer Nanodegree program to equip learners with the skills to design, build, and operationalize AI-driven applications on Azure.

In this fully online program, learners will use AI models to solve real-world problems through automation, data insights, and generative AI solutions. Learners will develop custom Microsoft 365 Copilot applications and implement models through hands-on projects with personalized support from industry advisors.

Udacity’s Nanodegree program is available today and, for a limited time, learners can get started with the first lesson for free.

Deeper collaboration

Accenture will continue to support Microsoft in the transformation of its functions – including finance, sales, marketing and HR – with Copilot, leveraging its industry, functional and generative AI consulting services.

Microsoft also selected Accenture to establish four Microsoft Co-Innovation Labs to incubate blueprints for AI-led design and experimentation, with immersive showcases within Accenture offices and Microsoft Co-Innovation Labs.

At the labs, the companies will build pilots, leveraging assets developed by the Copilot business transformation practice, to inspire reinvention.

Having seen the value of Microsoft Copilot in its business, Accenture will leverage the latest Copilot agentic technologies to facilitate AI enabled augmentation and help transform clients’ business processes across corporate functions and industry-specific operations.

[Source]

These autonomous agents, which will function with minimal human oversight, are termed “apps for an AI-driven world” and are expected to automate key business processes such as customer service, sales lead identification, and inventory management.

Microsoft’s new offering will be accessible through its Copilot Studio application, which is designed to be user-friendly, requiring little technical expertise to create and deploy the AI agents. The agents will leverage a mix of proprietary AI models and those developed by OpenAI, increasing the scope of possibilities for businesses eager to simplify operations.

As part of its initiative, Microsoft will also roll out 10 pre-built AI agents which will handle common business tasks, such as managing supply chains, tracking expenses, and enhancing client communications. 

McKinsey & Co., a consulting firm with early access to these tools, showcased how an AI agent could effectively manage client inquiries, track interaction history, assign tasks to the right consultant, and schedule follow-up meetings. 

The technology will bolster businesses, as Microsoft envisions a future where every employee has a personal AI assistant, or “Copilot,” to interact with the various autonomous agents within their organisation. 

Charles Lamanna, corporate vice president of business and industry Copilot at Microsoft, explained that the Copilot will serve as a personal interface for employees, providing them with a tailored experience while interacting with an array of AI tools to boost productivity.

This latest development comes despite a slight dip in Microsoft’s stock performance for the September quarter, the company is focused on the long-term prospects of its AI technologies.

Microsoft’s announcement follows Meta’s unveiling of its “Self-Taught Evaluator” model, which focuses on creating AI that can train and evaluate itself with minimal human involvement. More autonomous systems capable of self-learning are being built, potentially reducing reliance on human-generated data and expert intervention.