Microsoft pitches Copilot cost advantage against AI rivals

Microsoft says its Copilot assistant costs businesses substantially less than competing products from OpenAI and Anthropic, intensifying competition over corporate spending on artificial intelligence.

Judson Althoff, chief executive of Microsoft’s commercial business, made the case at the company’s Ignite conference in San Francisco, arguing that customers could obtain comparable capabilities at a lower operating cost. His comments followed the publication of a Microsoft-funded comparison of enterprise AI subscriptions and usage charges.

The study, conducted with technology research company GigaOm, estimated that an employee using AI for everyday office tasks would cost $30 a month with Copilot, against $81 for ChatGPT Enterprise and $205 for Claude Enterprise. Those figures included estimated usage charges rather than subscription prices alone.

For more demanding users deploying autonomous AI agents, the estimated monthly bills rose to $96 for Copilot, $149 for ChatGPT and $299 for Claude. Microsoft described these differences as evidence that its pricing model was better suited to companies introducing AI across their workforces.

Microsoft said Copilot was 84% cheaper than Claude and 63% cheaper than ChatGPT for everyday users, based on the study’s assumptions. The corresponding savings for advanced users were 68% and 35%, respectively. The company also maintained that the products delivered broadly comparable quality across the tasks evaluated.

However, the figures are not an independent measurement of what all corporate customers pay. Microsoft commissioned and funded the research, performed the testing and cost analysis, and worked with GigaOm on the methodology. The estimates used October list prices or assumed prices where published rates were unavailable, excluding negotiated discounts.

Anthropic has challenged the picture suggested by the comparison, saying its Claude Enterprise customers who do not use coding tools spend less than $75 a month on average. That figure is substantially below the $205 monthly estimate assigned to Claude’s everyday users in Microsoft’s model, although the underlying workloads may differ.

The study modelled 100 monthly tasks for a typical office user and 140 for an advanced user, including 20 agent workflows. It assessed activities such as finding information, preparing meetings, analysing data and producing presentations. Responses were checked against task-specific criteria, with incomplete answers treated as failures.

Microsoft’s central argument rests on the way its software is billed. Routine AI interactions and work within Word, Excel and PowerPoint are covered by the Copilot subscription, while more complex agent activity attracts consumption-based charges. Rival enterprise offerings can combine seat fees with additional charges linked to model usage.

The company also pointed to its ability to direct requests automatically towards different AI models, varying computing effort according to the complexity of a task. Its Work IQ system retrieves information from Microsoft 365 applications and files, potentially reducing repeated searches and the volume of data processed.

Neither mechanism guarantees lower costs for every organisation. Actual expenditure depends on usage patterns, contracted prices, available integrations and the kind of work employees assign to AI systems. Microsoft itself advised customers to evaluate their own workloads rather than treating the comparison as a universal price ranking.

The pressure to contain spending extends beyond these three suppliers. A Capgemini survey published this month found that nearly six in ten executives at large companies planned to increase use of lower-cost, open-weight AI models over the following year. Such models offer another alternative for businesses reassessing their corporate budgets.

The pricing campaign comes as Anthropic and OpenAI expand further into workplace applications traditionally dominated by Microsoft’s Office suite. Anthropic has introduced spreadsheet and other productivity capabilities, while Microsoft is integrating fuller Word, Excel and PowerPoint functions directly into Copilot.

Jacob Andreou, Microsoft’s executive vice-president responsible for Copilot, argued at Ignite that competitors’ lighter document and spreadsheet tools could compromise functionality and collaboration. Microsoft’s pitch combines its established productivity applications with access to AI models developed both internally and by outside laboratories.

That creates an unusual commercial relationship: Microsoft competes with OpenAI and Anthropic for workplace AI customers while also distributing models from both companies. The company is developing its own specialist models, seeking more flexibility over performance and the cost of providing AI services.

Alongside its pricing claims, Microsoft introduced additional tools for controlling enterprise spending. Agent 365 Business Analytics is designed to connect AI expenditure with work completed, while Cost Management allows administrators to track consumption, set limits and receive alerts across users, agents and models.



Notice an issue?

Arabian Post strives to deliver the most accurate and reliable information to its readers. If you believe you have identified an error or inconsistency in this article, please don't hesitate to contact our editorial team at editor[at]thearabianpost[dot]com. We are committed to promptly addressing any concerns and ensuring the highest level of journalistic integrity.


Loading next story…