Microsoft is satisfying the agents

Microsoft He has The largest ecosy system is builtIt is now providing it with strong new capabilities that put the company in front of it in one of the most exciting Enterprise Tech sectors.

The company announced on Tuesday evening two important additions to its Copilot Studio platform: the possibilities of deep thinking that enable the agents to address complex problems through precise and methodological thinking and the agent’s flows that combine artificial intelligence flexibility with automation of the inevitable business process. Microsoft has also unveiled the specialized deepest thinking factors for Microsoft 365 Copilot: Researcher and Analyst.

“We have clients with thousands of agents already,” Microsoft Vice President of Microsoft Copilot Charles Lamanna told Venturebeat in an exclusive interview on Monday. “You start getting this type of workforce as agents, regardless of the job, you may have an agent that can help you accomplish it faster.”

Microsoft’s distinguished analyst for Microsoft

While the researcher’s agent reflects the potential of competitors such as Deep Opays and deep Google research, the Microsoft analyst’s agent represents a more distinctive offer. The analyst’s agent, designed to work like a personal data world, processing various data sources, including Excel and CSVS files and streams integrated into documents, and generating visions through the implementation and perception of software instructions.

“This is not an essential model outside the shelf,” “This is a large part of the extensions, control and training at the top of the basic models.” Microsoft has benefited from its deep understanding of the Excel Work and data analysis patterns to create an agent in line with how the institution’s users have already operated with data.

Analyst can create a Python icon automatically to process uploaded data files, produce perceptions, and provide commercial visions without the need for technical experience from users. This makes it particularly value for financial analysis, budget prediction and operation of operating reports use cases that usually require intensive data.

Deep thinking: Bring critical thinking to the agents of the Foundation

The ability of Microsoft’s deep thinking expands the capabilities of agents beyond completing the simple task to complex judgment and analytical work. By combining advanced thinking models such as Openai’s O1 and connecting them to institutions’ data, these agents can systematically address mysterious work problems.

The system defines a dynamic when deep thinking is called, either based on the complexity of the task or explicitly when users include claims such as “reason for this” or “think very about this”. Behind the scenes, the statute analyzes the instructions, evaluates the context, and chooses the right tools based on the requirements of the task.

This allows the scenarios that was previously difficult to automate. For example, a large communication company in the field of deep communication uses deep thinking agents to create complex RFP responses by collecting information from all internal documents and knowledge sources. Likewise, Thompson Reuters employs these capabilities for due care in integration and acquisition reviews, and to process unveiled documents to identify ideas. Watch an example of a logical customer working in the video below:

https://www.youtube.com/watch?

The agent flows: re -visualize the automation of the operation

Microsoft has also introduced the agent’s flows, which effectively develops automated automation (RPA) by combining the workflow -based workflow tasks and with artificial intelligence thinking. This addresses the customer’s demands to integrate the inevitable logic of business with the possibilities of flexible artificial intelligence.

“Sometimes they do not want the model to free. They do not want artificial intelligence to make its own decisions. They want to have highly coded work rules.” “At other times they want the agent free and make calls.”

This hybrid approach allows scenarios such as the prevention of smart fraud, as the agent’s flow may use a conditional logic to direct the higher value requests to the artificial intelligence agent for a deep analysis against policy documents.

Pets in Home, a pet Dying Company in the UK, has already published this technology to prevent fraud. Lamana revealed that the company had saved “more than a million pounds” through implementation. Likewise, Dow Chemical realized “millions of dollars provided for transportation and shipping” through the improvement of the agent.

Below is a video clip showing the flow of the agent at work:

https://www.youtube.com/watch?

Microsoft graph feature

It is essential in the Microsoft agent strategy in the integration of institutions data through Microsoft Graph, a comprehensive mapping fee for relationships in the workplace between people, documents, email messages, calendar events and business data. This provides agents with the contextual awareness that general models lack.

“The less well -known secret capacity of Microsoft’s graphic drawing is that we are able to improve the importance of the chart based on participation and the extent to which some files are delivered.” The system defines the documents that are referred to, shared or commenting on, ensuring that agents refer to reliable sources instead of old copies.

This Microsoft approach gives a great competitive advantage to independent artificial intelligence providers. Although competitors may offer advanced models, Microsoft combines them with the context of the workplace and explicitly improvement explicitly for cases of institutions and microsoft tools.

Microsoft can benefit from the same web data and models technology that competitors can, that Lamanna has noticed, “but we also have all the content within the institution.” This creates the effect of the budget wheel as it enriches every new interaction of the agent to understand the graph of the workplace patterns.

The adoption of the institution and the ease of access it

Lamana said that Microsoft has given priority to making these strong capabilities within the reach of various technical resources. The agents are displayed directly within Copilot, allowing users to interact through natural language without quick engineering experience.

Meanwhile, Copilot Studio provides a low -code environment to develop the dedicated agent. “It is in our DNA tool for everyone, not just people who can run Python SDK and make calls, but anyone can start building these agents,” confirmed Lamana.

The accessibility approach has fueled this rapid adoption. Microsoft previously revealed that more than 100,000 organizations have used Copilot Studio And that More than 400,000 agents have been created in the last quarter.

Competitive scene

While Microsoft appears to be the publication of the institution’s agent today, the competition is intensifying. Google has expanded its Gemini capabilities For agents and Coding agentwhile The Openai’s O1 and Agents SDK model provides strong thinking and developers agent tools. Large institutions applications such as Salesforce, Oracle, Servicenow, SAP and other platforms have launched all agents for their customers during the past year. Also on Tuesday, Amazon’s Aws, Amnesty International’s agent, called Amazon Q in Quicksight, issued to allow employees to engage in natural language to conduct data analysis without specialized skills.

Employees can use the natural language to conduct data analysis at the level of experts, ask questions to questions, obtain implementable recommendations, and help them open new visions and make decisions faster

However, the Microsoft feature lies in its most comprehensive approach-a strong conjugation with the leading thinking model, Openai, with the selection of models as well, the infrastructure at the level of institutions, the integration of broad data through the workplace tools, and focus on business results instead of the raw capabilities of AI. Microsoft has created an environmental system similar to best practices by combining subscribers who understand individual work patterns with specialized agents for specific commercial operations.

For decision makers in the institution, the message is clear: the agent technique has matured beyond experimenting with practical business applications with a measurable investment. The selection of the statute is increasingly dependent on integration with the current tools and data. In this field, Microsoft has a feature in many app areas due to the number of users, for example, at Excel and Power Automation.

Watch my full interview with Charles Lamana, which is below to hear how Microsoft leads the agent’s strategy, what these new capabilities mean to the institution’s users, and how institutions benefit from agents to provide measuable business results:

https://www.youtube.com/watch?