Microsoft Intros ROI Tracking for AI Agents, Expands Copilot in Forms
Microsoft is presenting brand-new capabilities focused on revealing organizations whether artificial intelligence representatives provide quantifiable value while expanding how Copilot can create and examine studies.
The company recently revealed ROI for agents in Microsoft Foundry, a personal preview capability created to link a representative’s operating expense with company results. Independently, Microsoft revealed an updated Copilot experience in Microsoft Forms that brings the familiar Microsoft 365 Copilot chat user interface straight into types and quizzes.
Together, the new functions show Microsoft’s broader effort to move business AI beyond isolated triggers and speculative implementations. The Foundry is intended primarily at designers and IT leaders responsible for examining representatives, while the Types update brings more agent-like actions into an application used by business users.
Microsoft said ROI for representatives “translates the cost of running an agent into the business worth it produces.” The feature tracks metrics such as job completion rates, time saved, and cost efficiency, showing operational expenses and service results together in the Foundry website or through an API.
Organizations can compare various agent versions, screen day-to-day trends, and examine private traces related to bad results. Microsoft stated the information can assist stakeholders “validate investment and prioritize what to improve next.”
The capability resolves a growing obstacle for companies moving AI representatives into production. Although designers can track token use, latency, errors, and other technical measures, those signals do not always reveal whether an agent conserves employees time or finishes enough work to require its cost.
ROI for representatives constructs on Microsoft Foundry’s tracing, examination, and keeping an eye on abilities. The company introduced the function at Build 2026 in June and said it is offered in private preview. Microsoft has not announced when it will become broadly available.
On the efficiency side, Microsoft is including what it called “seamless integration with a Microsoft 365 Copilot chat experience” to Microsoft Forms.
Users will see a Copilot button in the lower-right corner when producing or opening a kind or test. Opening the chat pane grounds Copilot in the present kind, permitting it to examine questions, recommend structural changes, modify settings, and evaluate sent responses.
For example, users can ask Copilot to recognize missing out on concerns, reorganize areas, or flag settings that contravene a survey’s designated purpose. Copilot can also make bulk modifications, such as replacing a placeholder throughout a type or marking multiple concerns as needed.
The upgraded experience adds assistance for standard branching logic, customized thank-you messages, closing dates, and follow-up concerns about survey outcomes. Microsoft said the analysis functions are meant to offer “clear insights and actionable takeaways,” although users need to continue evaluating branching setups before dispersing a kind.
Existing abilities, consisting of Draft with Copilot, concern rewriting, and quiz response explanations, will remain offered.
The new Copilot experience in Kinds is rolling out worldwide to customers with Microsoft 365 Copilot industrial licenses.