Microsoft shared results on November 12 from IDC research into the business value of artificial intelligence. The study, sponsored by Microsoft, is based on a survey conducted in July 2026 among 5,204 senior technology and business leaders across 28 industries. The data compare AI adoption, expansion of use, measurement of results and the organization of responsibilities within businesses.
The main distinction in the published results is between obtaining benefits and reproducing them across more parts of a business. According to Microsoft's presentation, 66% of respondents say AI has delivered substantial or transformative results. The proportion scaling that value across multiple business units is 48%. These percentages describe different indicators of adoption rather than the same measure.

Who responded and how advanced firms are defined
The research uses the category “Frontier Firms” for organizations in the top 10% of IDC's AI maturity scale. The slow-adopter group represents the bottom 30%. This classification belongs to the study and supports comparison between levels of maturity; it is not a certification awarded to companies, a legal category or a measure of business size.
Respondents are senior leaders in IT and in business or operational functions. The results reflect their answers about their organizations. The November publication presents findings from the July survey, separating the collection period from the communication date. Microsoft's article says that the full IDC InfoBrief will follow, rather than presenting the complete research document at this stage.
IDC's maturity scale combines investment discipline, breadth of use, agent adoption, skills, confidence in managing risks and outcomes. The comparison between advanced firms and slow adopters is built from this combination of dimensions. A high score concerns the organization of AI use as a whole, rather than simply holding a subscription or access to a particular model.
Usage and value measurement are separate indicators
On measurement, 61% of surveyed leaders report high confidence in their ability to measure return on AI investment. Among Frontier Firms, that proportion reaches 93%. The indicator concerns confidence in measurement, rather than the percentage of profitable projects. It cannot be equated with the number of active users or cost reductions achieved within an operational process.
Microsoft cites an average reported ROI of four times investment among Frontier Firms, compared with 2.7 times among slow adopters. Within those two groups, 39% and 4% of respondents respectively report returns of at least five times investment. These figures remain outcomes declared in research sponsored by the company and shared through its official presentation.
Separately from returns, the research describes adoption breadth: organizations use AI in an average of 6.5 business areas, according to the data presented. IT infrastructure, cybersecurity and customer service are among the leading areas. The average number of areas indicates spread within the organization, separately from ROI and the percentage of projects that are profitable.
Different IDC survey indicators presented by Microsoft
| Indicator | Published result | What it describes |
|---|---|---|
| Substantial or transformative AI results | 66% | Benefits declared by respondents |
| Value scaled across multiple units | 48% | Extent of adoption across the organization |
| High confidence in measuring ROI | 61% | Perceived measurement capability |
| Comprehensive AI skilling strategy | 20% | Existence of an organization-wide plan |
AI agents introduce an organizational dimension
Another part of the research concerns AI agents, systems that can perform tasks through tools and workflows. According to the results presented, 36% of organizations are piloting such systems. Piloting describes an implementation and evaluation stage, separate from widespread operational use within a business or the coordination of several agents within the same business process.
Among Frontier Firms, 96% of respondents indicate that they are deploying agents at scale, while 53% mention orchestration. These percentages refer to IDC's advanced group, rather than all 5,204 responses considered together. The change in reference population is essential when the figures are compared or used in a presentation describing the wider AI market.
Microsoft also presents a reported return for agents of 3.8 times investment among advanced firms, against three times among slow adopters. This indicator adds to the comparison between groups while retaining its character as a declared outcome. It does not specify the cost of an application for an individual business or its investment payback period.
In integrating AI into business processes, the difference between an available tool and an operational workflow is concrete: the application receives data, applies rules and produces an output used by people or other systems. That progression from isolated use to connected processes is directly relevant to the scaling examined by the research.
Governance and training develop unevenly
Across the general results, 65% of organizations report governance processes for AI agents, while 60% declare high confidence in security and governance. The existence of a process and confidence in that process are separate indicators. The presentation does not turn them into an individual technical audit of the applications used by each respondent's organization.
Differences between groups appear here too: 87% of Frontier Firms report agent governance, compared with 51% of slow adopters. The research places these practices alongside the scale of deployment. It describes how businesses in the compared groups say they organize usage, rather than a list of controls directly verified within each system in the study.
On employees, 67% of respondents indicate workforce trust in AI, while only 20% report a comprehensive, organization-wide skilling strategy. Trust and training plans therefore have different shares in the published data. People accepting a tool does not, by itself, describe the skills available to use that tool effectively within professional tasks or more complex business workflows.
The comparison of training strategies between groups is pronounced: 80% of advanced firms report an organization-wide AI skilling strategy, against 3% of slow adopters. The group definitions remain the same for this comparison. These percentages do not represent the proportion of individually trained employees across all businesses or the number of training hours completed by each person within those organizations.
What the publication adds to business decisions
Platform selection is assessed through multiple criteria. In Microsoft's presentation, trust and security, advanced technology and model performance are each cited by 44% of respondents as leading criteria. The percentages are responses to different criteria, rather than shares of a total of 100%. Cost is considered alongside accuracy, security, compliance and the ability to scale deployment.
For businesses preparing investments, the study offers a comparison of practices and leader-reported results, with the period, population and sponsorship identified. The data currently available are those presented by Microsoft. Publication of the full IDC InfoBrief is announced for later, defining the stage at which the results and methodology are being communicated to the public.