Software integration is becoming a larger operational issue as organizations connect hundreds of applications, APIs, automation workflows, and AI capabilities. The available research points to a consistent pattern: integration is widespread, but connectivity, governance, and collaboration still lag behind demand. Postman’s 2025 State of the API Report surveyed more than 5,700 developers, architects, and executives globally, while MuleSoft’s 2026 Connectivity Benchmark Report surveyed 1,050 IT leaders.
Contents
- The scale of modern integration
- API adoption and organizational maturity
- Connectivity gaps and integration challenges
- Automation across business teams
- Integration tools, gateways, and protocols
- AI, APIs, and agentic transformation
The scale of modern integration
Enterprise software environments are large before new AI systems are added. MuleSoft’s 2026 Connectivity Benchmark Report found that organizations managed an average of 957 applications. Organizations that had fully adopted agentic transformation managed an average of 1,057 applications, 10% above the general mean.
The application count was also high in the MuleSoft 2024 Connectivity Benchmark Report, which described more than 900 applications in the average enterprise. These figures describe application estates rather than the number of successful connections between systems, so they should not be read as a measure of integration quality.
The number of applications is only one part of the integration workload. In the 2025 State of the API Report, 69% of API practitioners said they spent at least 10 hours per week on API-related tasks. Within that group, 43% spent 10 to 20 hours weekly and 26% spent more than 20 hours. The remaining 31% spent less than 10 hours per week.
API work is embedded in everyday engineering activity. In the same Postman report, 81% of respondents used APIs for testing, 73% used them for developing, and 58% used them for documenting. Respondents were primarily connected to software delivery: 73% worked in engineering or software development, while 22% were principals or tech leads. Executives, vice presidents, and directors accounted for 8%, including 4% at the C-level.
API adoption and organizational maturity
API-first development has become common, but complete adoption remains less typical. Postman reported that 82% of organizations had adopted some level of an API-first approach. The breakdown was 57% somewhat API-first, 25% fully API-first, and 18% not at all API-first. The fully API-first share was up 12% from 2024.
API programs are also connected to commercial outcomes. In the 2025 State of the API Report, 65% of organizations said they generated revenue from their API programs. That result places APIs in both an engineering and business context: they may support internal integration, partner connectivity, or products and services that generate income.
API teams are usually small. Eighty-four percent of API teams consisted of one to nine people, according to Postman. At the same time, 93% of API teams faced collaboration blockers. The combination suggests that integration responsibility often sits with compact teams while the systems they support span many departments and applications.
API adoption also has a regional dimension in the Postman respondent base. Thirty percent of respondents were in North America, 7% were in Latin America, 20% were in Europe, the Middle East, and Africa, and 43% were in Asia-Pacific. These figures describe the report’s respondent distribution, not the global distribution of all API teams.
Connectivity gaps and integration challenges
The most direct connectivity measure in the 2026 Connectivity Benchmark Report was the share of applications connected. Only 27% of applications were connected on average. Among organizations that had achieved agentic transformation, the connected share reached 32%.
This leaves a substantial difference between the size of an application estate and the number of applications connected to the broader digital environment. In the 2024 benchmark, only 28% of applications were integrated, despite the average enterprise having more than 900 applications. The two measurements come from different report years and should be treated as separate snapshots rather than a calculated trend.
Integration can also create dependency concerns. Seventy-one percent of respondents to MuleSoft’s 2026 benchmark agreed that their IT infrastructure made systems overly dependent on one another. Eighty-two percent of IT leaders identified data integration as one of their biggest challenges when using AI.
The earlier MuleSoft 2024 Connectivity Benchmark Report found that 81% of respondents saw integration challenges as roadblocks to digital transformation. Among public-sector respondents, the share was 86%. Sixty-five percent of public-sector respondents had some form of API strategy.
Public-sector delivery pressures were also quantified in the 2024 benchmark. Nineteen percent of public-sector respondents expected budget cuts in 2024, compared with 6% overall. Forty-two percent said project demands were rising. Only 29% of public-sector respondents said they delivered on all projects, compared with 47% overall.
Legacy evidence provides additional context but is not independently verified here. A MuleSoft and Simplus summary of 2020 benchmark coverage reported that 85% of digital transformation initiatives were hindered by integration problems. Software AG’s APIs and Integration Report 2022 surveyed 1,150 IT executives, while its Annual APIs and Integration Report 2021 surveyed 950 IT executives. A DZone trend report on software integration reported that 38% of respondents spent time managing API integrations, and a Software AG hybrid integration survey reported that 64% of organizations were already running integration in a hybrid environment.
Automation across business teams
Automation is moving beyond isolated technical workflows. The Workato 2024 Work Automation Index analyzed 1,055 companies and 82,000 automated processes. It found that 44% of all automated processes were built outside IT, showing that integration work is increasingly distributed among business teams.
Enterprise automation was used across multiple departments by 74% of teams, while 52% used it in four or more business teams. One measure of expansion showed that the share of companies with automation in five or more departments rose from 19% one year earlier to 37%.
Business operations teams automated 27.7% of all processes in 2023, and data operations accounted for 24% of automated processes. Customer support and operations grew 226% in automated processes in 2023. These figures describe process automation activity, not the proportion of employees or applications automated.
| Automation measure | Reported result |
|---|---|
| Processes with generative AI growth in 2023 | 400% |
| Generative AI endpoint growth in 2023 | 500% |
| Revenue operations share of generative AI processes | 48% |
| IT share of generative AI use cases | 31% |
| Automated processes that were complex or highly complex | 61% |
| Automated processes with humans in the loop in 2023 | 11% |
Generative AI activity grew rapidly in the Workato index: processes with generative AI grew by 400% in 2023, while generative AI endpoints grew by 500%. Revenue operations was the top adopter, representing 48% of generative AI processes; IT was second with 31% of use cases. The index also found that 61% of automated processes were complex or highly complex, and the share classified as highly complex increased from 13% to 24% over two years.
Integration tools, gateways, and protocols
The Postman 2025 State of the API Report measured several layers of the integration toolchain. For continuous integration and delivery, 54% of organizations used GitHub Actions, 34% used AWS DevOps, and 29% used Azure DevOps.
Monitoring coverage was varied. Grafana was used by 36% of organizations, Sentry by 20%, and Elastic by 20%. Seventeen percent used no monitoring tools at all. The monitoring figures indicate reported tool use and do not establish whether monitoring was comprehensive.
API gateway use was led by AWS API Gateway at 47%, followed by Azure API Gateway at 26%. Another 23% used other gateway solutions. Thirty-one percent of organizations used multiple API gateways: 20% used two gateways and 11% used three or more.
Teams also used multiple ways to connect systems and services. Half of teams used webhooks, 35% used WebSockets, and 33% used GraphQL. These percentages describe the reported use of particular technologies and may overlap because organizations can use more than one integration method.
AI, APIs, and agentic transformation
AI is creating new integration requirements while exposing existing governance gaps. In Postman’s 2025 report, 89% of developers said they used AI in daily work. Yet only 24% of developers designed APIs for AI agents. Sixty percent still designed APIs primarily for humans, 13% designed equally for humans and AI agents, and 7% primarily designed for AI agents or machine consumption. Another 5% were actively transitioning from human-first to AI-first API design, while 16% had not yet considered AI agents as API consumers.
Postman recorded 7.53 million calls to AI APIs in the prior 12 months, an increase of 40% year over year. At the same time, 51% of developers cited unauthorized or excessive API calls from AI agents as their top security concern.
Awareness and regular use of emerging protocols were not the same. Seventy percent of developers were aware of the Model Context Protocol, but only 10% used MCP regularly in daily work. Twenty-four percent planned to explore MCP.
MuleSoft’s 2026 benchmark measured a later and broader organizational view of agentic capabilities. Ninety-eight percent of respondents said their organization planned to adopt agentic capabilities, and 88% were moving toward agentic transformation. Forty percent used Agent-to-Agent protocols and 39% used MCP. Fifty-four percent had a centralized governance framework for agentic capabilities, while 27% of APIs remained ungoverned.
The transition is not frictionless. Sixty-eight percent of IT leaders said it was challenging to stay current with emerging standards and protocols for AI agents. The same report found that 82% of IT leaders considered data integration one of their biggest AI challenges. Together, these findings connect AI adoption with practical integration work: organizations are planning agentic capabilities while still addressing connectivity, governance, and standards.