93% of IT leaders will implement AI agents in the next two years

Businesses are seeing greater adoption of agentic AI, enhancing their ability to deliver greater value at the speed of need by leveraging digital labor across all lines of business. 

According to the 2025 Connectivity Benchmark Report by MuleSoft and Deloitte Digital, 93% of IT leaders report intentions to introduce autonomous AI agents within the next two years, and nearly half have already done so. 

Also: AI agents will match ‘good mid-level’ engineers this year, says Mark Zuckerberg

MuleSoft’s 10th annual Connectivity Benchmark Report is based on survey data from interviews with 1,050 IT leaders across the globe. Here is the executive summary:

• Data is trapped across siloed enterprise apps: The average number of apps used by respondents is 897, with 45% reporting using 1,000 applications or more — further hindering IT teams’ ability to build a unified experience. Only 29% of enterprise apps are integrated and share information across the business.  
• IT workload is increasing: With IT leaders being pushed to incorporate AI, 86% of respondents anticipate their teams’ workload will rise in the next year, all while still maintaining existing systems, including 70% who report governing enterprise-wide automations.  
• Integrated user experience is elusive: Enterprise-wide automation is on the rise; even so, 66% of respondents still don’t provide an integrated user experience across their channels.  
• Integration is a challenge for accelerated adoption of AI agents: Nearly all (95%) of IT leaders report 95% integration as a hurdle to implementing AI effectively, but APIs have helped. Fifty-five percent of IT leaders report APIs improve their IT infrastructure, while 45% recognize the ability for APIs to support enhancing user experiences.  

<!–>

–>

Here are my top 20 big takeaways that every CIO and IT leader must review and self-assess from the report:

<!–>

AI adoption surpass projections

• Adoption of autonomous agents is a key business priority: Introducing autonomous agents within the next two years is on the roadmap for 93% of IT leaders; nearly half have already done so.
• Productivity gains drive agent adoption: The vast majority (93%) feel that AI will increase developer productivity over the next three years, which is up seven percentage points since last year’s report.
• More AI models will be deployed to drive agent adoption: The average number of AI models estimated to be used doubled from 2024 (9 to 18), and IT leaders predict a further increase of 78% in the next three years to an average of 32 models.
• IT budgets are growing in 2025: As demand for AI grows, so does the budget: 85% of IT decision-makers expect an increase in their overall budget in 2025, while 11% anticipate that their IT budgets will stay the same.
• Investments in data infrastructure are 4X AI investments: To prepare for the expanded use of AI, enterprise CIOs are allocating 25% of their budgets to data infrastructure and management (compared to 5% to AI).
• IT workload will increase significantly in 2025: Along with an increased budget, 86% of IT leaders anticipate workloads to rise as well. On average, they expect an 18% increase in the number of projects their organization will need to deliver. To meet this growing demand, IT staffing budgets are also set to rise significantly, with an expected increase of 61.9% compared to last year’s predicted amount, with an estimated budget nearing $17M ($16.91M).
• Project delivery capabilities will decline due to higher demand from IT: The percentage of projects not being delivered on time has risen to 29% from 26% last year. Can AI help? This is one area in which IT leaders are highly optimistic about AI’s potential to boost productivity, with 93% expecting it to enhance developer productivity over the next three years, up from 85% in last year’s survey.
• The primary obstacle to AI implementation – integration and legacy debt: Integration remains the most significant barrier to AI implementation, with 95% of organizations facing challenges when integrating AI into their existing processes. Cybersecurity and data privacy are paramount concerns for IT leaders during AI integration.  

Disconnected data strains IT resources

• Disconnected data remains an overwhelming blocker to legacy modernization for the majority of organizations: Today, 83% of organizations report that integration challenges are a significant barrier to their legacy modernization efforts. And 97% of IT leaders acknowledge that their organizations struggle with integrating end-user experiences.
• User experience across all channels is not managed: Around two-thirds (66%) of respondents don’t provide an integrated user experience across all of their channels.Integration challenges prevent optimal user experiences – Integration challenges are hindering digital transformation at their organization for 83% of respondents.
• Connections are key to improved experiences: Around half (49%) report improved return on investment when creating a connected end-user experience.
• The top challenges to providing a seamless customer experience are: 1. adopting AI tools (44%), 2. reusing software components to create new products and services (42%), 3. leveraging APIs (37%), 4. implementing microservice architecture (36%), and 5. adopting event-driven APIs (35%).  

–> <!–>

Rising demand for AI agents fuel the need for robust integration

• The number of apps in the enterprise is growing: Organizations are leveraging a staggering number of applications, with an average of 897. Remarkably, 46% of these organizations report using 1,000 applications or more. Concerningly, only 2% of IT leaders report their organizations have integrated more than half of their applications.
• Customer experiences are mostly digital: With an estimated 71% of all customer engagements occurring in a digital format, the need for application integration is key to success.
• Data silos are innovation roadblocks: Only 10% of respondents report experiencing no challenges due to data silos, while 74% of organizations find their IT systems to be overly interdependent.
• Data mobility is the biggest integration challenge: Moving data from source systems to the data warehouse is the top challenge, followed by correlating data in the warehouse to deliver insights and reusing data sources across different user-facing applications.
• IT sees a rising need for integration across all departments: According to IT leaders, at least 80% or more of data science, engineering, sales, and customer service departments are reporting a need for integration.
• IT teams face new hurdle with widespread AI: A vast majority (81%) of IT leaders report that their companies are grappling with challenges in leveraging AI for system integrations. The average application lifetime within organizations is five years, highlighting the need for robust and adaptable systems that can integrate new technologies – without compromising performance or security.
• AI needs data to deliver results: Without integration, AI does not have access to the critical data for it to function at its best. Eighty-one percent of respondents identified data integration as one of the most significant challenges their organizations face when implementing AI for systems integrations. Outdated infrastructure is a significant obstacle, with 41% of participants reporting that their organization’s old IT architecture and infrastructure hinder the use of data for AI applications.  

The 2025 Connectivity Benchmark Report also highlighted the importance of automation, with central IT teams governing 70% of automation. A well-rounded automation strategy is considered essential for effectively integrating AI into an organization.

Also: The future of sales? These AI agents offer 24/7 ABC energy for SMBs

Automation can expand beyond IT delivery teams – 65% of organizations have developed a complete or nearly complete strategy to empower non-technical users with low-code and no-code solutions. The demand for automation is abundantly clear, with 98% of IT leaders highlighting their need for automation within their organization.

To learn more about the 2025 Connectivity Benchmark Report, you can visit here.