As IT resource demands spike, new research from Mulesoft finds that 93% of enterprise IT leaders have implemented or plan to implement AI agents in the next two years.
However, integration challenges hinder companies from fully realising the technology’s potential to create a limitless digital workforce, which can significantly alleviate IT workloads.
The findings, gleaned from a survey of 1,050 enterprise IT leaders, reveal that 95% of respondents struggle to integrate data across systems. Only 29% of applications are typically connected within organisations, on average, impacting the accuracy and usefulness of AI agents.
“As limitless digital labour, AI agents are set to transform enterprises. We see forecasts of new levels of efficiency and productivity with the unified human and agent workforce,” says Linda Saunders, Salesforce director: solutions engineering Africa. “However, to unlock this new capacity, integration and APIs are critical to building an agent-ready foundation.
“Agents depend on unified data and enterprise actions, powered by integration and automation, to provide accurate, trusted responses and perform complex, multi-step tasks.”
Key findings from MuleSoft’s 10th annual Connectivity Benchmark Report include:
Increased IT demand allows AI agents to improve efficiency, lower cost
IT teams already responsible for maintaining current systems are struggling to meet new demand for AI solutions.
- The percentage of projects not delivered on time has risen to 29% in 2024 – up from 26% in 2023.
- What’s more, 86% of IT leaders expect workloads to rise in the future. On average, surveyed leaders expect an 18% increase in projects from 2024 to 2025.
- To meet increased demand, enterprise IT leaders spent an estimated average of $16,9-million on IT staff in 2024, a massive 61% increase from 2023 (estimated at $10,5-million).
AI agents and integration solutions present an opportunity to improve productivity, increase efficiency, and lower costs by taking action on behalf of overwhelmed IT teams.
- Most IT leaders (93%) say that AI will increase the productivity of their developers in the next three years. And those using AI agents are even more confident, with 98% agreeing.
- Around two-fifths (39%) of IT’s team time, on average, is spent designing, building, and testing new custom integrations between systems and data to enable digital capabilities and solutions for the business — something that can be automated with the right integration solution.
Complex infrastructures hinder AI agent efficacy
The growing number of applications and AI models used by enterprises creates data silos that can limit agents’ ability to operate effectively.
- On average, 90% of IT leaders say data silos are creating business challenges in their organisation. The challenge is greater at organisations using agents (94%) compared to those not using agents (87%).
- Organisations today are using an average of 897 applications. Organisations with AI agents use even more applications (1 103) – 45% more than those that don’t use agents (762).
- The estimated number of AI models used by organisations has doubled (from an average of nine in 2024 to 18 in 2025). Organisations using agents are leveraging an average of 22 AI models, while those not using agents are leveraging 15 AI models.
APIs, connectivity can enhance IT and AI agent performance
By aligning integration and AI strategies, organisations can simplify and unify their data infrastructure, unlocking the full potential of their data to efficiently power AI agents. Using these tools, agents can access critical, business-specific data to function optimally and extend across the enterprise to interact directly with existing systems and automations.
- Among organisations with agents, those using APIs are taking advantage of their capabilities to: improve IT infrastructure (56%); enable data sharing across teams (49%); and integrate disparate systems (46%)
- On average, 50% of an organization’s internal software assets and components are available for internal reuse. This is an opportunity for companies to leverage their reliable data and existing investments with APIs, instead of building from scratch to create trusted AI agents.
APIs drive broader business value
In addition to enabling AI agents to access services, databases, and functionalities provided by applications, IT leaders also highlight the significant business value of application programming interfaces (APIs) – from increasing efficiencies to impacting revenue.
- Organisations using APIs report the top benefits as: increased productivity and speed to meet business needs (49%); increased employee engagement and collaboration (47%); and increased innovation (46%) and teams managing their own IT needs (46%)
- More than a third (37%) of IT teams are leveraging APIs to increase the efficiency of their application development process.
- IT leaders estimate an average of 40% of their company’s revenue is generated from APIs and API-related implementations. The number is even higher (45%) at organisations using agents.