Forrester predicts only 15% of firms will use agentic AI by 2026

The automation industry stands at a pivotal moment. For decades, businesses have relied on deterministic automation—rule-based systems that execute predefined tasks with mechanical precision. These systems excel at repetitive processes like data entry, invoice processing, and routine customer service responses, but they lack the ability to adapt when circumstances change.

Now, artificial intelligence is fundamentally reshaping this landscape. Agentic AI systems can reason through complex scenarios, make contextual decisions, and adapt their behavior based on changing conditions. Unlike traditional automation that follows rigid scripts, these AI agents can analyze situations, weigh options, and choose the best course of action—much like a human employee would.

This transformation represents more than just a technology upgrade. It signals a shift from automation that simply executes tasks to automation that can think, reason, and solve problems independently. According to Forrester Research, a leading technology research firm, this evolution will define the automation landscape through 2026 and beyond.

However, the transition won’t be smooth. Organizations face significant challenges in governance, return on investment, and technical complexity as they navigate this new terrain. Here are three key predictions that will shape how businesses approach automation over the next two years.

1. Less than 15% of firms will activate agentic features in automation platforms

The automation industry is splitting into two distinct approaches, and most companies will stick with the familiar path despite vendor pressure to embrace more advanced capabilities.

The first approach embeds AI assistants within existing workflows, creating what Forrester calls “agentish” systems. These operate like sophisticated helpers—they can handle variations in data formats, respond to simple questions, or flag unusual patterns, but they still follow predetermined processes. Think of a customer service system that can understand different ways customers phrase complaints but still routes them through the same escalation procedures.

The second approach puts AI reasoning at the center, allowing agents to dynamically decide how to complete work. These truly agentic systems can break down complex goals into smaller tasks, choose which tools to use, and adapt their approach based on results. For example, an agentic system tasked with “resolve this customer issue” might decide whether to process a refund, schedule a service call, or escalate to management based on the specific situation.

Despite the promise of these advanced capabilities, most organizations will hesitate to flip the switch. Traditional process automation vendors are undergoing massive transformations of their core products while simultaneously repositioning their brands around AI. This creates uncertainty about platform stability and long-term support.

More critically, businesses struggle with fundamental questions: How do you measure the return on investment for a system that makes autonomous decisions? How do you maintain compliance when an AI agent chooses its own path through a process? These governance challenges will keep most organizations running familiar, predictable automation systems through 2026, even as vendors heavily promote their new agentic features.

2. Strategic robot innovation will unlock 20% of new enterprise use cases

While software-based AI agents capture headlines, physical robots are quietly becoming more capable and easier to integrate into business operations. This convergence of smarter robots and digital workflows will dramatically expand what companies can automate.

Traditional industrial robots required extensive programming for specific tasks and couldn’t adapt when conditions changed. Modern robots leverage machine learning to handle variations in their environment, learn new tasks through demonstration, and coordinate with software systems. A warehouse robot can now learn to handle different package sizes, communicate with inventory management software, and adjust its behavior based on real-time demand patterns.

This flexibility creates what researchers call an expanded “automatability surface”—the range of work that becomes feasible to automate. Manufacturing companies can automate quality inspection processes that previously required human judgment. Hospitals can deploy robots that navigate complex environments while coordinating with electronic health records. Retail operations can integrate robotic fulfillment with customer service systems.

However, this expansion comes with complexity. When smart robots work alongside AI software agents in shared environments, coordinating their activities becomes exponentially more challenging. A customer order might trigger a software agent to process payment, a warehouse robot to pick items, and a delivery system to schedule shipment—all requiring seamless coordination.

This complexity drives convergence across previously separate automation tools and platforms. Companies need unified architectures that can orchestrate diverse automation technologies to deliver value across complete business processes, not just individual tasks.

3. Process intelligence will rescue 30% of failed AI projects

Process intelligence—technology that analyzes how work actually flows through an organization—sits at a crucial intersection in automation’s evolution, yet most vendors have barely scratched the surface of its potential.

Currently, process intelligence tools primarily serve as diagnostic instruments. They analyze system logs, user interactions, and data flows to create visual maps of how processes actually work versus how they’re supposed to work. Business analysts use these insights to identify bottlenecks, compliance gaps, and improvement opportunities.

However, the real opportunity lies in transforming these insights into active guidance for AI agents. Instead of just showing humans where processes break down, process intelligence can provide AI systems with real-time context about normal versus unusual situations, compliance requirements, and operational constraints.

Consider an AI agent handling insurance claims. Without process intelligence, it might approve a claim that technically meets policy requirements but violates unwritten business rules learned through years of experience. With process intelligence integration, the agent understands that similar claims historically require additional documentation, that this customer segment has higher fraud rates, or that claims of this type typically take longer to resolve.

Most process intelligence vendors have approached AI cautiously, offering incremental improvements like chatbot interfaces for asking questions about process data. The vendors that will thrive are those that embed their insights directly into AI reasoning engines, creating feedback loops between process understanding and automated decision-making.

This integration becomes particularly valuable for rescuing AI projects that fail due to lack of context. Many agentic AI implementations struggle because they operate in isolation from organizational knowledge about how work actually gets done. Process intelligence provides the missing link between AI capabilities and business reality.

The path forward

The convergence of agentic AI, advanced robotics, and process intelligence represents a fundamental shift in how organizations think about automation. Rather than simply replacing human tasks with software, this new paradigm creates collaborative environments where AI agents, robots, and humans work together on complex business challenges.

Success in this environment requires more than just deploying new technologies. Organizations must develop new approaches to governance, risk management, and performance measurement that account for autonomous decision-making. They need architectures that can coordinate diverse automation technologies while maintaining visibility and control.

The companies that navigate this transition successfully will gain significant competitive advantages through more flexible, responsive, and intelligent operations. Those that cling too tightly to familiar approaches risk falling behind as the automation landscape fundamentally reshapes around them.