Sustained interest and experimentation in AI will support learning and steady progress in 2025. Generative AI (genAI) and edge intelligence will drive robotics projects that will combine cognitive and physical automation, for example. Citizen developers will start to build genAI-infused automation apps, leveraging their domain expertise.
Despite obvious benefits, however, implementation challenges will hinder 2025 gains. AI agents will find moderate success in employee support applications. GenAI’s ability to create autonomous, unstructured workflow patterns and adapt to the dynamic andunpredictable nature of real-world processes will have to wait.
Key developments in 2025 include the following:
• One-quarter of robotics projects will work to combine cognitive and physical automation. Physical robots continue to expand beyond the high-volume and low-variance tasks they’ve tackled in the automotive sector and related industries since the 1960s. Innovations around genAI and edge intelligence are encouraging developers of physical robots (and adjacent technologies like autonomousvehicles) to take a fresh look at embodied AI.
This will enable robots to sense andrespond to their environment instead of following preprogrammed rules andworkflows, exposing them to more complex and unpredictable situations. Decision-makers in asset-intensive industries will leverage the combination of cognitive and physical automation to enhance their operational efficiencies.
• Citizen developers will deliver 30% of GenAI-infused automation apps.
As the citizen developer train continues to roll, the vast majority of firms either have or areworking on a citizen developer strategy to empower workers outside of IT to deliver apps. Our ongoing research has uncovered another clear pattern:
Citizen developers outside of IT have the necessary inspiration and domain expertise to imagine what a genAI solution might look like, effectively prompt LLMs, and infuse results into useful applications. This means citizen development is the most practical path to the scaled experimentation genAI requires.
Our latest data showsthe market is also beginning to see this: Director- and higher-level developers most selected AI-infused applications as the apps that their citizen developers would be allowed to build with low-code tools. Our conclusion: Citizen developers will deliver a large ratio (e.g., 30%) of genAI-infused automation apps in 2025.
• Implementation challenges will stall 25% of agentic and AI agent efforts.
Unlike traditional automation, which relies on well-defined processes, genAI-based solutions introduce new conceptual and technical complexities. Vague businessobjectives and premature integration in decision-making will create confusion.
Determining the optimal level of autonomy to balance risk and efficiency will challenge business leaders. Integrating human oversight and ensuring reliable access to enterprise data for AI agent training are additional hurdles.
Furthermore, a fragmented vendor landscape, characterized by overlapping features, will complicate genAI platform selection. Vendors will hastily rebrand offerings to align with AI trends without providing substantial value-added capabilities. To effectively leverage AI agents, enterprises need to reevaluate processes designed for humaninteraction and replace outdated technologies with alternatives that support AI-driven automation.
• GenAI will orchestrate less than 1% of core business processes.
GenAI will affect process design, development, and data integration, reducing design anddevelopment time as well as the need for desktop and mobile interfaces. Businessusers will develop initial workflows, create forms, and visualize the process.
But, this genAI efficiency still leaves current digital and robotic process automation platforms orchestrating the core process, subject to their deterministic and rule-driven models. To prepare for 2025, recognize that deterministic automation will remain in control of the core long-running process, while AI models will support bursts of insight and efficiency.
• Over half of successful genAI projects will be for employee support.
Despite AI agent enthusiasm for customer and consumer support, employee support for operational processes will have the most success. Financial and healthcare sectors will be most aggressive. AI agents offer potential but with caveats.
AI agent implementation won’t be a cakewalk: Companies must map the multiple sources of enterprise and customer data (which is likely siloed) to simultaneously drive AI agent decision-making, develop integrations with key business and technology systems that AI agents need to access, and shepherd employees along a change management journey.
Summary
Unwavering interest and ongoing research will fuel advancements in AI in 2025. AI will continue to progress in robotics, citizen development, and AI agents for employee support. To prepare for 2025, decision-makers should balance AI innovation with the scale and reliability of traditional automation tools and methods.
— Forrester Research, USA & Australia.