Artificial intelligence is rapidly becoming a key driver of productivity, product quality and business transformation, with companies increasingly deploying AI agents across development, operations and customer-facing activities, according to Simona Almajan, Country Manager of NXP Semiconductors Romania.

Speaking at the Workplace of the Future conference organized by The Diplomat-Bucharest, she outlined how organizations are adopting different forms of AI, ranging from generative and agentic AI to edge and physical AI applications.

“When we talk about generative AI or agentic AI, we are referring to agents that companies implement to become more efficient,” Almajan said. “This type of AI focuses on optimizing development workflows, streamlining operations and improving daily activities across the organization, helping teams become more productive, more efficient and ultimately deliver higher-quality products and results.”

She distinguished between enterprise-focused AI applications and the AI technologies embedded in products themselves.

“We also talk about edge AI, which is a more technical and tangible form of AI that exists directly in chips,” she explained. “These are the models that process and classify data and the tools that analyze information coming from the cloud.”

Almajan noted that in sectors such as automotive, AI-powered decision-making is increasingly moving closer to the device.

“In a vehicle, the system must decide whether to stop or react to an object that appears in its surroundings and could represent a danger,” she said. “The data and models required for that decision can be stored in the cloud, but they can also reside directly on the target device or on the chip itself. This is what we refer to as physical AI.”

According to Simona Almăjan, AI agents are becoming relevant across virtually all industries and business functions.

“These agents can be implemented in any company, regardless of whether it develops physical products or delivers services,” she said. “AI can be applied everywhere, from customer interactions and product promotion to more advanced use cases involving customized AI agents connected to enterprise platforms and services.”

He described AI-powered customer engagement as one of the most significant opportunities for future business growth.

“This is part of what we see as the future of product delivery and customer relationship management,” she said. “Organizations are looking for new ways to make information, services and expertise available to customers through intelligent, customized interfaces.”

Beyond customer-facing applications, AI is also reshaping internal business processes, particularly in areas such as human resources, sales and supply chain management.

“For years, companies have invested in dashboards and reporting systems,” she said. “With these tools and AI agents, we can extract data much faster, with higher quality and greater diversity of insights. However, these agents must be programmed and structured properly. It is not enough to simply prompt them; they need to be designed so that they consistently deliver the same result every time they are executed.”

Almajan also highlighted the growing role of AI in engineering and product development, where companies are evaluating opportunities to automate previous manual processes.

“Organizations are analyzing existing workflows step by step and identifying areas that still rely on manual checklists,” she said. “Many of these activities were never automated, even during the DevOps era. Today, companies see an opportunity to introduce AI agents into those workflows.”

Software development is among the areas expected to benefit significantly from AI assistance, he added.

“We already have agents capable of writing code, and software developers will increasingly use these tools to accelerate development,” Almajan said. “The quality of the output depends on how effectively the system is guided, but the code must still be verified, especially in automotive and other industries where functional safety and cybersecurity requirements apply.”

Maintaining trust and compliance will remain essential as AI-generated outputs become more common, he warned.

“We must continue to comply with standards that require verification and trust,” Almajan said. “These AI agents themselves will ultimately need to be trustworthy. If code is generated automatically rather than written by humans, the challenge becomes ensuring that the entire process meets certification requirements and provides the confidence people need to use a vehicle or any other technology safely.”

She concluded that while AI will accelerate innovation and automation, trust, validation and regulatory compliance will remain critical components of future AI-enabled systems.