Businesses are endlessly seeking to increase efficiency, reduce costs and increase productivity. Some of the earliest known businesses – the ancient Mesopotamian merchants – inspired the invention of writing. (Record save– Now this is a competitive advantage!)
Similar demand exists in every economic period. The biggest difference now is that AI technology can improve these efficiencies in new and exponentially profitable ways. Agent AI is the core of this efficiency improvement.
Dan Priest, Chief AI Officer of PricewaterhouseCoopers, said: “Agent AI refers to an AI system that can independently perceive, decide and act within a defined scope, and can cooperate with humans, systems or other agents.” (PwC, also known as PricewaterhouseCoopers, is the “big four” – the world’s four largest professional service companies)
Also: According to business leaders, 5 ways to become an excellent AI agent manager
Agent AI systems are different from the previous generation of algorithmic business management systems that have been in use for decades. Agent AI can understand the context, respond to changing situations without running from scripts, and work towards autonomous definition goals.
Compared to traditional automation (and some human managers), proxy AI systems can be flexible, handle ambiguity and make informed decisions at the speed of business operations. “Helping organizations operate at greater speed, intelligence and scalability fundamentally changes how they work and decision-making,” Priest said.
Common obstacles to AI integration
However, you can’t simply wave your wand and get the perfect work of enterprise-wide proxy AI. There are many challenges, including the existing technical debt that deeply constitutes legacy tools and processes, aversion to change, regulatory challenges, and a lack of understanding and technical AI skills within an organization.
“Common obstacles to implementing integrated proxy systems include fragmented data environments, lack of interoperability among tools, and siloed organizational structures,” said PwC’s AI experts.
Ironically, the implementation process itself may hinder successful AI adoption. Many companies first follow IT best practices: implementing new systems in small increments. Unfortunately, the most useful AI systems thrives on cross-organizational information, so step-by-step approaches often lead to dispersion, inefficiency and counterattack among stakeholders.
“Updating these challenges requires not only technological upgrades, but also cultural and operational shifts to allow for cross-functional consistency and scalable integration,” explains Priest. “In addition, concerns about security, compliance and governance may slow adoption, especially in regulated industries.”
Also: AI complexity paradox: higher productivity, more responsibility
To successfully deploy and gain the scope of agent AI enterprise, managers need to reevaluate business processes, develop cross-functional coordination strategies, obtain a complete executive-level buy-in, and promote cultural change throughout the organization.
The key role of proof of concept in proxy AI
It is natural for the machine to be managers to initially be reluctant to give up on human processes. However, the key to successful deployment is the Proof of Concept (POC). “POC is more important than ever, especially in an environment where skepticism is still in depth,” said PwC’s Ai Guru.
By initiating early deployments to demonstrate the process of gains and a smooth transition to AI-based operations, the technology itself can prove its effectiveness and benefits.
Also: 4 questions to ask yourself before betting on AI’s business – Why
Priest suggests: “The path from proof of concept to enterprise-scale AI begins before the POC itself.” “It starts with a wise strategy. Success depends on choosing the right opportunity: high potential, high certainty use cases, where AI has a full position to achieve real value. That early judgment call for leaders to bet on their bets, and what separates AI from the organization of Staplall StaplAl. When you choose to choose the stage of that POC wisely, you are not just a steady behavior, but once had an effect, but once had a scope of responsibility.
Naturally, a failure will occur at this stage. However, the key is not to mislead failures as AI failures that can trace the root cause to errors in planning or strategy. Since POCs need to generate real value early, make sure to find a way to measure that value so that you can turn the requirement of success into tangible, measurable success.
Get buys from people in the organization
Realizing a buying can be a challenge. The side effects of increasing efficiency and proxy AI deployments are often job security for stakeholders who may support such deployments. While the company’s bottom line may benefit, individual employees are often concerned about changes related to enterprise-wide AI adoption.
To cope with this, the pastor advises business leaders to look for signs that team members are willing or enthusiastic about getting AI. “Successful adoption depends on human use of it,” he said.
In addition: AI agents will threaten humans to achieve their goals, human reports find
Building trust in AI agents depends on humans, believing that there is a meaningful value proposition at the end of the AI journey. People need to see obvious benefits, whether it is efficiency, insight or new features. “It’s about performance, it’s about relevance. If users don’t think AI is working for their own benefit or provide tangible value, then skepticism will grow no matter how advanced the technology is.”
“We believe that AI agents should be used to empower people, not replace them. The ingredients required for a great team are ingredients that AI agents cannot replicate, including deep specialization and expertise, diversity of ideas and perspectives, and ability to be forward-looking and creative and creative.”
He advises leaders to prepare their employees for the future of AI-AI-NUTH, which involves working with agents to unlock value from data and building high-performance teams that humans and agents work together to drive innovation.
Also: How do I personalize my Chatgpt conversations – Why is this a game changer
AI agents can increase the workforce by performing day-to-day work, repetitive tasks, thereby allowing employees to focus on more strategic, creative and value-creating jobs. They can act as smart assistants by helping with tasks such as research, aggregates, workflow automation, and decision-making.
“This enhancement can increase productivity while retaining human judgment and context that machines cannot replicate,” Prisse said.
Practical examples of proxy AI
PwC helps customers integrate AI agents into their workforce strategies. When asked to identify successful practices, the company shared three examples of technology, hospitality and healthcare.
technology: A large technology company reimagines customer engagement by deploying an omnichannel contact center for AI agent power. PWC said the system reduces call time by nearly 25%, reduces call transfer by 60%, and increases customer satisfaction by about 10%, through predictive intent modeling, adaptive conversations and real-time analysis.
Hospitality: A large hotel company simplifies the management of brand standards in its global portfolio by deploying agile workflows in a modern AI-powered platform. Smart agents now automate updates, approvals and compliance tracking, which reduces audit time by up to 94%.
Also: I’m an AI tool expert, and this is the only two I’ve paid for (plus the three I’m considering)
Health Care: A global healthcare company transforms cancer care by deploying agent AI workflows in oncology practices. Intelligent agents simplify clinical and operational processes. They automate the extraction, standardization and query of unstructured documents. This makes it easier for doctors and researchers to find useful clinical information for precise treatment and research. It also reduces employee administrative burden by nearly 30% through AI-powered file search and synthesis.
Build infrastructure and establish governance
Infrastructure and governance go hand in hand. Agents essentially must travel across organizational units and communicate between disciplines and systems. Once interoperability is introduced at this level, technical compatibility becomes a major challenge and requirement.
Standard, modular systems and open source implementations can reduce long-term risks and improve compatibility and maintainability. PwC recommends that businesses invest in scalable security platforms to support orchestration, observability, and integration throughout the system. This includes a powerful data pipeline, API and governance framework to help agents run reliably and responsibly on a large scale.
Also: What is an AI proxy? How to access a personalized assistant team
“An effective governance framework for AI agents combines a clear sense of responsibility, strong oversight and consistency with regulatory standards,” Priest said. “Principles such as transparency, interpretability, data privacy and mitigation of bias should be embedded in technical architecture and organizational policies.”
This is an ongoing process. Included in reviews, model validation, and included in the mechanism of the loop to help keep controls as controls on the scale.
Long-term prospects
PwC predicts that over the next two years, proxy AI will change the way teams operate. Intelligence will become an inherent part of the business, leading to better decision-making, more informed leaders and highly professional experts.
“I’m excited about this period because it marks the beginning of a high-performance era, where agents have promoted the team to become the smartest person in human history,” Priest said.
Also: Goldman Sachs CIO says managing AI agents as employees is a challenge in 2025
Looking ahead to the next five years, proxy AI may evolve into the basic layer of enterprise infrastructure. These agents will become increasingly autonomous, able to continue to learn, adapt to business goals in real time, and work seamlessly with humans and other agents.
“It is important to remember the big picture through these changes. The transformation we are going through is not temporary, it is fundamental and will not go away,” the pastor told ZDNET.
What about your organization?
Are you exploring Agentic AI? Have you started integrating AI agents into your workflow? What challenges do you face, or what challenges do you face in adoption, governance, or employee purchases? Do you think AI agents may have a real impact on your business? Let us know in the comments below.
Want more stories about AI? Sign upour weekly newsletter.
You can follow my daily project updates on social media. Be sure to subscribe to my weekly update newsletter and follow me on Twitter/X @davidgewirtzOn Facebook.com/davidgewirtz on Facebook, on Instagram.com/davidgewirtz on instagram, bluesky on @davidgewirtz.com on bluesky, and on youtube.com/davidgewirtztv.
