Five key takeaways about AI agents and their impact

Imagine it’s a Wednesday in the near future – a normal workday.

You sit down at your desk at 8:30 a.m. and open your laptop. A chat window opens, and an artificial intelligence (AI) agent presents you with a summary of the emails you received overnight, highlighting several that need immediate replies, which it already has drafted. You make a couple of quick edits, which the agent learns from, and send them off.

One of your clients has hit a snag on a project you’re working on, so the agent has rescheduled your morning and set up a call at 10 a.m. In your diary, the agent has summarized the client’s message about the problem and provided you with a list of suggested solutions.

The agent has also scanned the morning headlines. As well as providing a news round-up tailored to your diary for the day, it has seen stories about proposed EU regulatory changes that could present significant growth opportunities for another client. It has directed a separate agent to use the client’s latest sales figures to create a presentation with a range of projections and outline a possible marketing plan. You review the document, send it to your client and ask the agent to find time to discuss with her as soon as possible.

By 9 a.m., you’ve already accomplished more than you would have in an entire day just a few years ago.

While bringing AI agents into workflows represents a massive change for organizations, we see it as a huge opportunity – and one with humans firmly at the center. Here are the most important things to know about AI agents.

1. AI agents are here and spreading rapidly

The launch of ChatGPT in late 2022 marked a significant breakthrough in AI technology, leading to a surge in the development of AI agents. These agents are designed to work independently and perform complex tasks, and they are expected to become increasingly common in the next few years.

AI agents can be categorized into five types:

• Data agents analyze and combine information from various sources.
• Tool agents perform specific tasks, such as creating presentations or managing schedules.
• Coding agents assist in software development by generating code.
• System agents plan and manage workflows for both agents and humans.
• Validation agents review and check the work of other agents and individuals.

According to Statista, the market value of AI agents is projected to grow from US$5.1 billion today to US$47 billion by 2030. Organizations that embrace this change will gain a competitive edge.

2. Incorporating agents is a business change, not just a technology change

The integration of AI agents has far-reaching implications across various business functions, including human resources, risk management and pricing. Organizations must focus on identifying areas where agents can have the most impact.

At EY, we have developed methods to help professionals create AI agents that streamline workflows. By encouraging innovative ideas, we help empower our teams to identify challenges and develop solutions that improve efficiency. This approach not only improves productivity but also fosters a culture of innovation.

3. It’s humans working with agents, not humans versus agents

AI agents excel at performing repetitive tasks, often surpassing human capabilities in efficiency. However, their true strength lies in enhancing human skills. The most successful organizations will be those that foster collaboration between humans and agents, creating what we call “agentic workflows.”

In these workflows, humans remain at the center, responsible for generating ideas and strategies, while agents handle execution. This partnership allows individuals to focus on higher-level tasks that require creativity and critical thinking. Training employees to effectively use these new AI tools is essential, as is rethinking traditional team structures and performance metrics.

4. Risk management is crucial

With the introduction of AI agents comes a new set of risks. Many businesses are unprepared for these challenges. According to the EY Responsible AI Pulse survey, nearly half of C-suite executives believe their organizations are not adequately equipped to manage the risks associated with AI.

To reduce these risks, it is vital to keep humans involved in workflows that include agents. Agents can exhibit unintended biases or make errors, which can be addressed through targeted training and strict controls over their learning processes. Transparency is also essential; organizations must be able to demonstrate how agents arrive at their conclusions and decisions. A robust risk management framework is key to success.

5. Organizational change will be required

While the technology behind AI agents is advancing rapidly, organizational structures, security protocols, and compliance frameworks are hindering implementation. Additionally, employee willingness to work with AI agents plays a significant role in adoption rates. The EY AI Sentiment Index Study revealed that only 57% of individuals felt comfortable using AI tools.

At EY, we have designed our AI tools to provide a collaborative experience, helping employees understand how to integrate these technologies into their daily tasks. Four-fifths (81%) of our workforce is already using our generative AI (GenAI) ecosystem, EY.ai EYQ.

Reshaping the future of work

The changes brought about by AI agents affect every sector and type of business simultaneously. Organizations that can build hybrid teams of humans and agents will thrive in this new landscape. By empowering their workforce and modernizing management frameworks, businesses can harness the full potential of AI agents.

It is crucial for organizations and leaders alike to embrace the opportunities AI agents present. By understanding their capabilities and implications, we can reshape the nature of work for the better.

Agentic AI use cases

GenAI was used to develop an iteration of this article. In accordance with EY editorial guidelines, the end product was reviewed and edited by EY professionals before publication.