The Rise of Agentic AI: A New Era of Automation
A New Level of Automation
Generative AI is starting to deliver promising but limited results. However, the IT industry is pushing full speed ahead to the next level of automation, agentic AI. Since AI can’t yet design, build, and deploy agents, it’s up to humans to learn to create and target these agents productively. But this development is going to take some time.
Scaling AI-Based Services
Recent research from consultant Accenture suggests that scaling AI-based services to achieve sustainable business value is a challenge, with only 13% of projects delivering significant results in this area. The Accenture study covered 3,400 executives and more than 2,000 client projects.
The Need for New Skills
The rapid evolution to agentic AI calls for a new type of talent trained in AI and model development, as well as business acumen. "Our work with leading clients on talent and skills substantiates this," said the report’s authors, led by Jack Azagury, group chief executive for consulting at Accenture.
The Skills Architecture
Talent readiness is one of the biggest barriers to scaling and unlocking value for companies, so a proper skills architecture is needed to win in the age of Gen AI. "Currently, organizations are spending three times more on technology rather than on people — this must change," Azagury said.
The Gap in Training
Most companies are not training people for the age of AI. While Accenture found 94% of workers want to learn about Gen AI, only 5% of companies provide training in this area. "That gap must be closed," he said. "One can invest in all the available Gen AI tools, but if your employees don’t know how or why to use them or put trust in them, the value will simply not be realized."
Types of AI Agents
Azagury said three types of AI agents require companies’ attention at this time:
- Utility agents: "Perform routine, high-frequency tasks that enhance operational efficiency." Examples include a function in an autonomous vehicle or a dynamic pricing system.
- Super agents: "Combine multiple functions, synthesizing data to drive strategic workflows." An example is a marketing agent that assembles data from relevant sources and determines the sequence of steps to execute a campaign or report.
- Orchestrator agents: "Oversee end-to-end processes, breaking down silos and enabling seamless collaboration." An example is an agent that brings together multiple agents across different services, such as a production system that coordinates individual agents handling specific tasks, such as order supply, inventory, and scheduling.
Building and Deploying Agentic Architecture
Azagury said building and deploying agentic architecture requires a special breed of teamwork. This approach means assuming a "dual role of driving a step change in the software development lifecycle," he said. Technology professionals also need to ensure their organizations and employees "thrive using Gen AI overall."
Conclusion
The rise of agentic AI brings new challenges and opportunities for companies. To unlock the full potential of Gen AI, organizations must invest in upskilling their IT workforce with a methodical, skills-based, and structured talent strategy. This will require CEOs and other leaders to be brought up to speed on the implications of AI agents.
FAQs
Q: What is the main challenge in scaling AI-based services?
A: The main challenge is achieving sustainable business value, with only 13% of projects delivering significant results in this area.
Q: What are the three types of AI agents?
A: Utility agents, super agents, and orchestrator agents.
Q: What is the skills architecture required for the age of Gen AI?
A: A proper skills architecture is needed to win in the age of Gen AI, with a focus on AI and model development, as well as business acumen.
Q: What is the gap in training for Gen AI?
A: 94% of workers want to learn about Gen AI, but only 5% of companies provide training in this area.
