AI Agents: What if AI performs tasks itself?

What is agentic AI?

Agentic AI, also known as AI agents, is the next step in leveraging artificial intelligence within organizations. While traditional AI primarily responds to questions or commands, Agentic AI can take initiative on its own. An AI agent understands a goal, reasons about the steps needed to achieve that goal, and then executes them. Think of a digital colleague who not only provides answers but also completes tasks and makes adjustments. These agentic AI systems consist of multiple components: language models, databases, and integrations with other tools such as calendars, email, or a web browser.

What should I think of?

Agents can assist with various types of work:

• Analysis and understanding: processing data and recognizing patterns.

• Control and validation: checking results against rules or quality standards.

• Search and match: linking relevant information, people, or documents.

• Structuring and organizing: categorizing and summarizing information.

• Action and decision-making: independently performing actions to achieve a goal.

Each level requires clear boundaries, good context, and human oversight. This way, you work safely, responsibly, and effectively with agentic AI in the cultural sector.

Opportunities in the cultural sector

Agentic AI may seem far off, but many organizations are already working with it. Here are some concrete examples:

• Appointments and minutes: the agent reads new emails, schedules appointments, creates a brief summary of the meeting, and saves the summary in the file.

• Audience analysis: an agent automatically reads visitor data from the ticketing system, analyzes it, and combines the data with other information, such as demographic information or visitor feedback.

• Fundraising: The agent automatically searches online for new subsidy opportunities based on your organization’s profile. The specifications are read and partly completed by the agent.

• Education: Cultural education is personalized based on the target group. The teacher or student provides a profile, and the agent personalizes the educational material.

• Marketing and communication: Schedule content on social media or in your CMS, manage multiple channels, and optimize posts based on results.

• Customer service: An agent partially manages a mailbox. Is it a simple question? Answer via the FAQ. Can another colleague handle it better? Forward it. Is it urgent? Send a push notification immediately.

• Event planning: An AI agent automatically creates a detailed script for an event. The agent also sends automatic reminders and updates to all involved via email or a messaging app like Microsoft Teams.

Agentic AI can help cultural organizations work smarter, more consistently, and more efficiently. Take inventory of which processes in your organization are repetitive, data-driven, or planning-oriented. These are the areas where agentic AI offers the most potential.

Three levels of autonomy

AI agents evolve from simply reactive to independently acting. There are roughly three levels of autonomy (agency):

Getting started with agentic AI

Agentic AI is new territory, but the potential is vast. Want to get started with agentic AI? Follow this step-by-step plan:

1. Start small: choose one task that takes a lot of time or recurs frequently. Build a simple agent, such as a mailbox manager that prioritizes or filters messages.

2. Observe, learn, and improve step by step. This way, you gain experience without significant risks. For example, start with an agent in a conversational tool or a simple workflow in an automation tool.

3. Expand functionality once the prototype works well.

4. Share what you learn. By sharing experiences and results within the sector (e.g., in the DEN Community (opens in new tab)), collective insights grow.

5. Involve colleagues, partners, and the public to ensure applications align with your organization’s values.

6. Establish clear agreements and evaluation moments.

Small steps, shared knowledge, and human oversight make the use of agentic AI safe, effective, and valuable. Together, we can discover how agentic AI can make cultural institutions smarter, more accessible, and future-proof.