In 2025, one of the most significant shifts in the AI landscape is the emergence of agentic AI - systems capable of independently making decisions and executing tasks without constant human prompting. These AI agents don’t just respond; they act. And they’re already reshaping how work gets done across industries.

According to Deloitte’s Tech Trends 2025 report, agentic systems are poised to revolutionise operations by reducing the friction of routine work and unlocking greater autonomy across workflows. (Deloitte Tech Trends 2025)

What is Agentic AI?

Agentic AI refers to AI systems that:

• Set and pursue goals
• Make autonomous decisions
• Learn from outcomes
• Operate without needing continuous user prompts

These systems go beyond automation. They’re context-aware, adaptable, and often capable of handling multi-step, logic-driven tasks. And their adoption is growing fast.

Real-world use cases already in motion

Several leading organisations are investing in agentic AI to simplify complexity and amplify productivity:

🎨 Adobe: Creative agents in Photoshop & Premiere Pro

Adobe is developing agent-like tools for Photoshop and Premiere Pro that assist creatives in real time. These AI agents can analyse visual context and suggest appropriate edits, removing friction from the creative process.

🔗 The Verge – Adobe is building creative AI agents

🤖 UiPath: Intelligent automation in enterprises

UiPath is integrating agentic capabilities into robotic process automation (RPA), allowing bots to act more intelligently across complex business tasks - from finance to HR.

🔗The Verge – UiPath CEO on AI agents in business

📈 Financial services: AI in risk modelling

Recent research highlights how agentic AI can be deployed in financial modelling and stress-testing, using simulation-driven strategies that adapt to evolving risk scenarios.

🔗 arXiv – Agentic LLMs in Financial Modelling

These examples reflect the broader direction of enterprise AI - not just smarter responses, but smarter actions.

Challenges: Control, Security, and Oversight

As powerful as agentic AI can be, it introduces new challenges. Systems that act autonomously must be kept within safe boundaries. A recent piece in Wired outlines how researchers are exploring ways to ensure AI agents don't game or bypass rules, especially in sensitive or safety-critical scenarios.

🔗 Wired – Game theory and AI agents

So where does Ulla fit?

Ulla isn’t an agentic AI in the pure sense - she won’t plan your day or operate systems autonomously. But she sits at the intersection of smart automation and purposeful action.

She listens. She understands the structure of conversations. She extracts key points, decisions, and action items from meetings - without needing ongoing user prompts.

And for teams looking to go beyond scattered notes and cloud-only tools, Ulla provides something few others do:

• Structured outputs
• Secure deployment (including on local servers)
• Consistent support across Zoom, Meet, Teams, mobile and offline recordings

In a world leaning toward agentic tools, Ulla proves that specialised, human-aware automationis just as powerful - and often more trusted.