No Audio

From Automation to Autonomy

Download report

The Capabilities and Complexities of AI Agents

AI PULSE SURVEY | VOL 3

6 min read

Nearly one in four organisations have already integrated Agentic AI into their operations. This rapid adoption rate highlights the urgency and enthusiasm with which businesses are embracing AI technologies to enhance decision-making and collaboration. Dive into the latest results of our AI Pulse survey and see how Agentic AI is reshaping the business landscape.

What your peers are saying about agentic AI

Key Findings

23 %

of organisations have integrated agentic AI and multi-agent AI into their operations

27 %

of organisations plan to integrate agentic AI into their operations within six months

38 %

of respondents prefer semi-autonomous agents that act within defined limits

46 %

of organisations are leveraging existing platforms with pre-built agentic AI capabilities as part of their agentic AI integration strategies

77 %

of AI-mature organisations are using or expect AI agents to manage repetitive tasks

Stages of AI Adoption and Maturity

At which of the following stages would you place your organisation regarding AI adoption?

Stages of AI Adoption and Maturity
Stages of AI Adoption and Maturity
Stages of AI Adoption and Maturity

Agents of change

Future Expectations for Autonomy

What level of complexity and autonomy does your organisation envision for its AI agents in the future?

Future Vision: AI Agents and Their Autonomy
Future Vision: AI Agents and Their Autonomy
Future Vision: AI Agents and Their Autonomy

*Roughly 6% of respondents are undecided on the level of complexity they envision for their AI agents.

Key takeaways

  • Despite the ambition for full autonomy, most organisations favor semi-autonomous agents.
  • AI maturity influences autonomy goals with organisations further along in adoption are more open to implement fully autonomous agents.

Leading the Semi-Autonomous Way

Key takeaways

  • Aerospace & defense, manufacturing, and financial services show strong support for semi - autonomous agents.
  • Context-aware agents are favored in sectors where human oversight, regulatory compliance, and personalisation are most critical.
  • 31% of executives foresee fully autonomous agents while mid-level leaders remain more cautious, with only 16% expecting full autonomy possibly due to their closer proximity to operational risks and implementation challenges.

AI agents are rewiring how work gets done

AI Agents are Reshaping Decision Making

To what extent have or do you expect agentic AI systems to affect the decision-making capabilities of your organisation? 

Agentic AI and the Future of Decision-Making
Building the Future: Integrating Agentic & Multi-Agent AI
Building the Future: Integrating Agentic & Multi-Agent AI

Key takeaways

  • Real-time insights get a boost as across industries, agentic AI is widely expected to enhance data processing and accelerate decision-making.
  • Healthcare and tech lead in leveraging AI for insights, while government and manufacturing remain cautious, likely due to regulatory and operational constraints.
  • 77% of the most AI mature organisations expect AI agents to take over repetitive decision tasks, streamlining operations. 

The Sprint to Onboard AI Agents

What is your organisation's expected time frame for integrating agentic AI and/or multi-agent AI systems into your core operations?

Key takeaways

  • The next 6 months are pivotal, as organisations with mature AI capabilities are rapidly integrating agentic systems to drive performance and strategic impact.
  • Adoption follows experience, with advanced organisations already deploying agentic AI, while earlier stages build momentum through deliberate scaling.

AI Agents are Powering Core Functions

In what functions or areas of your business are you currently using, piloting or expecting to use agentic AI and/or multi-agent AI systems?

Key takeaways

  • Strategic expansion is underway, as mature organisations deploy agentic AI across marketing, HR, finance, legal, and compliance functions.
  • Early results exceed expectations, with leaders reporting improved performance and continuous enhancements driven by data feedback.
  • Adoption patterns vary by industry, with tech, healthcare, and financial services leading in IT and risk management, while retail and manufacturing focus on operations and supply chain.

Building Smarter: AI Integration Planning

How is your organisation planning to develop or integrate agentic AI and/or multi-agent AI capabilities? 

Building the Future: Integrating Agentic & Multi-Agent AI
Building the Future: Integrating Agentic & Multi-Agent AI
Building the Future: Integrating Agentic & Multi-Agent AI

Key takeaways

  • Speed is strategy, as companies race to integrate agentic AI to stay competitive and avoid falling behind.
  • Partnerships fill skill gaps, with many organisations leaning on vendors and platforms like AWS, Microsoft, and Salesforce to accelerate development.
  • Maturity shapes the roadmap, with advanced organisations driving internal innovation, while earlier stages rely on external support and pilot programmes.

Meet the minds behind the report and insights

Tom Andreesen, Managing Director and AI Leader

Tom is a managing director with over 33 years’ experience helping organisations develop and implement a variety of business and technology solutions to enhance their operations. Tom has also helped companies establish risk management capabilities and overall governance programmes to help address operational risks, technology risks, and regulatory compliance requirements. Tom is the leader of Protiviti’s Global Microsoft Alliance programme.

“Invest in infrastructure, talent and training, and leverage available frameworks that can be customised to meet the agentic needs of your organisation.”

Connect on LinkedIn

Patrick Anderson, Managing Director

Patrick Anderson has over 25 years of experience in IT and Technology, specialising in AI, cloud computing, cloud economics, data strategy and analytics, programme and project management, agile transformation, and organisational change management. He holds numerous certifications, including Lean Six Sigma Black Belt, Agile Scrum Master, Data Protection Officer, and Generative AI Solutions with Azure OpenAI Service. Patrick has been working with linguistic processing since 1998 and Machine Learning since 2017.

“A balanced approach is essential so that organisations can transition to agentic AI systems successfully. Systems risk failing in unforeseen ways without such a balance. Leaders should chart a deliberate course and proactively seek support to identify and overcome gaps.” 

Connect on LinkedIn
Loading...