AI agents are already being deployed by some of the world’s largest companies. WPP, for instance, has built close to 30,000 agents to enhance marketing effectiveness by encoding and scaling WPP’s domain expertise, tailoring solutions to client needs, and improving personal productivity within its AI-enabled marketing services platform, WPP Open. It is likely there will be many billions of AI agents deployed by the end of the decade across industries, automating increasingly complex tasks and projects without the need for human oversight. Verifying these agents continuously to ensure their safety and functionality will be an essential part of an emerging market that has been predicted to be worth trillions of dollars. 

Conscium has developed a platform where organisations will be able to send their AI agents to be tested within virtual environments and certified as fit for purpose. The platform uses a multi-agent storyteller concept similar to gaming, with non-player characters (NPCs) testing the subject agent for its suitability to perform the role for which it has been designed.  

Conscium was founded in 2024 by Daniel Hulme, who previously built the UK’s largest independent AI consultancy, Satalia, before it was acquired by WPP in 2021. Conscium’s mission is to develop safe and efficient AI that aligns with ethical standards and human values. It has recruited some of the world’s leading neuroscientists and computer scientists to its advisory board, including Karl Friston, Mark Solms, Anil Seth, and Megan Peters. AI agent verification will be the company’s first commercial offering.

Daniel Hulme, CEO and Co-Founder of Conscium, said: “2025 is the year of the AI agent. While Generative AI models can write you a Shakespearean sonnet or show how you would look as a Barbie doll still in its packaging, they cannot alter the real world. But that is changing.

“In a not-to-distant future, AI agents may be able to manage a company’s inventory based on supply and demand, reconcile expense receipts with your bank account, or create compelling personalised social media messages for thousands of customers, consistent with your company’s brand guidelines. Currently, they still require human supervision, but soon they will be able to operate independently in the real world. It is imperative that these AI agents perform as designed, without making mistakes, and that they do not deviate from their intended missions. In other words, AI agents need to be verified to ensure they are deployed safely.”

Stephan Pretorius, Chief Technology Officer at WPP, said: “The deployment of AI agents at scale presents both incredible opportunities and significant responsibilities. At WPP, we're committed to ensuring our AI-powered solutions deployed through WPP Open are not only innovative but also safe and reliable. Our partnership with Conscium and its AI agent verification platform is a crucial step in that direction, allowing us to rigorously test and validate the performance of our agents, ensuring they deliver value responsibly.”

For more startup news, check out the other articles on the website, and subscribe to the magazine for free. Listen to The Cereal Entrepreneur podcast for more interviews with entrepreneurs and big-hitters in the startup ecosystem.