Agentic AI (sometimes referred to as multi agent AI systems) is set to revolutionize business operations. Joe Dunleavy, Global SVP and Head of AI Pod at Endava, explains how this exciting technology will pave the way for a more transparent, auditable, and sustainable use of AI and how its impact will transform businesses on a large scale.
Until now, humans have needed to be in the driver’s seat when it comes providing granular instructions to AI technology. This ensured that AI is not only pointed in the right direction in terms of outcome but also helped to minimize any potential risks such as hallucinations, misinformation or biases. Additionally, organizations that deploy AI often merely enhance the efficiency of singular tasks, reaping short-term value rather than targeting large-scale autonomous automation.
However, AI systems are now able to handle more ambitious business processes, decision-making, and data transformation. Agentic AI is set to revolutionize how organizations across industries can leverage this technology to their advantage. With the help of agentic AI, they will soon be able to automate processes in entirely new, more efficient, and autonomous ways, allowing them to address complex business problems at scale and speed. But how do they get there in the most effective, secure and compliant way?
Global SVP and Head of AI Pod at Endava.
The three stages of AI transformation
To achieve this level of performance, automation, and autonomy, AI requires a solid foundation. The transformation evolves over three phases. The first stage focuses on enhancing day-to-day work by assisting with tasks such as summarizing documents or generating assets like presentations, leading to faster, more cost-effective and accurate results. In the next stage, automation processes become more integrated with business objectives. At this point, AI takes on more responsibility for task sequences, working alongside people rather than just following individual commands. This way, AI evolves from a tool into a trusted partner.
In the third stage, the technology achieves an even higher degree of autonomy. At this point, AI is no longer ‘just’ a teammate that collects, summarizes, and analyses information. Instead, it takes on an advisory, more ‘proactive’ role. This is made possible by AI-based, autonomously acting agents (agentic AI) that can work without direct human intervention within any environment, including with different large language models (LLMs) and cloud platforms. Unlike traditional AI models, which are programmed specifically for singular processes, agentic AI approaches can handle far more complex tasks.
In a team of autonomous agents (multi-agent system), each agent is assigned an individual role and fed the necessary knowledge. These agents can communicate and interact with each other as well as with their environment, react to changes, and contextualize their tasks to make holistic decisions and achieve the best possible outcome. All of this works with minimal human oversight, without the need to manually provide input at every step of the process.
Although agentic AI technology is still in its early stages, these systems can safely drive workflows forward with minimal supervision. While autonomous agents automatically perform time-consuming, mundane and repetitive tasks, they can accelerate the amount of work done in a specific timeframe, which can be applied across the business to drive large scale efficiencies. This frees up employees who can in turn focus on more complex strategic and creative challenges. This approach nurtures every employee’s potential, increases employee job satisfaction, and drives business growth and value.
Benefit from autonomous agents – but not without transparency
Autonomous agents can be applied to tackle complex and nuanced workflows in any imaginable industry. However, AI systems are usually built as so-called black boxes with their functions and processes neither visible nor comprehensible to their users. As a result, strictly regulated industries such as healthcare, financial services, insurance and energy —where strict rules govern the collection, processing and storage of sensitive data — are often reluctant to implement the technology in their daily business operations. After all, they have to adhere to specific requirements when collecting, processing, utilizing, and storing (sensitive) data. Just as it's important to not only reach the right answer but also demonstrate the steps taken in fields like law or accounting, these industries must be able to clearly show how AI arrives at its results to meet compliance requirements.
The solution to this challenge is a data-first approach. In order for these industries to use AI in their favor and optimize their processes, they must be able to break open the black box and disclose its contents in a transparent and auditable manner. An autonomous multi-agent system that reveals how AI agents ingest and transform data is ideal for addressing this challenge as each time an agent acts upon data, the system captures the relevant information surrounding the operation, creating a clear line of sight and understanding of the decision the agent makes as an audit trial. This breakdown allows both data and processes to be made visible, comprehensible, and common AI-related issues like AI hallucinations can be effectively circumvented.
With the help of agentic AI, businesses can automate sophisticated processes and solve complex business problems on a large scale, all while remaining compliant. As a result, the technology is key to unlocking productivity, satisfaction, business growth, and maintaining competitive advantage. This does not mean that employees will be replaced by the technology. Whilst it requires less human intervention and oversight, users remain in complete control of the AI system and at the heart of operations. AI might be in the driver’s seat, but users dictate the direction and can step on the brakes at any point.
We've featured the best cloud storage.
This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro
