Harnessing the Potential of Agentic AI: Lessons from the AI Summit 2024

2024 was all about how AI is transforming the business landscape. 70% of firms believe generative AI has either begun disrupting their business or will do so in the next 18 months, according to our new report with IDC, Navigating the Digital Shift. This is prompting business and tech leaders to proactively assess the future of AI.?

To close out 2024 on a high, in December, we exhibited at the 2024 AI Summit to hear over 250 guest speakers set the standard for AI excellence across six stages at New York City’s Javits Convention Center. In its ninth year running, the event saw technology leaders discuss the latest trends and research – as well as innovative AI applications – that are transforming industries globally.??

Here is what we learned from key conversations over the two-day event:??

Harnessing agentic AI capabilities?

A key topic of conversation was the impact of agentic AI, what it is and how it will impact businesses. With multi agent AI capable of autonomous action and decision-making, there was general excitement that it has the potential to create new opportunities in efficiency and flexibility. However, this raised concerns regarding the quality of internal data and how it is managed to successfully harness the potential capabilities of AI.???

Secondly, there were concerns about model accuracy and putting solutions in front of customers that were not fit for purpose and could ultimately put a brand at significant risk. Delegates were also skeptical of point solutions for agents and the ability to manage multiple agents simultaneously to leverage scale.?

Another issue raised was the number of internal generative AI proof of concepts (POCs) that lacked a unifying theme, plan to deliver or clear method of measuring, often creating false starts or a lack of direction.? There was also a lot of frustration about how to clearly articulate the business benefits and ROI from these investments.?

While there was a lot of excitement about agentic AI throughout the event, conversations often headed in the direction of how best to set guiding principles, key objectives and aims that focus on the business issues that matter.??

Our answer: Focus on business outcomes first?

To address these issues, it’s critical to focus on?the type of organisation you are aiming to be, the customer experiences you want to provide and?how to operationally meet them. ?From there, you can pinpoint the best use cases that will have a measurable impact on your business. We discussed with delegates a framework developed at Endava to identify and prioritise high-value use cases, based on impact and readiness, and an accelerator to business case through rapid prototype and discovery.?

Many delegates expressed concern that use cases offering the largest benefits also posed the most significant risk, potentially damaging brand image with customers. Our research and client engagements over the past two year involving multi-agent AI?led to Endava developing its Morpheus framework which utilises multiple techniques to ensure robust solutions including flexible model choice, agent reflection and memory management.? At the conference we shared several demos showcasing this capability.?

Many discussions of benefit realisation quickly turned to internal data quality. ?Many attendees expressed concern that poorly managed data assets would inhibit their AI efforts. ?Clearly, well managed proprietary data is critical as for models to perform successfully as they require accurate, consistent, and mostly complete data. ?I discussed with delegates ways that AI can speed up core transformation efforts, and the Endava accelerator Compass allows a rapid diagnosis of legacy systems from which to develop a modern core architecture, forming the essential foundations from which AI can be successfully leveraged.??

Despite the potential challenges and headwinds of successfully leveraging new AI capabilities, particularly agentic AI, having a clear roadmap for understanding the most valuable use cases to focus on is a great place to start. This allows rapid prototyping to really understand these capabilities further to maximise their potential.? This is a great opportunity for technology teams to open up a dialogue with their operational colleagues on the exciting potential of AI.?

Get in touch with any questions you have, and we’ll be happy to help.?

?

?