CareCredit - Women's Health April 2024

How End-to-End Agentic Software Will Transform AI for Business 

Change happens slowly, and then all at once.

And when it comes to business software, we are firmly in the “change happening all at once” phase.

That’s according to Nvidia CEO and founder Jensen Huang, who on Wednesday (Feb. 21) heralded the start of a new era — that of accelerated computing.

“We have hit a tipping point,” Huang said during his company’s latest financial earnings call.

And he would know. Nvidia’s computer chips and both on-premise and cloud-based data factories were responsible for the computing power and inference speed behind the advent of generative artificial intelligence (AI), and the next generation of chips are already being used to power the next wave of advancements: agentic AI software.

Existing GenAI software took the business world by storm in just one year with its scalable prompt-response architecture that was able to move between, and across, text, image and video, to produce dynamic responses to static user queries.

Agentic software is positioned to take the capabilities of an automated software program to the next phase: a request-action architecture.

As the name implies, agentic AI can conduct business operations from start to finish without the need for human intervention. For example, agentic programs will be able to attach a file to an email and send it, rather than just creating the text of the email or retrieving the file to be attached.

Companies like OpenAI are already working on these next generation digital assistants, and their commercialization will transform existing workflows and drive new engines of productivity across sectors.

Read moreAI Copilots Usher in the Service-as-Software Era

What Intelligent Software Holds for Payments

“Agentic AI systems — AI systems that can pursue complex goals with limited direct supervision — are likely to be broadly useful if we can integrate them responsibly into our society,” wrote OpenAI researchers, explaining that such systems have “substantial potential.”

That’s because agentic software can leverage AI to make decisions, learn from data, and interact with users or other systems.

And this ability to complete tasks from end-to-end holds vast implications for the payments landscape.

Agentic AI software can automate routine and repetitive tasks related to payments, such as invoice processing, data entry, and transaction reconciliation. This automation can lead to increased efficiency while reducing the likelihood of errors associated with manual processes, a goal that has long been prioritized within finance departments across sectors.

Within fraud prevention and detection, agentic AI can analyze patterns and detect anomalies in payment transactions, helping to identify potential fraud while learning from historical data to improve accuracy in flagging suspicious activities — all without the need for human intervention.

AI agents can also analyze historical payment data to make predictions about future trends, enabling better financial planning and decision-making within the enterprise. This includes forecasting cash flow, predicting payment delays, and optimizing payment schedules.

Read more12 Payments Experts Share How AI Changed Everything in 2023

Shift to Agentic Architectures

Speaking to PYMNTS this February, Ahsan Shah, senior vice president of analytics and AI at Billtrust, predicted a shift toward agentic, or multi-agent, generative AI in the near future, explaining that it involves a layer above individual processes, where AI acts as an orchestrator across various domains and can carry out tasks rather than just surface information.

This approach will optimize various steps in the accounts receivable process and enable personalized recommendations, campaign management and strategic decision-making. “This is a transformational opportunity for B2B,” Shah told PYMNTS.

The nature of the upcoming shift represents a fundamental phase shift in software capability.

“It won’t be something you ask and get an answer back, but a system you can ask to do things for you,” James Clough, chief technology officer and co-founder of Robin AI, told PYMNTS during a conversation for the “AI Effect” series this January. “Instead of just drafting that email, it might draft the email and get the attachment and put it in your outbox and then click send as well. I think that shift from chats to agents is one of the most exciting things we’ll see in the next year.”

As revealed in the PYMNTS Intelligence report “Consumer Interest in Artificial Intelligence,” consumers interact with about five AI-enabled technologies every week on average, including browsing the web, using navigation apps and reviewing online product recommendations.

Nearly two-thirds of Americans want an AI copilot to help them do things like book travel, and AI is already helping to transform and personalize the in-car experience.

Still, as Shaunt Sarkissian, founder and CEO of AI-ID, told PYMNTS, “It’s important to consider the endpoint of the automated interaction, where the AI leaves off.”

That’s because while the benefits are substantial, it’s crucial for businesses to consider ethical and privacy implications, data security, and the need for ongoing monitoring and adjustment of agentic AI systems to ensure optimal performance and compliance.