Accenture Plc is doubling down on its agentic artificial intelligence partnership with Nvidia Corp., announcing a new AI agent builder within its AI Refinery platform.

Using Accenture’s AI agent builder, business users can quickly create their own, highly customized AI agents, using the standard preconfigured ones already available in the AI Refinery platform as a kind of template. Users can adapt these new AI agents on the fly, in accordance with their changing business needs and market conditions, such as if there are policy changes or new products launched that they need to be aware of.

AI Refinery was announced by Accenture in January. It’s a platform for managing preconfigured AI agents, which are more advanced digital assistants that can perform tasks autonomously on behalf of their users with minimal supervision. In addition to Accenture’s initial 12 AI agents, it also provides access to third-party agents created by other developers. It’s built on the Nvidia AI Enterprise software and is available on all of the major public cloud infrastructure platforms.

Accenture says customers can now use AI Refinery to create and modify AI agents rapidly without any coding skills, meaning they won’t have to wait for support from their company’s engineering teams. Users will be able to make changes with just a few clicks, and this new capability applies to both Accenture’s own preconfigured agents, and also those created by third parties.

According to Accenture, dozens of major companies are already using the AI Refinery to cater to their specific agentic AI demands. For instance, ESPN used the platform to build an AI-powered avatar called FACTS, which is meant to interact with sports fans during its live coverage of events, pulling data from the ESPN Analytics service to answer their questions.

Hewlett Packard Enterprise Co. has used the platform to create AI agents for its HPE Private Cloud platform, while Beauty Tech Lab Ltd.’s Noli worked with Accenture to create digital assistants integrated within its website to provide hyper-personalized shopping experiences.

The United Nations is another fan of the AI Refinery, having used it to build a multilingual research agent that supports an incredible 150 languages. The agent is designed to help promote local economic development and sustainability, and it acts like a researcher, tapping into the U.N.’s vast Sustainable Development data.

More AI agents on show

Accenture announced its AI agent builder alongside an expanded lineup of preconfigured AI agents that are being published in the AI Refinery. These new agents are all powered by Nvidia’s Llama Nemotron models, which themselves were only announced today at that company’s GTC 2025 event.

The Llama Nemotron models were built on Meta Platforms Inc.’s open-source Llama family. They offer enhanced AI reasoning performance that makes them better suited to serve as a foundation for agentic AI, enabling them to act independently of humans.

Accenture said it’s planning to publish more than 100 industry- and task-specific AI agents in the AI refinery by the end of the year, having started with just 12 when it launched in January. Some of the newest ones announced today include an agent that’s focused on telecommunications assistance, helping call center operators to prepare for incoming calls and handle them more efficiently. The agent follows each conversation, acting as a kind of concierge for the human operators, providing real-time insights and recommendations to help them deal with the caller’s problem faster.

There’s also a new insurance underwriting AI agent that can perform much of the mundane, manual work involved in the underwriting process – such as data retrieval, research and analysis, validation and risk assessments. The new order-to-cash agents are aimed at finance managers and accounting teams, helping to streamline financial operations tasks such as order validation, invoice reconciliation and accounts receivable management.

Last, Accenture announced an AI agent that’s focused on providing commercial credit sales intelligence to the banking industry. It works by automating data extraction and rule-based decision making to accelerate commercial banking credit sales for credit underwriters.

Accenture Chief AI Officer Lan Guan said the company’s AI agents are all designed to tackle very specific business challenges.

“Accenture empowers organizations and decision-makers with the flexibility to drive business value faster with agents that can observe the environment, apply reason, continuously improve and take action,” she promised.

Image: Accenture