Rather than building artificial intelligence (AI) agents from the ground up, many smaller organizations are looking to technology partners to supply this capability to them in a way that doesn’t require a significant amount of expertise to build, deploy and maintain.

Prosus N.V., a provider of an e-commerce platform used by more than 2 billion people, for instance, has made available ToqanClaw, a general-purpose AI agent that enables any end user to create an application or automated workflow without requiring them to write any code, and Zapia, an AI assistant that enables end users of its platform to make reservations and coordinate schedules.

Instead of deploying, for example, an open source AI agent such as OpenClaw that then needs to be secured and maintained, ToqanClaw provides an alternative approach based on a proprietary AI framework, dubbed Toqan, that is integrated across more than 20 different AI models.

That approach offloads the cost of investing in an agentic AI platform to the same company that organizations, such as restaurants, already rely on to manage e-commerce transactions, says Euro Beinat, global head of AI for Prosus.

Prosus has already been using the same AI agent internally to give 40,000 employees access to a Large Commerce Model (LCM). The company trains the model using data from 500 million daily interactions generated by more than 1 billion end users on its e-commerce platform, which runs on Amazon Web Services (AWS).

It’s not clear how many organizations will prefer to build their own AI agent versus making use of ones that are provided to them by the providers of applications they already rely on. While there is little doubt that there are already millions of users of OpenClaw, many organizations might determine that the overhead associated with securing and governing an AI agent might best be handled by a third-party that has more AI expertise. In many cases, organizations in the retail sector simply don’t have the skills and expertise required to build and maintain AI agents, notes Beinat. “In a lot of the organizations that we serve the IT literacy is low,” he says.

At the same time, those same organizations want to be able to take advantage of technology advances to compete with what are often larger rivals.

It’s too early to say what the ultimate economic impact of AI agents is going to be. On the one hand, there is little doubt that organizations will become more productive as more employees are able to “vibecode” applications that previously would have required a significant amount of expertise to create. On the other hand, it won’t be too long before AI simply becomes yet another table-stakes capability that almost every organization has.

Each Digital CxO, of course, will need to determine how much AI expertise their organization will actually require. The one thing that is certain is the pace of digital business transformation is already starting to accelerate in ways that may soon be bubbling up more from the bottom of the organization versus necessarily waiting for someone at the top of the corporate ladder to officially sanction it.