Consider the implications of this. Today, most workers use multiple applications, from half a dozen to more than fifty. As these applications evolve and agents are added, we will face the problem of “agent sprawl,” where users have as many virtual agents as applications.
At re:Invent, we had a CIO from a major bank who talked about how they were building virtual assistants for their own apps, as well as using Teams Copilot and Salesforce agents. After the session, I asked him what he thought the future would be like, and he told me he foresaw a day when users would have a “tapestry” of agents to choose from. I then asked him what he thought it would be like to work in such an environment, and he said, “Chaos, probably.”
Multiple agents create several problems. The first is that the agent or assistant only has the knowledge that is contained in the application. This can lead to fragmented insights. As bahamas mobile database example, consider a company that has a great website that best showcases a poorly made product. The web analytics and sales tools used before the sale may show high customer satisfaction rates because they measure satisfaction before the purchase. After the customer uses the product, their mood changes from happy to sad, and the contact center starts receiving calls for returns or repairs. Using the GenAI interface to understand customer sentiment would yield a different result.
Additionally, as agents move from reactive to proactive interactions, users will be inundated with messages from these systems as they try to keep you informed and up to date. I expect applications to have controls, just like today, so we can control how we interact with them, with most users keeping mission-critical applications turned on. It’s like a CEO having a team of advisors in every part of the company whispering in his ear at the same time.