Tech buyers swamped with agentic AI choices

It's all so new

Analysts, vendors and their customers are still sorting out the function of agents, how they can be deployed in CX, and which vendor or vendors a particular organization will need, said Lou Reinemann, an IDC analyst. Different agents playing different roles will be needed by different-sized companies, and their needs will also depend on the maturity of the products they're selling.

"Early on, customer experience can be a really good differentiator; service can be a good differentiator," Reinemann said. "As you grow, and the product and the brand become more known, they're going to need to reinforce the brand qualities as you want to grow."

In CX, agentic AI seems to be moving toward more refined use cases, vendor by vendor, said Ross Monaghan, Adobe principal at Perficient, a global consultancy. Salesforce is working with CRM data in its own platform, while Adobe appears to be marketing-focused at the outset of its foray into the technology.

While agents might someday communicate with each other, he said, they will probably stay within their own platforms and role-based swim lanes.

Data, business outcomes key to AI success

Organizations will likely end up using multiple agentic AI platforms in the end, said Liz Miller, an analyst at Constellation Research. In her mind, it's unavoidable; the key to success will be a single data schema for all the agents to use.

"How am I ensuring the data that is delivered to each of these services comes from a single source and is curated?" Miller said. "It's literally that fundamental supply chain of data that we really have to pay attention to."

Ultimately, it's the job of management and not IT to decide how to employ AI, Adobe customer and JPMorgan Chase CEO Jamie Dimon said in a sit-down interview session with Narayen. Management doesn't need to comprehend nuts-and-bolts data science to decide how to use the tech; instead, it should focus on business outcomes AI can deliver.

For JPMorgan Chase, that means using automation to prospect for customers, manage documentation, purchase ads, detect fraud and perform other tasks across its divisions, offices and bank branches. A dedicated AI team decides which use cases to tackle next.

"As [AI] gets used, people are coming up with more and more wonderful stuff. It's just got to be part of your DNA, and it's hard," Dimon said. "A lot of people don't want to do it. They think it's the job of tech. It's not the job of tech; it's the job of management. You don't have to understand exactly how it works."

Don Fluckinger is a senior news writer for Informa TechTarget. He covers customer experience, digital experience management and end-user computing. Got a tip? Email him.