The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX · The Agile Brand

Tray.ai's Stephen Stouffer on AI adoption and why marketers should care about the integration layer

December 22, 2025·26 min·2 clips
Why are companies spending billions on AI only to see projects stall before impacting customers?
The conversation begins with a question about why so many companies pour money into AI and still fail to see customer-facing results. The host frames that gap as an agility problem, not just a tooling problem. He argues that scaling AI depends on connecting systems and data sources that usually sit in separate lanes. Stephen Stouffer joins to walk through the integration layer that makes that connection workable in practice. He treats integration as the part that lets a pilot turn into something an organization can actually run. The discussion keeps coming back to operational reality rather than broad AI slogans. Stouffer recommends starting by documenting repetitive tasks that drain time and attention. That exercise helps teams identify obvious automation wins and gives them a concrete way to build buy-in. The episode also separates internal efficiency gains from outcomes that can be tied directly to revenue. Sales ops and marketing ops come up as obvious places to make that case. Case deflection is another example the conversation uses to show how automation can pay off. The host presses on how teams move from one or two pilots to something broader. Stouffer points to a front-end interface that makes building at scale more approachable. He says that matters when a company needs to create hundreds of integrations. The conversation notes that code alone does not solve that problem. A scalable infrastructure underneath the interface is part of what makes large-scale adoption possible. The Yext example gives the discussion a concrete benchmark. Stouffer cites a 60 percent reduction in integration costs as a sign that the operating model can produce real savings. The episode treats AI adoption as a coordination problem as much as a technology problem. It ends with the idea that one well-chosen use case can prove value and open the door to the next.

As heard by us

AI adoption stalls when systems, data, and workflows do not scale together.

The episode keeps coming back to a plain problem: AI work slows down when systems, data, and workflows do not move together. Greg Kilstrom and Stephen Stouffer spend the conversation on the practical side of that gap, especially the work of connecting scattered systems and data…

Read the full review in PlayNext →

Why you'd press play

If AI pilots keep stalling, this gives you the integration-layer view behind scaling AI beyond the pilot stage.

Read the full recommendation in PlayNext →
Listen to the show on