Synopsis: KJ Kusch, global field CTO for WalkMe, a unit of SAP, explains how the rise of artificial intelligence (AI) agents will transform the application user experience.
Hybrid work stacks now include a parade of AI helpers—SAP’s Joule, Google’s Gemini, IBM’s Watson and more. Kusch tells Mike Vizard that most employees bounce among them without a traffic cop, leading to contradictory prompts and stalled workflows.
The root issue, he says, is that enterprises still switch context every time a task spans multiple systems. “AI needs to be orchestrated, not just turned on per application,” Kusch argues, because business processes rarely live inside a single tool. WalkMe’s answer is to watch for hesitation—three mis-clicks, a long pause—and then surface the missing policy, knowledge-base article or regulatory rule so the user can keep going instead of opening yet another tab.
That safety net matters as organizations reassess last year’s AI rush. Teams now question data provenance and model accuracy; Kusch predicts another “bump” as companies insert extra checkpoints to balance speed with quality. At the same time, leaders are eyeing app consolidation. If machine-learning can deliver the niche insight a standalone SaaS once provided, why keep paying for overlapping interfaces? The long game, Kusch says, is fewer screens and more reusable data services behind a common experience layer.
Yet many firms haven’t automated even basic tasks. Kusch is still surprised to see factory workers typing delivery codes by hand—proof that starting small beats chasing “solve-everything” moonshots. He advises deploying AI on safe, repetitive chores first, then expanding once the ROI is clear.
The takeaway: before rolling out another agent, map the real user journey, set guardrails for when models disagree and measure whether AI actually removes clicks. Otherwise, the promised productivity boost turns into just another layer of digital noise.