IBM's announcement of expanded watsonx Orchestrate capabilities at Think 2026 has exposed a structural problem the WEM industry has been avoiding: no platform currently orchestrates AI agents across vendor boundaries at enterprise scale. Contact centers are deploying thousands of AI agents—virtual customer-facing systems, QA scoring tools, scheduling bots, coaching assistants—each from different vendors, each operating with limited awareness of the others. The result is a collection of point solutions running in parallel rather than a governed, coordinated system. Research from MIT Sloan and BCG found that 47% of enterprises deploying AI have no strategy for managing their agents, a gap that becomes critical when Cisco forecasts agentic AI will handle 68% of contact center interactions by 2028. The mathematical foundations of traditional WFM have collapsed under this new reality: classic Erlang models assume random, independent arrivals, but AI agents fail in clusters, escalating bursts of already-frustrated customers into human queues simultaneously. This is not a future problem—it is an operational liability today.
The WEM vendors closest to addressing this gap—Verint-Calabrio, NICE, and Genesys—are all moving in the right direction, but each has the same fundamental limitation: their orchestration layers manage agents within their own ecosystems. Verint-Calabrio's reframing of the WFM operator's role from queue manager to "AI technician" and its Long-Term Capacity Planner for budgeting AI headcount represent genuine architectural progress. Yet none of these platforms can govern a NICE virtual agent running alongside a Genesys scheduling bot alongside a third-party QA tool. The question facing CX leaders is whether the WEM industry will close this gap in the next 18 months, or whether enterprises will turn to horizontal AI platforms—IBM, Microsoft, ServiceNow—to orchestrate their multi-agent workforce from above, leaving WEM vendors managing the edges of a system they no longer control end-to-end. IBM has not entered the WEM market, but it has just demonstrated what genuine multi-agent governance looks like at scale.
The implications for teams already running Zendesk, Freshdesk, or Salesforce are immediate and uncomfortable. Your current platform likely orchestrates its own agents well but cannot see or govern agents from competitors. As AI density in contact centers increases, this visibility gap becomes a control problem: you cannot manage what you cannot observe, and you cannot optimize a workforce you do not fully govern. The vendors that solve heterogeneous agent management—real-time conflict resolution across overlapping workflows, unified observability across infrastructure, and governance controls that travel with agents regardless of where they run—will own the next generation of WEM competition. Those that do not will find themselves managing the operational consequences of a multi-agent workforce they never designed to coordinate.
IBM’s Think 2026 framework reveals a coordination problem the WEM industry has been quietly avoiding. With thousands of AI agents now operating across enterprise contact centers – each from a different vendor, built for a different task – the question of who, or what, orchestrates