June 26, 2026
8×8 Thinks Call Routing Has Been Broken for Decades. Its New AI Intelligent Routing Service Aims to Fix It
Skills-based call routing has been the backbone of contact centre operations for the better part of thirty years. An admin assigns skills to agents; those skills are periodically updated; incoming interactions are matched to the best available fit from the resulting pool. The problem, as most contact centre managers quietly acknowledge, is that the skills matrix is almost always out of date. Additionally, the pool almost always stops at the contact centre door. This is where intelligent routing powered by AI can come into play.
8×8 launched AI Routing this week at Customer Contact Week in Las Vegas. It pitched it as an organisation-wide routing engine that matches customers to the best available resource across the entire business, and in real time. Crucially, this is not just within the contact centre agent pool. The product spans 8×8’s Contact Center, Engage, and Work platforms, meaning it can direct an interaction to a back-office specialist or a subject matter expert as readily as to a front-line agent.
The core argument is about architecture. Hunter Middleton, Chief Product Officer at 8×8, put it plainly:
“The routing problem has always been about architecture. The best person for a given customer interaction might sit anywhere in the business, but most systems were never designed to look beyond their own boundaries.”
Rather than relying on manually maintained skill profiles, AI Routing analyses interaction transcripts, historical data, and real-time sentiment to build and update agent skill profiles automatically. Admins can review and accept suggested changes. Customer intent is detected without manual IVR mapping. Every routing decision is logged with exportable audit trails and confidence scores, which are relevant for any organisation operating under quality or compliance frameworks.
The system connects to 8×8 Intelligent Customer Assistant, AI Studio, IVRs, and third-party bots. Teams can begin on a single queue and expand from there. It is currently available to select 8×8 customers.
What the Market Could Make of 8×8’s New Intelligent Call Routing
The problem 8×8 is describing is tangible and well-documented. According to Natterbox’s 2026 Contact Centre Benchmarks report, pre-conversation hunting time, the period customers spend in routing before reaching a queue, fell 54% year-on-year between 2024 and 2025, from 5.15 minutes to 2.37 minutes. This was accredited to organisations replacing static IVR menus with conversational AI routing. The gains are measurable. The gap between organisations that have modernised routing and those still running manual skills administration is conspicuously widening.
What is notable about the 8×8 announcement is the skill profiling element, specifically. Sheila McGee-Smith, founder and principal analyst at McGee-Smith Analytics, identified this as the gap that AI routing vendors have historically failed to close:
“AI-powered routing has long promised a lot, but vendors have largely failed to address one key challenge: automatically assigning and maintaining agent skills and proficiency levels.”
If the system can genuinely maintain accurate, dynamic skill profiles from live interaction data rather than from an administrator’s best estimate updated quarterly, that alone would represent a meaningful operational improvement, independent of the broader enterprise-wide routing ambition. The two claims are worth separating. The skills automation is the more immediately verifiable one, while the cross-enterprise routing vision is the more ambitious.
The notable caveat is availability. AI Routing is in controlled early access for existing 8×8 customers. Organisations running significant third-party CCaaS infrastructure will want to examine the integration story carefully before drawing conclusions about fit.
What It Means for Operations and CX Teams Looking to Modernise Their Call Routing
Perhaps the key question to this development is whether AI routing decisions are visible and challengeable in real time, not just logged after the fact. The audit trail feature addresses the accountability concern. Meanwhile, the admin review step in skill profile management addresses the workforce relations dimension. Agents are not being assessed invisibly, and management retains sign-off on any suggested changes.
The promise is fewer transfers and faster resolution for customers. Whether that holds will likely hinge on how well intent detection handles genuinely complex, multi-issue queries. These are often the ones that defeat simple classification. Those are precisely the calls that matter most to CSAT and first-contact resolution rates, and precisely where the architecture faces its proper test.
