Ibex's Sierra Tie-Up Shifts the Real Test From AI Chat to Who Owns the Handoff
Ibex, the contact center outsourcer with more than 36,000 human CX staff and 138 million managed interactions a year, announced a partnership with autonomous-agent vendor Sierra on May 5, 2026, layering Sierra's AI agent engine onto ibex's existing Wave iX stack and human operations. Philippine Airlines is already running the combined offer in phases. On ibex's May 6, 2026 earnings call, chief executive Bob Dechant said digital and omnichannel services had grown 18% and now make up 82% of revenue, and drew a clean line between the two vendors: Wave iX handles agent assist, training, translation, verification and routing, while Sierra supplies the autonomous conversation layer ibex decided not to build itself. What's not yet public is joint performance data for a full journey — AI agent to human agent to resolution — reported by CX Today.
Who's actually accountable when the AI-to-human handoff fails?
Ibex is contractually accountable to the client, but the failure itself could sit inside Sierra, Wave iX, a CCaaS platform, or a client-owned connector.
Dechant's own description of the commercial structure supports this reading: clients contract with and are billed by ibex, which builds the implementation and pays Sierra through a negotiated, resolution-based cost arrangement. That's a sensible way to keep one name on the invoice. It does nothing, though, to prove that a failed transaction — a payment that doesn't post, a booking that doesn't confirm — gets diagnosed and fixed by one team rather than passed between two vendors' support desks while the customer waits.
Does the Philippine Airlines rollout prove the handoff works?
It proves the offer is live, not that escalations resolve cleanly — no published data yet covers a full AI-to-human journey.
A production customer running the joint stack "in phases" is a meaningfully stronger signal than a roadmap slide, and buyers evaluating ibex should weight it accordingly. But phased rollout is also exactly the stage where handoff problems hide: early phases tend to route the easy, well-bounded conversations, leaving the messy multi-system escalations for later. Until ibex publishes escalation-rate or resolution-cost figures for the combined offer specifically, the airline deployment is evidence of adoption, not of the handoff being solved.
Why didn't ibex build its own autonomous agent?
Management concluded it couldn't match a specialist vendor's autonomous-agent engine, so it partnered with Sierra instead of building in-house.
That's a defensible build-versus-buy call, and it mirrors a split this site has already flagged: CX vendors are increasingly separating into orchestration/operations players and specialist AI-engine providers rather than trying to own both layers. The trade-off is that ibex's pitch of "one journey, one operating model" now depends on integration quality between two independently-run product roadmaps, not on ibex's own engineering.
Frequently asked questions
What does ibex actually own versus Sierra?
Ibex owns the managed service, client contract, Wave iX layer, implementation, analytics and human contact center operation. Sierra owns the autonomous agent platform underneath it.
Is there published performance data for the combined offer?
Not yet. Ibex has reported company-wide figures — 36,000 CX experts, 138 million interactions a year, 18% growth in digital/omnichannel to 82% of revenue — but no joint metrics for a full AI-to-human journey.
What should a buyer ask ibex before signing?
Ask ibex to force a failed transaction in a demo, show every data object handed to the human agent, and name the single party contractually responsible for recovery.
Source: CX Today, "Ibex CX Integration: Is AI-to-Human Handoff the Real Test?" (August 27, 2026).