Cisco Resolves 145,000 Support Cases Using Agentic AI

What happened

Cisco has disclosed that its agentic AI systems have autonomously resolved 145,000 customer support cases, giving the networking giant one of the more concrete, publicly quantified benchmarks yet for AI-driven service resolution at enterprise scale. The figure was reported by CX Today, citing Cisco's own account of how its AI-enabled support operations have performed.

Rather than simply assisting human agents, the systems described are said to have closed out cases end-to-end without escalation — a distinction that separates this disclosure from the more common industry claims of AI "handling" or "deflecting" queries. Cisco has not published a detailed breakdown of case types, timeframes, or the proportion of total support volume this represents, but the headline number positions the company as an early reference point for agentic AI's role in enterprise technical support.

Why it matters

Agentic AI — systems capable of taking multi-step action toward an outcome rather than just answering a query — has been heavily marketed but thinly evidenced at production scale. A named enterprise vendor citing a six-figure count of autonomously resolved cases gives buyers, analysts and rival vendors a tangible data point to test claims against, rather than relying on vague productivity narratives.

For technology and operations leaders, the significance lies less in the number itself and more in what it signals about where agentic AI is being deployed first: high-volume, procedural support environments where resolution paths are well-documented and repeatable. That is a useful proof point for organisations weighing where to pilot similar systems, and a reminder that credible AI-in-service claims increasingly need to be backed by hard resolution data rather than adoption or usage metrics alone.

By the numbers

• 145,000 customer support cases reported as autonomously resolved by Cisco's agentic AI systems.

The Renascence take

A single headline figure like this is easy to over-read in either direction — as proof AI has "solved" support, or as marketing spin with no substance. The more useful question for operators is what sits behind the number: which case types qualified, what "resolved" actually meant to the customer, and whether satisfaction and repeat-contact rates held up once humans were removed from the loop.

Resolution volume is a vanity metric unless it's paired with outcome quality — a case can be closed autonomously and still leave the customer dissatisfied or quietly re-contacting through another channel. Before citing numbers like Cisco's as a benchmark, service leaders should demand the denominator, not just the numerator: what share of total volume this represents, and how satisfaction, effort and re-contact rates compare to human-handled cases. The organisations that win with agentic AI won't be the ones who deploy it fastest, but the ones who instrument it honestly.