In its October 2026 Relentless Defense report, Cisco pressed 8,000 security practitioners with a brutal question: if you could run your last defense again, what would you change? The answer was surprising — nearly six in ten did not point at weak technology. They pointed at their own organization: siloed teams, scattered data, skill gaps. Around the same time, Microsoft’s latest reporting ranked regions by their exposure to cyber threats in the first half of 2026, placing one of the world’s most targeted geographies in the top four. A separate local survey found that within the past 12 months as many as 76% of companies had suffered a serious security incident — while only 3% qualify as top-tier “relentless defenders,” a slice of the 8% elite globally. The organizations getting knocked down were never the ones without weapons. They were the ones with the weapons, moving too slowly.
Cisco breaks defensive capability into three parts: Coverage at 25%, Speed at 25%, and Friction carrying half the score — because the 8,000 surveys voted for it. The starkest data point: only 14% of teams can update their defensive controls in real time, while 60% need a day or more to roll out a new control. Attackers win lateral movement in minutes; defenders run on a rhythm of meetings, assessments, approvals, and waiting. Another pair of numbers: only 30% of organizations have continuous automated threat detection as the default, versus 69% among the 8% elite defenders. The gap is not about budget. It is about whether the defense is continuously running.
The report’s most vivid moment is a scenario: an AI detection agent captures lateral movement matching ransomware precursor behavior within 47 seconds, recommends “isolate the host, confidence 71%,” and then — stops at “awaiting human authorization.” That is the accurate portrait of most enterprises in 2026: AI already performs detection and decisioning in seconds, but the last step, a human sign-off, holds everything hostage. More than a third of last year’s incidents involved AI techniques — attacker-side automation is a settled fact. On the defender side, fewer than half of organizations have widely deployed AI capabilities, not for lack of will, but because they still do not trust that they can control AI.
The same data points to the way out, though: defenders who use AI consistently report faster detection, faster response, and lighter operational load. This is not a vision statement — it is an experience repeated across 8,000 questionnaires. The key move is not exotic: turn the path from “detected” to “blocked” from a human workflow into an automated pipeline. At the edge, WAF and bot defense stop automated attacks and spoofed traffic first; anomaly detection continuously flags behavior that drifts from baseline; when rules trip, blocking and rollback verification happen automatically; humans keep only review and exception calls. It is not people who are slow — it is the pipeline that turns “one day of waiting” into “forty-seven seconds.”
76% hit, 8% standing. The most direct reading of that ratio: the next wave of incidents will not be starved of technology — they will be starved of continuity. Attackers do not pick the day your security team has fewer meetings; defense cannot run only while humans are awake. Chaining detection, decision, execution, and verification into one automated pipeline that never stops and does not depend on shift memory is the bridge that pulls the 76% back toward the 8% — with “a human notices the problem” on one end and “the system closes the problem itself” on the other.
The gap between 76% and 8% is not budget and not tools — it is the word “continuous.” LAFA’s AI operations detect, defend, and remediate automatically around the clock, turning “47 seconds of waiting for a human” into “no waiting at all” — if the attack never clocks out, the defense cannot clock out either.