How AI-Native SIEM Could Help Security Teams Reduce Attacker Dwell Time

· Original article ↗

Summary

The article explains how security teams can use MTTD, MTTR, and investigation-workflow metrics to identify delays and assess whether AI-native SIEM helps reduce attacker dwell time and potential business impact.

Key points

  • Dwell time is the period between an attacker gaining access and a team containing the threat; the article says a recent incident-response report put global median dwell time at 14 days, up from 11.
  • The article frames dwell time as the combined result of mean time to detect (MTTD) and mean time to respond (MTTR).
  • Attackers may use legitimate credentials and ordinary tools, while individual events can appear harmless unless analysts connect activity across systems.
  • AI-native SIEM is presented as a way to help analysts correlate signals and investigate faster, not as a replacement for analysts.
  • The article recommends measuring high-severity incidents and delays between the first useful signal, investigation, and containment, rather than relying only on average MTTD and MTTR.
  • It cites IBM’s 2026 breach report: average breach cost of $4.99 million and average identification and containment time of 247 days; breaches lasting over 200 days averaged $5.65 million, versus $4.32 million for shorter ones.

Article Details

Defense Focus
Reduce attacker dwell time by improving detection, investigation, and containment, and measure the time spent at each stage.
Detection Methods
  • Correlate related events across systems, such as an unusual sign-in, access to a rarely used server, process execution, and creation of another account.
  • Evaluate detection and response using MTTD and MTTR, including for high-severity incidents.
  • Measure time from the first useful signal to analyst investigation, time spent gathering context, and delays before containment.
Data Sources
  • Security logs and events across multiple systems
Platforms
  • AI-native SIEM
  • Traditional SIEM
Defensive Actions
  • Establish baseline MTTD and MTTR, identify workflow delays, and measure them again after changing the workflow.
  • Assess high-severity incidents separately so averages do not conceal prolonged intrusions.
  • Track manual investigation and decision delays, and use AI-assisted detection and investigation to reduce them.
  • Measure containment time and prioritize reducing the interval in which an attacker can continue operating.