The AI landscape doesn't move in one direction — it lurches. Some techniques leap from experiment to table stakes in a single quarter; others stall against regulatory walls, technical ceilings, or organisational inertia that no amount of hype can dislodge. Knowing which is which is the hard part. The State of Play cuts through the noise with a rigorously maintained index of AI techniques across every major business domain — classified by maturity, evidenced by real-world adoption, and updated daily so you always know where you stand relative to the field. Stop guessing. Start knowing.
A daily newsletter distilling the past two weeks of movement in a domain or two — delivered to your inbox while the index updates in the background.
Each dot marks the weighted maturity of practices within a domain — hover for a brief summary, click for more detail
AI that executes predefined incident response playbooks automatically, containing threats and preserving evidence. Includes SOAR platform automation and orchestrated containment; distinct from automated remediation in IT ops which restores service rather than containing threats.
Incident response automation is transitioning from deterministic SOAR platforms to agentic AI systems. Traditional SOAR platforms achieved mainstream adoption across large enterprises and MSSPs with documented 40–90% reductions in manual effort and sub-ten-minute resolution for common incidents, but maturity plateaued around governance and playbook maintenance costs. Beginning in Q1 2026, major cloud vendors (AWS, Microsoft, Arctic Wolf) released agentic incident response systems executing autonomous investigation and containment in milliseconds. Autonomous incident response deployment surged 412% in 2025 (8% to 41% adoption), driven by 89% increase in AI-enabled attacks and threat-accelerated urgency. The new tension is calibration: agentic systems reduce MTTR from hours to minutes but require human-in-the-loop guardrails to prevent false-positive over-triggering and scope violations. Real-world data shows 53% of organizations have already experienced AI agent scope violations, and legacy SOAR achieves only 40–55% alert automation in practice despite 95% claims. The practice is proven and mainstream, but success depends on disciplined scope, human oversight, and acceptance that autonomous systems require continuous tuning and governance.
The vendor ecosystem has completed transition from legacy SOAR (Cortex XSOAR, Splunk SOAR) to agentic platforms as table-stakes. Q1 2026 saw vendor consensus: AWS DevOps Agent (GA March 31) documented 77% MTTR reduction at WGU (2h→28min) and 75% at Zenchef (1–2h→20–30min); Microsoft's Agentic SOC enables parallel autonomous investigation across identity, endpoint, email, and cloud; Arctic Wolf Aurora (GA March 2026) orchestrates 300+ specialized agents automating 90% of Tier-1/2 tasks with 85% alert fatigue reduction. Palo Alto shipped Cortex XSIAM Autonomous Playbooks with zero-customization governance and auto-updates; Arvo AI's Aurora Actions and incident.io enable natural-language scheduled/post-incident/manual automation across 22+ integrations. Named enterprise deployments confirm operational maturity: Target and Shopify automated triage/investigation at Google Cloud Next 2026; Google Cloud's phishing containment achieves MTTC <60 seconds with 95% SOC hour reduction (1,200 to 50 annually). SOAR market grew to $1.87B (2025), forecast 18.6% CAGR to $4.4B by 2030. Analyst sentiment shifted: KuppingerCole named Cortex AgentiX market leader; Rod Trent's deployed-platform analysis shows 3–10 minute investigations (vs. 20–40 human minutes), 85–90% MTTR, 97–98% accuracy, 3–4% escalation rates.
Critical governance gaps have emerged as deployment barrier. CSA survey (600+ orgs): 53% experienced AI agent scope violations, 47% had AI agent incidents, 97% expect major AI agent incident within 12 months, 65% already experienced one. Proofpoint research (1,400+ professionals): only 33% fully prepared to investigate AI-related incidents, 52% lack confidence controls detect compromised AI agents. Practitioner consensus identifies confidence-based case formation, deception-based validation, and tiered response (human-in-loop for edge cases) as prerequisites for safe automation; one-way automation architectures amplify coordination failures. Legacy SOAR achieves 40–55% alert automation in practice vs. claimed 95%; 83% of analysts report alert burden despite deployments. The practice is operationally proven but requires permanent playbook governance, authorization scope discipline, forensic artifact capture for agent environments, and acceptance that deployment outpaces governance capability—creating structural risk for organizations without explicit AI incident response procedures.
— Consulting firm documents 3-layer IR automation architecture deployed across 50+ production clusters: 40–70% MTTR reduction, automating 60–80% of investigation time with LLM-driven triage and approval policies.
— Commercial agentic SOC platform orchestrating 300+ specialized agents for parallel investigation and response with human-in-the-loop guardrails; demonstrates production-ready autonomous execution architecture.
— Arvo AI Aurora Actions ships agentic playbook automation in natural language with manual, post-incident, and scheduled triggers across 22+ integrations, demonstrating L4/L5 agentic execution framework.
— Analyst synthesis of deployed platforms (Stellar Cyber, Torq HyperSOC, Prophet): 3–10 minute investigations vs. 20–40 human minutes, 85–90% MTTR reduction, 97–98% accuracy, 93%+ true positive reliability.
— Deep practitioner analysis of IR automation maturity: identifies confidence-based case formation, deception-based validation, and tiered response as prerequisites for mature automation execution.
— Arctic Wolf cuts 250 employees (8.3%) to fund Agentic SOC automation platform; third major vendor (after CrowdStrike, Palo Alto) explicitly redirecting hiring budget from analysts to AI-driven incident response.
— Technical guidance on domain-specific agent IR challenges: identifies 6 forensic artifacts (prompt history, context, tool calls) and three incident families distinct from traditional IR, evolving automation practices.
— Palo Alto Networks GA Autonomous Playbooks for XSIAM 3 with zero customization required, auto-updates, and analyst-approval gates for sensitive actions; signals managed automation maturity.