#RSAC26CISO SafeSpace at RSAC26
Self-Learning Platform

Security that earns your trust. Automatically.

Behavioral baselines built and refined automatically. No rules, no tuning.

Supported by security leaders from
Kaiser PermanenteSnowflakeYouTubeThe VenetianEquinixRubrikTalkDeskKaiser PermanenteSnowflakeYouTubeThe VenetianEquinixRubrikTalkDesk
Behavioral Baselining

Your environment is the baseline.

  • Baseline built per entity within 7 days of deployment
  • Models update automatically as your environment evolves
  • Detection gaps from configuration drift close themselves
alaris / entity-baselines
ENTITY BASELINE CONFIDENCE
3,847 entities modeled · updated 2m ago
USERj.morrison@corp
91%
HOSTWIN-WKST-0847ANOMALY
31%
58% deviation from 30-day baseline · flagged for review
PROCsvchost.exe
87%
NET10.0.1.0/24 (DMZ)
77%
APPsap-erp.corp.local
55%
Monitoring 3,847 entities across 12 segments · 1 active anomaly
Analyst Feedback Loop

Every verdict makes detections smarter.

  • False positive dismissals reduce similar alerts automatically
  • Escalation confirmations increase sensitivity to that pattern
  • Accuracy improves continuously without analyst effort
alaris / analyst-feedback-loop
LATEST ANALYST VERDICT
DISMISSEDScheduled task at 02:15 on WKST-042
Known maintenance window
Model update: 47 similar alerts suppressed
DETECTION ACCURACY TREND
78%
Wk 1
84%
Wk 2
91%
Wk 3
96%
Wk 4
THIS MONTH
312
Dismissed
94
Escalated
441
Confirmed
847 verdicts processed this month · model updated 6h ago
Threat Intel Integration

Fresh intel applied instantly.

  • Real-time ingestion from commercial and open-source feeds
  • New IOCs mapped against behavioral baselines immediately
  • MITRE ATT&CK coverage updated without engineering intervention
847
Active IOCs
12
Feed Sources
94
MITRE TTPs
LIVE INTEL FEED
RECORDED FUTUREIP IOC185.220.101.47BLOCKED
MISP COMMUNITYDOMAINupdate-cdn.malicio.usWATCHING
VIRUSTOTALFILE HASHd41d8cd98f00b204e980...BLOCKED
MANDIANTTTPT1059.001 PowerShellACTIVE
MITRE ATT&CK COVERAGE
Initial Access89%
Execution94%
Persistence86%
Defense Evasion78%
Last ingestion: 18 seconds ago · 3 new IOCs applied to baselines
Automatic Re-Baselining

Environment changes handled automatically.

  • Post-acquisition re-baselining without manual configuration
  • Remote work policy changes absorbed without alert storms
  • New application rollouts calibrated within days, not weeks
alaris / rebaseline-event-log
REBASELINE EVENT LOG
Change event detected09:14:02
143 new hosts joined during acquisition onboarding
Re-baselining triggered09:14:04
Isolation window opened for 143 entities
New baselines computed09:21:38
143 entities re-baselined in 7.6 minutes
Detections restored09:21:39
FP rate: 4% — no alert storm occurred
FALSE POSITIVE RATE — ACQUISITION EVENT
Pre-acquisitionSpikePost-rebaseline
7.6m
Re-baseline time
0
Alert storms
4%
Current FP rate
No alert storm · 143 entities re-baselined in 7.6 minutes
How It Works

Smarter from the first alert

Building Baseline
Users0%
Hosts0%
Applications0%
Network0%
01

Baselines Form

  • Statistical baselines built for every entity over seven days
  • Covers users, hosts, applications, and network flows
  • Calibrated to what is normal in your specific environment
Detections Active
Lateral movement91%ENV-TUNED
Credential abuse88%ENV-TUNED
Data exfiltration94%ENV-TUNED
Persistence86%
02

Detections Activate

  • Behavioral detections tuned to your environment activate automatically
  • More accurate than generic signature rules from day one
  • No manual tuning or threshold configuration required
Accuracy Over Time
Month 1
81%
detection
12%
false pos.
Month 3
90%
detection
6%
false pos.
Month 6
97%
detection
1.8%
false pos.
Accuracy compounds automatically
03

Accuracy Compounds

  • Every analyst verdict feeds the learning engine
  • Detection accuracy improves month over month
  • False positive rates drop continuously without analyst effort
Integrations

Works with your existing stack

CrowdStrike
Splunk
Microsoft Sentinel
Elastic
SentinelOne
Okta
Jira
ServiceNow
PagerDuty
Microsoft Teams
Microsoft Defender
IBM QRadar
Palo Alto
Fortinet
Zscaler
Cisco
Check Point
Darktrace
AWS Security Hub
Azure Security
Google Cloud
Wiz
Orca Security
CyberArk
Microsoft Entra
Slack
Recorded Future
MISP
VirusTotal
Mandiant
CrowdStrike
Splunk
Microsoft Sentinel
Elastic
SentinelOne
Okta
Jira
ServiceNow
PagerDuty
Microsoft Teams
Microsoft Defender
IBM QRadar
Palo Alto
Fortinet
Zscaler
Cisco
Check Point
Darktrace
AWS Security Hub
Azure Security
Google Cloud
Wiz
Orca Security
CyberArk
Microsoft Entra
Slack
Recorded Future
MISP
VirusTotal
Mandiant

+ 100 more integrations

See how fast Alaris learns your environment.

95% accuracy in 30 days. Zero tuning.

Why Self-Learning

The gap static rules can't close

Alaris Self-Learningadaptive AI
Manual Processstatus quo
Detection rules
Continuously learning adaptive detections
Engineering backlog required
False positive rate
94% reduction vs. baseline at deployment
High and increases over time
Threat intelligence
Real-time feed integration, auto-applied
Ad hoc on analyst request
Tuning overhead
Zero manual tuning, self-calibrating
Continuous analyst hours
Configuration drift
Automatic re-baselining after environment changes
Periodic manual review cycles
Analyst feedback
Every verdict improves future accuracy
Institutional knowledge only
Further Reading

Additional Resources

Gets smarter every day it runs in your environment