No detection rule fired. The hunt found 9 compromised hosts.
Query Workers ran a 7-day Living-off-the-Land hunt across Windows, Linux & AWS — banking trojan, keylogger, staged wiper, all pre-detonation.
Full audit trail: https://t.co/NWUY37lQLQ
#SIEM#cybersecurity
More time is spent pulling data than analyzing it. Each Query Worker automates a SecOps job across every connected source, builds the evidence chain, and flags what it couldn’t verify. Your analysts make the call.
Full investigation gallery: https://t.co/Ur1cZ0MY4V
#AISOC
Query puts your security data to work. 50+ connectors. Unified data model. No pipelines to build or maintain.
Centralize the insights, not the data.
https://t.co/O6dSkxJnQS
'Detection coverage gaps' aren't detection problems. They're ingestion problems.
S3 logs, SaaS audit trails, EDR data tiered to cold — invisible to your SIEM, invisible to your rules.
Federated Detections runs the logic where data lives. No ETL.
https://t.co/OsfPL0d9hH
We gave a Query Worker one prompt: “hunt for OAuth app-consent abuse”
35 mins later:
• 25 federated queries
• 5 production-ready detections
• 3 critical telemetry gaps identified across Entra/JumpCloud/AWS
That’s AI-native threat hunting on the security data mesh
#AISOC
A Query Worker ran a threat hunt across 5 AWS accounts + Azure from one prompt.
12 queries later it found:
• Read-only roles stopping CloudTrail
• New IAM users created
• Unauthorized admin policy attachments
Then it generated detections + a remediation plan automatically.
What does it cost to index 6 months of CrowdStrike telemetry into Splunk?
More than storing it in S3 and searching it with Query. A lot more.
Dhiraj Sharan walks through how one customer did exactly that:
https://t.co/9YVC0Dv6v1
#cybersecurity#SecDataOps#federatedsearch
Are your SIEM retention periods driven by licensing pressure or investigative need?
20-question self-assessment to find out what your security data architecture is actually costing you: https://t.co/z6zaWHIgnx
#SIEM#cybersecurity#SecDataOps
3 ?'s you should ask your AI security vendor
• Can you see the actual query syntax?
• Can you independently rerun searches/verify results?
• Can you see what failed?
Matt Eberhart on why "explainable AI" usually isn't: https://t.co/TBnQHRkcmI
#AISOC#cybersecurity#SIEM
Here's a Worker running a threat hunt for OAuth app-consent abuse across Entra, JumpCloud, and AWS.
Result?
Thirty-five minutes. Structured hypothesis, 25 queries across three platforms, five detections ready for soak, and a gap inventory that makes the next hunt better.
Most AI security agents have a model. They don't have normalized data.
Okta says actor.alternateId. Entra says userPrincipalName. CrowdStrike says something else.
Agents guess. You get silent hallucinations that read well but aren't true.
The fix: https://t.co/Tv2FrVxfHV
We are excited to announce Mike Black as our new Forward Deployed Solutions Architect!
Mike is a seasoned cyber professional with extensive experience getting customers to value.
In this role, Mike will innovate on architectural success with Security Data Mesh clients.
Most security data pipeline projects start with the wrong question
"How do we move it?" vs. "What does this data need to do once it lands?"
If you move it right, big ROI. If you move it wrong... big mess
Best practices: https://t.co/7lBiuzl7yr
#SIEM#SecDataOps#cybersecurity
Splunk works. The economics don't.
Stop choosing between visibility and budget. Query federates search across 50+ sources — right from your Splunk console. No migration. No ingestion costs.
https://t.co/fF0W2utjoy
#splunk#SIEM#cybersecurity
Running a threat hunt for cloud API abuse across five AWS accounts/Azure.
1 prompt. 12 queries across five AWS accounts. 4 confirmed findings. 4 detections. And a remediation plan. Full attribution on every result.
That's a Worker running a hunt on the Query security data mesh
When you can't ingest all the data you need into your SIEM, detection gaps grow.
Query Federated Detections decouples detection from ingestion.
Write detection logic once, execute it directly against wherever the data lives. No data movement or pipeline project.
#cybersecurity
The average analyst experience is like being a flying cow in a tornado movie.
• Average of 11 consoles per investigation
• 20% time lost to manual data aggregation
• 42% of alerts uninvestigated
Want to help your analysts get back on solid ground?
https://t.co/amNFQVj1ie