Succes Story
%%Zip:%% Boosting SOC throughput with Prophet AI
Key highlights
20,182
alert investigations completed by Prophet AI in the last 12 months
6,727 hours
of manual investigation time avoided
~7 analyst-years
of investigation capacity freed (see methodology)
The problem
Rapid growth at Zip meant more users, more apps, and a steadily rising volume of identity, email, and endpoint alerts. Manual triage and investigation consumed valuable analyst time and created a real risk of backlog. Zip needed a way to process every alert quickly and consistently without expanding headcount.
The solution
Agentic investigation at scale
Prophet AI connects to Zip’s existing security stack, including AWS, Crowdstrike, Okta, Wiz, and more to triage and investigate alerts automatically, gather context across identity, email, endpoint, and SaaS sources, and escalate only the few that merit human attention. Analysts get transparent reasoning, evidence, and audit-ready steps for quick review and action.
The results
- 6,727 hours of analyst time returned:
At 20 minutes per manual investigation, Zip avoided 6,727 hours of repetitive toil.
- Equivalent to ~7 analyst-years of manual investigations:
If one analyst spends ~4 hours/day on investigations, that is ~2,880 investigations/year. Prophet’s 20,182 investigations equal ~7 analysts’ annual manual capacity.
- Per-analyst throughput up to 10×:
When AI performs the investigation and an analyst reviews in ~2 minutes, a single analyst can oversee ~28,800 investigations/year versus ~2,880 manually
- Analysts focused on real threats and projects:
With routine investigations off their plate, the team concentrates on higher-risk incidents, tuning, and proactive work.
Automation is the solution to Zip’s SOC since we’re headcount conscious. However, automation is inflexible, expensive to develop, and drifts as our business changes. Prophet AI delivers context to our analysts expediently so we can get back to designing and implementing security improvements and it doesn’t require the expensive overhead of a SOAR system.