Suhas Manangi
Co-Founder & CEO
Former Trust & Safety and AI product leader with experience at Snap, Airbnb, Amazon, Lyft, and Microsoft.
Controlled AI for investigation operations
Turn your policies into auditable investigation workflows.
Prepare more cases consistently while keeping ambiguous and high-risk decisions with your reviewers.
PreCog AI is configured with your policies, evidence sources, reviewer workflows, decision thresholds, permitted actions, and escalation rules. It prepares evidence-backed recommendations for each case while reviewers retain control.
Your team defines the rules, checkpoints, and final decision path.
Security & Compliance
Enterprise security and privacy controls, with current status maintained in the Trust Center.
SOC 2 Type IIAudit complete
ISO/IEC 27001Certified
HIPAACompliant
GDPRCompliant
The operating pressure
Investigation volume is not the only thing that scales.
Evidence sources, policy exceptions, and review obligations compound with it. PreCog AI gives teams a controlled preparation layer around the workflows they already own.
- 01
Case volume grows faster than reviewer capacity.
- 02
Evidence and prior review context are fragmented across tools.
- 03
Complex policies still require contextual judgment.
Designed for investigation teams
- Trust & Safety
- Fraud & Risk
- Marketplace Operations
- Customer Support
How PreCog AI works
One governed path from operating model to reviewer action.
- 01
Configure the operating model
Connect the policies, evidence sources, reviewer SOPs, decision thresholds, permitted actions, and escalation rules your team already uses.
- 02
Prepare each investigation
PreCog AI organizes case evidence, applies approved policy and configured decision rules, and prepares a recommendation with rationale.
- 03
Review, decide, and record
Reviewers approve, adjust, or escalate. Recommendation rationale and reviewer action remain connected for review.
Reviewer feedback can inform proposed, versioned workflow updates. Changes require approval.
ILLUSTRATIVE INVESTIGATION WORKFLOW
An illustrative high-risk investigation, prepared for accountable review.
A high-risk case arrives with evidence from approved sources. PreCog AI organizes the available evidence, attaches relevant policy context, prepares a recommendation with rationale, and records the human reviewer’s decision and resulting action.
Illustrative example · no customer data- 01Evidence prepared
Available case evidence organized for review
- 02Relevant policy context attached
Approved policy excerpts and decision context
- 03Recommendation prepared
Evidence-linked recommendation and rationale
- 04Human reviewer decision
Reviewer approves, adjusts, or escalates
- 05Rationale and action recorded
Decision rationale, reviewer action, and workflow version
Operational outcomes, retained controls
Move faster without separating the outcome from its control.
Expected operational outcomeControl that stays in place
01More consistent case preparation
✓Recommendations stay connected to evidence and approved policy.
02More reviewer focus
✓Ambiguous and high-risk cases remain at human checkpoints.
03Clearer operational oversight
✓Rationale and reviewer action remain connected for review.
Built by operators
Trust & Safety experience, built into the operating model.
Precognition Labs is led by product and engineering operators who have built integrity systems at scaled technology companies.
Xiuduan Fang
Co-Founder & CTO
Former Head of Platform Integrity Engineering at Snap and long-time Google engineering leader.
Company programs & backing
Frequently asked questions
How controlled investigation AI works.
What is a Trust & Safety investigation workflow?
It is the repeatable path a team uses to gather case evidence, apply approved policy and decision rules, route ambiguity or risk to reviewers, and retain the resulting rationale and action for later review.
How does PreCog AI use an organization’s policies?
PreCog AI is configured with the policies, evidence sources, reviewer SOPs, decision thresholds, permitted actions, and escalation rules the organization approves for a workflow.
Which decisions remain with human reviewers?
Organizations define the checkpoints. Ambiguous, high-risk, exceptional, or otherwise governed decisions can remain with reviewers who approve, adjust, or escalate the recommendation.
Does PreCog AI replace the existing case-management system?
PreCog AI is designed to work alongside the existing case-management system. The system of record and the organization’s reviewer decision path remain in place.
How are recommendations and reviewer actions audited?
The recommendation rationale, referenced evidence, reviewer action, and workflow version remain connected so teams can review how a case moved from preparation to decision.
Start with one workflow