Under the hood, Cali looks at the same things your team does: risk bands, typology hints, counterparties, entities, contracts, and OSINT posture. The AI layer makes that context legible, so it can help answer questions like:
- Why does this wallet or case appear high, medium, or lower risk?
- Which typologies and behaviors should we consider documenting?
- Which transactions and hops warrant closer review first?
- How do we explain this exposure to non-technical stakeholders?
AI does not replace analysts, investigators, or law-enforcement partners. It compresses time to understanding. Narratives are drafts, not filings. Risk rationales are suggestions, not final determinations. Every written output can be edited, redlined, and exported with clear authorship.
The goal is simple: less time reconstructing what happened, more time deciding what to do about it—and recording that decision in a way that stands up to review.
Bring your own model (BYO LLM)
Cali orchestrates prompts and workflows across modules, but enterprise teams can route AI calls through their own approved providers (for example OpenAI, Azure OpenAI, Anthropic, or internal gateways) where supported.
- Use your institution’s approved model provider and billing.
- Keep investigation data governance and retention policies under your control.
- Audit-friendly: outputs are tied to what Cali shows in the UI.
- AI assists drafting; analysts stay responsible for conclusions.
Model and provider availability depends on deployment and customer configuration.
Activity
Explains why activity lands in a risk band, which typologies may apply, and where to start reviewing. Helps describe structuring, layering, mixers, and service exposure without losing the underlying table.
Pivot
Generates narratives for buckets of rows and metrics, highlighting unusual flows, structuring-like behavior, and high-risk combinations. Supports multiple voices: executive, investigator, legal, SOC, and minimal technical summary.
Graph
Describes provenance and flow: mixers, bridges, intermediaries, clusters, and off-ramps. Helps connect hops into a coherent path that can be quoted directly in an investigation record or SAR.
OSINT
Summarizes sanctions, PEP hits, adverse media, complaints, filings, and selected social intelligence into a single posture for the entity or wallet in view. Useful for refreshes and updates as cases evolve.
Balance, Assets & Contracts
Explains portfolio composition, token quality, liquidity, and contract powers in plain language. Helps non-crypto stakeholders understand what a wallet actually holds and what those holdings enable.
Report Station
Drafts executive summaries, findings, and provenance sections that align with how teams already write SARs and investigation memos. Keeps text grounded in the evidence and exhibits already attached in the workspace.
Cali is designed for regulated environments. That means human review, clear provenance, and the ability to explain why a narrative or risk explanation looks the way it does.
- Analysts can regenerate, edit, and override any AI text before it is exported or filed.
- Narratives are tied to specific time windows, filters, and views—so you can reproduce the underlying context later.
- Written outputs are intended to be part of an investigative record, not a black box.
The AI layer is there to help your team move faster and write better, not to take responsibility away from the humans signing their names.