Approach

Decision-led, evidence-weighted.

Cipher’s approach prioritises what changes the decision. Evidence is weighted, uncertainty is stated, and residual risk is converted into controls rather than ignored.

Want a scoped plan for your case?

Send the decision, entities, jurisdictions and timebox. We respond with a proportionate plan and first gate.

See clarity Continue

The loop (why it holds)

Evidence can force scope narrowing, identity disambiguation, or a risk-control pivot. This is designed to remain defensible under constraint.

Engagement shape

A loop, not a line

Each phase produces an artefact. The loop is deliberate: evidence can force scope narrowing, identity disambiguation, or a risk-control pivot.

Pivots allowed

Gate

Define the decision

State what you are deciding, what would change the decision, and what cannot be assumed.

Output: Scope + acceptance thresholds

Accuracy

Anchor identity

Collect identifiers early to prevent misattribution and homonym errors.

Output: Identity anchors + exclusions

Defensible

Weight evidence

Triangulate sources and weight them; label uncertainty plainly where evidence conflicts.

Output: Confidence tiers + limits

Non-binary

Convert residual risk into controls

Where uncertainty remains, define controls, triggers and ownership so decisions can still progress.

Output: Controls + monitoring triggers

Where constraints are tight, we prioritise the checks that change the decision, not the checks that merely add detail.

Evidence weighting

A compact rubric: what carries most weight, what is supplementary, and what is lead-only.

Evidence rubric

Evidence strength and why it matters

We weight sources explicitly. High-confidence anchors reduce false positives and prevent implied certainty.

Tier	Strength	Typical use	Failure mode to avoid
Tier 1 — Authoritative	High	Identity anchors; legal status; registries; enforcement outcomes.	Treating absence as proof; stale filings misread as current.
Tier 2 — Owner-published	Med	Context, claims to validate, product/service descriptions; directional signals.	Self-serving assertions; unverifiable metrics; selective disclosure.
Tier 3 — Reputable media	Med	Chronology, attribution, corroboration; pressure-testing narratives and timelines.	Headline bias; republishing without updates; attribution drift.
Tier 4 — Self-published / aggregators	Low	Lead generation only; used to target primary checks and disambiguation pivots.	Amplification of errors; circular citations; impersonation/spoofing exposure.